Evaluation of the IMF's Role in the Poverty Reduction Strategy

ECIFF

ON

OITA

ULAVETNEDNEPEDNI

I n t e r n a t i o n a l M o n e t a r y F u n d

Evaluation Report

Evaluation of the IMF’s Role in Poverty Reduction Strategy Papers and the Poverty Reduction and Growth Facility

Evaluation of the IMF’s R

ole in PRSPs and the PR

GF

IMF 20

04Evaluation of the IMF’s Role in

Poverty Reduction Strategy Papersand the Poverty Reduction andGrowth Facility, 2004

IEO

Evaluation of the IMF’s Role in Poverty Reduction Strategy Papers and the Poverty Reduction and Growth Facility

I n t e r n a t i o n a l M o n e t a r y F u n d • 2 0 0 4

ECIFF

ON

OITA

ULAVETNEDNEPEDNI

Evaluation Report

© 2004 International Monetary Fund

Production: IMF Multimedia Services DivisionFigures: Lai Oy Louie

Typesetting: Alicia Etchebarne-Bourdin

Cataloging-in-Publication Data

Price: US$25.00

Please send orders to:International Monetary Fund, Publication Services

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

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

recycled paper

Evaluation of the IMF’s role in Poverty Reduction Strategy Papers andPoverty Reduction and Growth Facility / [prepared by a team headed byDavid Goldsbrough] — [Washington, D.C.] : International MonetaryFund, Independent Evaluation Office, 2004.

p. cm. — (Evaluation report)Includes bibliographical references.ISBN 1-58906-367-8

1. International Monetary Fund — Evaluation. 2. Developing countries —Economic policy. 3. Poverty. I. Goldsbrough, David John. II. Evaluation re-port (International Monetary Fund. Independent Evaluation Office).

HG3881.5.I58E93 2004

Foreword vii

Abbreviations and Acronyms ix

EVALUATION OF THE IMF’S ROLE IN POVERTY REDUCTION

STRATEGY PAPERS AND THE POVERTY REDUCTION AND

GROWTH FACILITY

Summary of Major Findings, Lessons, and Recommendations 3

1 Introduction: Origin and Objectives of the PRSP and the PRGF and Scope of the Evaluation 14

Origin and Objectives of the PRS Approach and the PRGF 14Initial Experience with the New Approach 15The Links Between Policies, Growth, and Poverty Reduction 17Scope and Methodology of the Evaluation 18

2 Implementation of the PRS Approach in Areas Relevant to the IMF Mandate 22

Tensions Underlying the PRS Approach 23The PRS Process 23Contents of PRSPs in Macro-Relevant Areas 31

3 Joint Staff Assessments 37

Contents of JSAs 38Value Added of the JSA 39

4 PRGF-Supported Program Design 43

Alignment with the PRSP 45Fiscal Flexibility and Fiscal Adjustment 49Strengthening Public Expenditure Management 51Pro-Poor and Pro-Growth Budgets 52Poverty and Social Impact Analysis 55Is Structural Conditionality More Selective? 57Some Program Design Issues 60Has Program Implementation Improved? 62

5 The IMF’s Way of Doing Business 63

What Was Expected of the IMF Under the New Initiative? 64Have the IMF’s Policy-Formulation Processes and Advice Changed

Under the PRSP/PRGF? 66

Contents

iii

CONTENTS

The IMF’s Contribution on the Ground: Evidence from Case Studies 68IMF Contributions to the Broader Partnership Framework 69Role of IMF Resident Representatives 70Role of Surveillance 71Obstacles to Further Progress in Adapting the IMF’s Role 72

6 Is the New Approach Making a Difference? 74

Changes in Nature of Domestic Policy Processes in Macroeconomic and Related Areas 75

Evidence on What Has Happened to Actual Policies and Final Outcomes 78

Boxes

1.1. Underlying Principles of the PRSP Process 151.2. Defining Characteristics of Poverty Reduction Strategy Papers 161.3. Some Common External Criticisms of the IMF’s Role in the PRS 172.1. Change of Government and Modification of PRSP Strategy:

The Case of Nicaragua 252.2. A Case of Observing PRS Principles Without Achieving Its

Objectives: The Guinean PRSP 262.3. Too Much Paperwork? 272.4. Was Macroeconomic Policy Discussed in the Participatory Process?

Insights from Case Studies 302.5. Why Is It Proving So Difficult for the PRS Approach to Provide an

Effective Operational Road Map on Macroeconomic Policy? 353.1. Key Findings of the Review of JSAs by the World Bank’s OED 403.2. Interactions with Local Stakeholders in the Preparation of JSAs 414.1. Key Features of PRGF-Supported Programs 444.2. Tanzania: Alignment, Fiscal Flexibility, and Program Design 464.3. Experiences with Setting the Medium-Term External Resource

Envelope 474.4. IMF–World Bank Benchmarking Exercise on Public Expenditure

Management 534.5. IMF Technical Assistance Support for Strengthening Public

Expenditure Management 554.6. PSIA in “Real Policy Time” 584.7. Do PRGF-Supported Programs Have an Excessive Disinflation Bias? 615.1. Internal Guidance to IMF Staff on the Implementation of the

PRS Approach 655.2. Views of NGOs: What Is the Nature of Their Policy Differences with

the IMF? 67

Figures

1.1. Logical Framework of the PRSP Approach and IEO Evaluation ofPRSPs and PRGF 19

4.1. Which Came First, the Chicken or the Egg? Sequencing of PRSPs and PRGF-Supported Programs 45

4.2. Relationship Between Initial Fiscal Balance at T–1 and the Targeted Fiscal Adjustment over a Two-Year Period (T–1 to T+1) 51

4.3. Assessment Results on Adequacy of Public Expenditure Management, 2001 54

4.4. Poverty-Reducing Expenditures 56

iv

Contents

6.1. Average Net ODA Flows to Groups of Low-Income Countries 816.2. Growth and Change in Terms of Trade 81

Tables

1.1. Status of PRSPs and PRGF by Region, March 2004 161.2. Country Case Studies for Evaluation 202.1. Possible Categories of Participation and Their Objectives 284.1. Alignment Between PRGF Structural Conditionality and the

PRSP 484.2. Targeted Fiscal Adjustment in ESAF- and PRGF-Supported

Programs, 1995–2003 494.3. Targeted Fiscal Adjustment in PRGF-Supported Programs,

2000–03 504.4. Actual Fiscal Adjustment Under ESAF- and PRGF-Supported

Programs, 1995–2003 524.5. Changes in Poverty-Reducing Expenditures, 1999–2002, and

Projected Changes, 2002–05 554.6. Number of Structural Conditions 586.1. Assessment of Progress in the Country Cases Against Two Possible

Intermediate Objectives 776.2. CPIA Ratings for PRSP and Non-PRSP Groups of Low-Income

Countries 796.3. Selected Policy and Outcome Indicators for PRGF-Eligible

Countries 806.4. Progress Toward the Income Poverty Goal 826.5. Progress Toward Meeting the Millennium Development Goals in

12 Mature PRSP Countries 82

Annexes

1 Survey of PRSP Stakeholders 832 Results from Survey of IMF Staff 863 PRSP Sourcebook: Does It Provide Adequate Space for Policy

Flexibility, and Is It Based on Strong Empirical Evidence? 884 Assessment of the Adaptation of the IMF’s Internal Policy Process 905 Modalities of Participatory Processes: How Broad-Based Has

Participation Been? 946 Qualitative Appraisal of Joint Staff Assessments 977 Determinants of Targeted and Actual Fiscal Adjustment 998 Composition of Targeted Changes in Expenditures and External

Financing in IMF-Supported Programs 1019 Changes in Domestic Policy Processes and Institutions: Evidence

from Case Studies 103

Annex Figures

A2.1 IMF Staff Responses on the Impact of the Key Features of PRGF-Supported Programs on Their Conduct of Fund Business 87

A8.1 Distribution of Programs According to the Targeted Change in Expenditures 102

A8.2 Distribution of Projected Changes in the Availability of Grants and Concessional Loans in ESAF- and PRGF-Supported Programs 102

Bibliography 107

v

Evaluation of the IMF’s Role in Poverty Reduction Strategy Papers and the

Poverty Reduction and Growth Facility

The Poverty Reduction Strategy (PRS) approachconsists of a series of process innovations de-

signed to encourage broader-based participation inthe development of a country-owned, long-term strat-egy for growth and poverty reduction that could alsobe a framework for coordinating donor support. Itwas accompanied by a transformation of the IMF’sconcessional lending facility into the Poverty Reduc-tion and Growth Facility (PRGF). It is too early toevaluate the success of the new approach in achievingits longer-term objectives, especially the extent of re-duction of poverty; progress in this dimension willbecome evident only over a longer period of time.The evaluation has, therefore, focused on intermedi-ate stage outcomes, that is, the quality of the broader-based policy formulation process, the nature of thestrategy and policy framework that has evolved, theinteraction between this framework and the PRGF,and the effectiveness of the IMF’s role. We summa-rize here our major findings as well as the lessons tobe drawn from them and make a number of recom-mendations for the future. The final section reflectson some implications of the evaluation for the longer-term role of the IMF in low-income countries.

Summary of Major Findings

The broad picture that emerges from our study isthat the PRS approach has the potential to encour-age the development of a country-owned and credi-ble long-term strategy for growth and poverty reduc-tion, which could provide an effective framework forcoordinating the efforts of donors and internationalfinancial institutions (IFIs), including the IMF. How-ever, actual achievements thus far fall considerablyshort of potential. This is partly because it is unreal-istic to expect quick gains given the initial condi-tions from which the process started in most low-income countries. But there were also shortcomingsin the design of the initiative that have reduced its ef-fectiveness, including a lack of clarity about the rolethat the IMF should play.

Participation in the formulation of Poverty Re-duction Strategy Papers (PRSPs) was generally

more broadly based than in previous approaches,and most stakeholders involved in the processviewed this as a significant improvement. However,the participatory processes were typically not de-signed to strengthen existing domestic institutionalprocesses for policy formulation and accountability(e.g., through parliament). In a few cases, institu-tional arrangements to sustain the process are begin-ning to develop around the budgetary cycle.

The PRS process has had limited impact in gener-ating meaningful discussions, outside the narrow offi-cial circle, of alternative policy options with respectto the macroeconomic framework and macro-relevantstructural reforms. This reflects in part the absence ofany mechanism to ensure that key issues were airedand the broader debate well-informed. Lack of clarityabout the role of the IMF in this area contributed tothis outcome. In the relatively few cases where abroader debate did occur, there was a positive impacton policy outcomes.

Results in terms of ownership are mixed. The ap-proach has often generated relatively strong owner-ship in a narrow circle of official stakeholders re-sponsible for driving the process, but much lessamong other domestic stakeholders. The perceptionthat the approach is overly influenced by proceduralrequirements of the Bretton Woods institutions(BWIs) is widespread.

In terms of diagnostics, the approach has gener-ally contributed to a significant improvement in un-derstanding the multidimensional nature of poverty,which has implications for designing poverty reduc-tion strategies. However, the approach has beenmuch less effective in identifying constraints to ac-celerating growth and making it more pro-poor. Theapproach has so far not contributed significantly tounderstanding the linkages between growth, povertyincidence, and macroeconomic policies at the indi-vidual country level. These issues present analyticalchallenges that are not necessarily resolved throughparticipation alone.

Strategies outlined in PRSPs generally constitutean improvement over previous development strate-gies, in the sense of providing greater poverty focus,a longer-term perspective, and some results orienta-

Summary of Major Findings, Lessons, andRecommendations

3

SUMMARY OF MAJOR FINDINGS, LESSONS, AND RECOMMENDATIONS

tion. However, most PRSPs fall short of providing astrategic road map for policymaking, especially inthe area of macroeconomic and related structuralpolicies. The focus of most PRSPs is on the compo-sition of public expenditures, especially social sectorspending, with much less emphasis on other aspectsof a broader strategy to encourage poverty-reducinggrowth. Even in the area of public expenditure, theoperational value of PRSPs is often limited, becauseof the still rudimentary nature of most costing andprioritization. In many cases, PRSPs also avoid ad-dressing key strategic choices involving “controver-sial” structural reforms. These weaknesses implythat in most cases PRSPs do not yet provide a policyframework in which PRGF-supported programs canbe anchored.

Except in a few countries where the process is be-ginning to be embedded in domestic institutions,there is limited feedback from initial implementationto policy design. This is particularly problematic inthe area of macroeconomic policy where the originalPRSP was often overtaken by events or proved unre-alistic, and there was little in the PRSP to guidechoices on key strategic trade-offs involved in recali-brating macroeconomic targets.

Capacity constraints have been a severe impedi-ment to progress in the implementation of the PRSapproach. There has been insufficient attention todeveloping a systematic plan of action to strengthencapacity, including in the IMF’s areas of primarycompetence. Budgetary processes are weak, and thelinkages between the PRSP, medium-term expendi-ture frameworks, and budgets are generally poor. Inparticular, public expenditure management (PEM)systems are generally too weak to allow the PRSP toplay a central role in implementing expenditure pri-orities or modifying them on the basis of feedbackon actual costs and outcomes. Strengthening PEMhas been recognized as central to the success of theinitiative and is one area where systematic monitor-ing by the BWIs on the basis of commonly agreedbenchmarks is being undertaken (at least for theHeavily Indebted Poor Countries (HIPCs)). The re-sults suggest moderate progress. In other areas, how-ever, capacity limitations, which constrain policy de-sign and implementation, have largely not beenaddressed systematically, including through BWIcontributions.

On balance, joint staff asessments (JSAs) do notperform adequately the many tasks expected of them.The clarity, candor, and comprehensiveness of theassessment are uneven, with scope for improvementeven in “good practice” cases. This partly reflects abuilt-in bias to reach a “yes or no” signal—which isalways “yes” in practice, encapsulated in standard-ized language. Other factors limiting the usefulnessof JSAs include the lack of explicit benchmarks in

most areas on which to base the assessment. Theirmain contribution has been in giving feedback to theauthorities on weaknesses in the PRSP, but JSAs arevirtually unknown outside the narrow official circleand consequently have no impact on the broader pol-icy debate. They do not incorporate systematic in-puts from development partners and, in practice,have played a limited role in informing lending deci-sions, including those of the BWIs.

The effectiveness of the IMF’s contribution hasvaried considerably across different components ofthe initiative and across countries—with marked dif-ferences between “good” and “average” practice. Itsoverall contribution has fallen well short of the (ad-mittedly very ambitious) goals it set for itself in theoriginal policy documents:

• IMF staff typically did not participate actively toinform the policy debate among domestic stake-holders during the PRS formulation process andto ensure that key macro-relevant issues wereaired. This is because IMF staff generally inter-preted the emphasis on country ownership asimplying that involvement on its part should belimited.

• IMF contributions to developing a better under-standing of country-specific micro-macro link-ages have also been fairly limited. The processhas led to much greater awareness within theIMF of the need for ex ante poverty and socialimpact analysis (PSIA), and this is evident in in-ternal IMF processes, but it has not yet trans-lated into a mainstreaming of such analysis inprogram design.

• On the positive side, there are signs that the“policy space” in the macroeconomic area haswidened—in the sense of greater openness onthe part of the IMF to considering alternativecountry-driven policies—at least in countrieswhere macroeconomic stabilization is no longera pressing issue.

Success in embedding the PRGF in the overallstrategy for growth and poverty reduction has beenlimited in most cases—partly reflecting shortcomingsin the strategies themselves. Nevertheless, programdesign under the PRGF has incorporated greater fis-cal flexibility to accommodate aid flows, and there isno evidence of generalized “aid pessimism” or a sys-tematic “disinflation” bias. Expenditures designatedas poverty reducing have increased markedly since1999, although there are questions about how “pro-poor” some of this spending is. IMF structural condi-tionality has been streamlined, but we have not beenable to reach a definitive conclusion on what has hap-pened to aggregate IMF–World Bank conditionality,which is not monitored by the two institutions. There

4

Summary of Major Findings, Lessons, and Recommendations

were only minor improvements in program imple-mentation under the PRGF.

In terms of outcomes, only tentative messagesemerge at this stage.

• BWI measures of the quality of policies and in-stitutions suggest that PRS countries generallystarted out in a better position than non-PRSlow-income countries, but did not improve at afaster pace.

• Short-term growth for PRS/PRGF countries isonly marginally higher than in the earlier period.However, these countries seem to have weath-ered the worsening of the external environmentin 2000–02 better than other low-income coun-tries, which experienced a decline in growth.

• Evidence on poverty-related outcomes, drawnfrom the parallel evaluation by the WorldBank’s Operations Evaluation Department(OED), is still too limited to draw definitiveconclusions. The most notable improvementsconcern various input- and output-related mea-sures (e.g., number of teachers, school enroll-ment, vaccination rates), but outcomes such asmaternal and infant mortality rates have gener-ally not improved.

To summarize, in cases where the PRS approachhas been integrated into domestic institutional struc-tures, there have been important improvements indomestic policy processes. In most cases, however,while the PRS approach has generally helped makeprogress in the right direction, achievements aremore tentative. Where the PRS approach has beenimplemented as an externally imposed, ad hoc exer-cise, it has acted more as a distraction from strength-ening the domestic policy framework.

Lessons

These broad findings suggest two sets of lessonsthat illustrate where the approach, and the IMF’s roletherein, need strengthening. The first set of lessonsrelates to the structure of incentives that has beengenerated by the PRS approach, which are not wellaligned with the intermediate objectives of the ap-proach. The second set of lessons relates to theIMF’s role.

Aligning the structure of incentives withintermediate objectives

A focus on improving fundamental domestic pol-icy processes in countries is more likely to yieldlonger-term gains, in terms of an improved policyenvironment for effective use of development aid,

than would a traditional focus on particular policymeasures. Whether this refocusing is achieved inpractice depends in large part on the underlying in-centives faced by countries under the initiative. Un-fortunately, actual incentives under the PRS ap-proach do not focus sufficiently on improvements indomestic policy processes and institutions. There arefour specific aspects to this problem.

Too much focus on documents and on BWI-driven procedures

The delicacy of attempting, from outside, to en-courage political systems to move in the direction ofgreater openness, accountability, and participatorypolicymaking has led to too much emphasis on pro-cedural components. In practice, therefore, whatcountries perceive as necessary to receive conces-sional financing and debt relief is to produce specificdocuments following various procedures, includingfor participation. At its best, the approach can stilllead to important gains in domestic policy processes,and it has done so in some cases. At its worst, how-ever, it risks being a distraction from the task ofmaking improvements in existing domestic proce-dures for policy formulation and implementation.

Insufficient scope for treating different countries differently

The philosophy underlying the original initiativewas to allow for considerable country variation inapproach. As implemented, however, there has beeninsufficient scope for treating countries differently.The value added of each of the principles underlyingthe PRS approach varies across countries dependingupon circumstances. Some countries had no experi-ence of strategic planning in the area of developmentand stood to benefit enormously from the exercise ofputting together an integrated poverty reduction andgrowth strategy that is results oriented and providesa long-term perspective. Others had a strong plan-ning tradition but limited experience with broad-based participatory policy formulation; in thesecases introducing a participatory element was poten-tially beneficial. Some countries had both elements,but donors (including the BWIs) were not alwayswilling to treat the country’s own strategy and do-mestic processes as a basis for their partnership withthe authorities. These differences in starting condi-tions illustrate the need for the PRS approach to beimplemented in a manner that capitalizes on existingstrengths and focuses efforts on “the missing ingre-dients” rather than starting anew in all areas. Other-wise, not only can ongoing efforts to improve exist-ing domestic processes be disrupted, but weakdomestic capacities may be used inefficiently.

5


Insufficient benchmarks to monitor progresstoward the intermediate objectives of improveddomestic policy processes, and consequent lack of clarity about the criteria for assessments andfinancing decisions

The PRS approach was largely defined in terms ofbroad principles related to processes and inputs. Morespecific milestones to monitor progress vis-à-vis in-termediate objectives, including for strengthening do-mestic processes according to the broad PRS princi-ples, were expected to be developed at the countrylevel. This has not happened and benchmarks formeasuring substantive improvements in policyprocesses are lacking, both at the country level andinitiative-wide. Moreover, these improvements arelargely not assessed in JSAs, which generally focuson the more procedural aspects of the participatoryprocess as well as on the specific policy content of thePRSP. Ambiguities about intermediate objectiveshave also contributed to diverging expectations of dif-ferent stakeholders and insufficient accountability.

More fundamentally, the PRS approach as de-signed and implemented does not sufficiently addresshow trade-offs between domestic ownership andBWI assessments of policies will be handled. Deci-sions on such trade-offs are still taken, but in a man-ner that (i) is not transparent, with countries left to di-vine what is “acceptable” by looking at variousguidelines and through exchanges with BWI staff;(ii) focuses too much on procedural compliance; and(iii) often leaves country authorities and other domes-tic stakeholders feeling that the BWIs control theprocess partly through unwritten rules of the game.

Asymmetry of commitments, with too little onwhat the BWIs are expected to deliver

PRSs that substantially meet the goals set out inthe original policy papers require considerable tech-nical and analytical capacities, not to mention finan-cial support. However, in the current setting, nothingensures that the capacity-strengthening priorities ofcountries will be addressed, and the links with lend-ing decisions can be ambiguous. While both the IMFand the World Bank are broadly committed to“aligning” their assistance programs to the PRSP, itis not clear in practice how much countries have togain by treating the PRSP as an effective strategicroad map, rather than as a procedural formality.

The role of the IMF

The bold expectations set out in the initial policydocuments require greater changes in the IMF’s wayof doing business than have occurred so far and areprobably overpromising what the IMF can deliver

with existing resources. There are several reasonswhy the IMF’s effectiveness under the new approachhas not matched the original expectations.

Lack of clarity about what the IMF should bedelivering in some areas

The original policy papers may have over-promised what the IMF could reasonably be ex-pected to deliver, given prevailing resource con-straints and its comparative advantage. One exampleconcerns the determination of a medium-term exter-nal resources envelope that strikes the right balancebetween needs, sustainability, and realism as well asthe associated efforts to mobilize the requisite donorsupport. Clearly, these are vital components of theoverall strategy, but it is not clear that the IMF iswell suited to deliver them.

There is also ambiguity about what donors re-quire in terms of a “signaling role” from the IMF, es-pecially when macroeconomic stabilization has beenachieved. This is at the heart of the issue of long-term program involvement, to which we return in thesection “Some Reflections on the IMF’s Longer-Term Role in Low-Income Countries.”

Weaknesses in the PRSPs inevitably impedePRGF-supported programs from adapting fully totheir expected role. In some countries, this will be atransitional issue as strategies are strengthened andbecome better integrated into domestic institutions.However, this will happen slowly in many cases, andhow the IMF role should adapt in such “difficult”cases remains unclear.

Moreover, some country authorities are reluctantto accept a more active role for the IMF, especiallywith regard to the broader policy debate.

Insufficient recognition of the changes the PRSapproach implies for the IMF’s “way of doingbusiness”

The implications of the PRS approach for theIMF “way of doing business” have not yet beenfully acknowledged or acted upon. The approachimplies a very different way of organizing IMF in-puts based on: a country-driven strategy that setspriorities within a long-term timeframe; emphasiz-ing contributions to informing a broader policy de-bate rather than traditional program negotiations;and operating within a “partnership framework”that recognizes explicitly that IMF contributionsare only one part of a broader picture. All of thesefactors mean that the IMF program and surveil-lance roles in PRSP countries will be qualitativelydifferent from those in other countries. But the con-sequences have largely not been spelled out, result-ing in an attempt to address a very different set of

6


challenges with an essentially unchanged institu-tional approach.

For example, it is clear from our discussions withstaff that many, while recognizing the value of thePRS approach, do not see it as implying fundamentalchanges in the way the IMF would contribute to abroad-based policy debate on the macroeconomicaspects of countries’ strategies. However, the evalua-tion indicates that, for the PRS approach to have ameaningful impact in the IMF’s areas of responsibil-ity, encouraging a well-informed debate is critical.

Moreover, the BWIs have not used the PRS ap-proach sufficiently as a mechanism for identifyingpriorities on what they should deliver and for coordi-nating key inputs from other partners, drawing on thecountries’ own priorities. For the IMF, this has meantthat scheduling and integrating inputs from the WorldBank into program design remains difficult.

Recommendations

Based on these lessons, we make six broad rec-ommendations. The first three concern the designand implementation of the PRS/PRGF approach.The other three are directed at clarifying expecta-tions about the IMF’s role and improving its effec-tiveness. Each recommendation outlines a broad di-rection of change rather than specifying a detailedblueprint. This approach seems preferable sincethere are a number of ways the recommendedchanges could be implemented that will involve,inter alia, choices on resource allocation that are bestleft to IMF management and the Executive Board.We give examples of possible steps to illustrate whatwe have in mind, but other approaches to imple-menting the recommendations are also possible.

Aligning incentives and objectives

The rationale underlying the first three recom-mendations is to bring the incentives built into thedesign of the initiative more in line with its interme-diate objectives. The approach we favor is one builtaround (i) greater country-driven flexibility (e.g., toallow countries to choose their own path toward im-proving domestic policy processes); (ii) trans-parency about those country-specific choices, so asto allow other stakeholders to express their views;(iii) monitoring on the basis of explicit benchmarksand strengthened ex post accountability; and (iv) aclearer framework in which donor support is de-cided, including through candid and graduated BWIassessments of countries’ strategies.1

Recommendation 1. Introduce greater flexibilityin the implementation of the PRS approach to fit bet-ter the needs of countries at different stages of theprocess and with different capacities and politicaland administrative systems.

The PRS approach would have greater valueadded if it was made more flexible in practice to fitdifferent country situations. Countries need to be puteven more firmly in the driver’s seat by determiningthemselves:

(1) How the policy formulation, implementation,and monitoring processes will be conducted and builtup over time, and with what rules of the game (e.g.,for opening up the policy process to previously ex-cluded groups or strengthening budgetary processes).Progress would be monitored against an explicit setof country-determined intermediate benchmarks (seeRecommendation 2).2

(2) What the output of these processes will be interms of documents (e.g., PRSP Progress Reports)and on what periodicity they will be prepared, rely-ing as much as possible on domestic institutionalarrangements and reporting vehicles. IMF processrequirements (e.g., linking reviews under the PRGFto completion of specific PRSP documents) shouldbe minimized and oriented around domesticprocesses (so that they do not conflict with domestictimetables and/or duplicate domestic instruments).

Recommendation 2. Shift the emphasis of the ini-tiative from the production of documents to the de-velopment of sound domestic policy formulation andimplementation processes.

This broad recommendation would involve thefollowing elements:

(1) Build in greater results orientation. Countriesshould be encouraged to establish—with help fromBWIs where needed—substantive criteria for judging

7

1An alternative approach to modifying incentives would be toestablish challenging universal minimum standards in a number

of areas (e.g., for participatory processes or ex ante PSIA). See,for example, Oxfam (2004). We do not favor such an approach.Even though driven by the best intentions, it would result in theimposition of even more procedural requirements on PRS coun-tries. Moreover, uniform standards do not allow for the diversityof countries’ situations or for necessary priority setting.

2To give a concrete example to illustrate what we have in mind,a country might propose to broaden participation and monitoringthrough strengthening the oversight functions and capacities ofparliament and at the local government level, with formulationand implementation of its medium-term strategy built around thedomestic budgetary process. In which case, the BWIs should ac-cept such an approach by not pressing for other possible proce-dural components and by accepting for their own internal pur-poses PRS-related documents oriented around these domesticprocesses. The country would establish benchmarks for monitor-ing progress in reaching the objectives it has set for itself, but theBWIs would provide their own frank assessments of the country’sstrategy and the policy processes by which it was derived.


progress toward key intermediate objectives such asdeveloping (i) an operational road map that providesstrategic guidance for setting priorities and resolvingtrade-offs; and (ii) effective institutional arrange-ments for formulating, implementing, monitoring,and updating this road map, with a firm link to bud-get processes.3 The choice of the criteria/benchmarksjudged to be most important would likely vary bycountry, but in some areas broad guidance could bedeveloped by BWI staff or others, upon which coun-tries could draw as a starting point.4

The existing PEM benchmarking exercise is anexample of what we have in mind: it combines coun-try-specific action plans for improving domesticprocesses; clear benchmarks for measuring progress,established with country consultation; and periodicex post assessments that provide inputs for country-specific and initiative-wide assessments (see Chap-ter 4, section on “Strengthening Public ExpenditureManagement”). Given the central importance ofgood PEM systems to setting and implementing pri-orities, this is a tool that all PRSP countries—notjust the HIPCs—should be encouraged to adopt.However, the types of benchmarks chosen by eachcountry will depend on the improvements that theirPRS process identifies as a high priority.

(2) Shift the emphasis of the incentives structurefaced by countries from procedural aspects and pro-duction of documents to achieving substantivechanges in domestic processes and policies objec-tively measured as described in (1) above. The newset of incentives would include:

(i) Transparency. Countries should present theirintentions and objectives, along with thebenchmarks selected to monitor progress in amanner open to public scrutiny.

(ii) Accountability. IMF (and World Bank) staffwould be responsible for providing clear andcandid assessments of the progress made byeach country in implementing the PRS ap-proach, both in relation to the goals set by the

country itself and against initiative-widebenchmarks (see Recommendation 3).

(iii) BWI support. IMF (and World Bank) staffwould help countries identify key constraintsin making progress toward PRS objectivesand support efforts to ameliorate them. Thiswould include not only domestic capacityconstraints but also obstacles stemming frompolicies in advanced economies and othertrading partners.

(iv) Selectivity in lending. Realistically, access todonors’ resources will remain the key incentivefor countries under the PRS approach. Ideally,donor decisions on the volume of resourcesprovided should be linked to the progresscountries are making under the approach. Tofacilitate this, IMF assessments in its area ofexpertise need to provide as clear and candid asignal as possible (see Recommendation 3).The criteria guiding the IMF’s own lending de-cisions under the PRS approach could also beimproved in this regard (see the section “SomeReflections on the IMF’s Longer-Term Role inLow-Income Countries”).

Recommendation 3. Clarify the purpose of the JSAand redefine the vehicle accordingly.

An objective assessment of a country’s efforts to-ward poverty reduction and growth, reflecting thejoint perspective of the staff of the IMF and WorldBank, can be of value to many audiences: (i) to theauthorities, as feedback on how to improve the effec-tiveness of their efforts; (ii) to BWI ExecutiveBoards, as input into their lending decisions; (iii) toother donors, as a signal of BWI views on the qualityof the poverty reduction and growth framework ofthe country; and (iv) to the public, for purposes ofmonitoring and accountability. In our view, there-fore, the JSA is a useful concept whose potential hasnot been realized and we recommend making thefollowing changes:

(1) JSAs should focus on the adequacy of domes-tic policy choices and the quality of domesticprocesses—including monitoring and implementa-tion arrangements, budget decision making, andpublic expenditure management systems—as well asactual progress toward intermediate objectives, andless on the quality of the PRSP as a document.

(2) To foster clear and candid assessments, wepropose: (i) making explicit the criteria and bench-marks used by staff to form their judgments (see alsoRecommendation 2); (ii) reporting on the views ofthird parties (especially local stakeholders anddonors) when available, and discussing differences

8

3This recommendation is similar to the call for an operationalframework made by staff of the European Commission (EC) andWorld Bank in their “Joint Note on Supporting the PRSP Processin Africa” (EC and World Bank, 2003).

4For example, benchmarks on various issues being developedby the New Partnership for Africa’s Development (NEPAD)might be drawn upon by African countries. In other cases, UNagencies (or even international nongovernmental organizations (I-NGOs)) might make suggestions or countries may choosebenchmarks tailor-made to their own priorities and circumstances(e.g., for participatory monitoring of local expenditures). The im-portant point is that the countries themselves should choose whatare the priorities for monitoring progress and not be overburdenedwith a multiplicity of requirements.


of view;5 and (iii) eliminating the need for JSAs toreach a binary (yes or no) conclusion as to the ade-quacy of the PRS as a basis for BWI concessionallending. JSAs should aim to provide a graduated as-sessment of the strength of the PRS and relatedprocesses as well as of the quality of policies.

(3) The effectiveness of the above recommenda-tions could be enhanced if JSAs were produced onan independent schedule (e.g., once a year), ratherthan being linked to a specific PRSP document—which inevitably creates a focus on the documentrather than the process, and creates another proce-dural requirement for the country. This would havethe added benefit that BWI monitoring of PRS im-plementation would continue regardless of PRGF-related developments.

(4) The JSAs would be more effective if, in addi-tion to flagging weaknesses in the PRS, they indi-cated clearly what are the main obstacles to over-come (including those not under the control ofcountry authorities); what the IMF proposes to do tohelp address them in its areas of responsibility; andwhat needs remain unaddressed, especially in thearea of capacity building.

There are a number of ways these key functionscould be carried out and an instrument like the JSA isonly one possibility. For example, an alternativewould be to rely on other instruments within the IMF(e.g., the surveillance process) and the World Bank(the Country Assistance Strategy (CAS)).6 However,we see merits in maintaining a joint instrument, sinceit is more consistent with the partnership dimension ofthe PRS approach.7 Indeed, looking forward, the as-sessment instrument may need to adapt further in thisdirection. For example, if general budgetary supportby bilateral donors to a number of PRSP countriescontinues to expand, a joint “partner” assessment—built around the common performance assessmentframeworks (PAFs) that have been developed for sev-eral countries (e.g., Tanzania and Mozambique), butexpanded to incorporate explicit IMF inputs—mightbe a logical development in the future.

Clarifying the IMF’s role and improving its effectiveness

The practical implications of the adoption of thePRS approach for IMF operations go well beyond

the formulation and negotiation of IMF-supportedprograms and have not been fully recognized. Inorder to orient our recommendations in this area, wenote that two types of decisions need to be takenwith respect to the IMF’s role:

• Decisions concerning the “architecture” of theIMF’s role in low-income countries. This in-volves the nature of the IMF’s signaling func-tion as well as the role of IMF conditionalityand financing in the longer-term PRS frame-work. These decisions go beyond the scope ofthe present evaluation and we do not make spe-cific recommendations. However, we discuss inthe section “Some Reflections on the IMF’sLonger- Term Role in Low-Income Countries”a number of principles that should be taken intoaccount.

• Within the present “architecture,” decisions con-cerning how much the IMF should be directlyinvolved in helping countries implement thePRS approach and changes in the IMF “way ofdoing business” that need to be made, irrespec-tive of choices on the scope of the IMF role, toimprove the effectiveness of the institution’scontribution. These issues are addressed in Rec-ommendations 4–6.

Recommendation 4. Clarify what the PRS ap-proach implies for the IMF’s own operations andstrengthen the implementation of the agreed role.

The thrust of this recommendation is to clarifywhat is expected of the IMF under the PRS approachand to outline changes that would improve its effec-tiveness in meeting these expectations.

(1) IMF engagement in the PRS process

• More emphasis should be given to IMF activi-ties that help to better inform broad-based pol-icy discussions in its areas of competence.Guidelines to staff need to be clarified so as toencourage more active inputs to such discus-sions, including analyzing alternative policyoptions and trade-offs. In exchanges with theevaluation team, some IMF staff expressedconcern that such activities go beyond theFund’s mandate and can potentially conflictwith domestic ownership. However, domesticownership should not be interpreted to meanlack of transparency in what the IMF wouldexpect on macroeconomic policy and the rea-sons why it takes this view. One possible ap-proach might be for the IMF staff to prepare ashort note on key macroeconomic issues thatneed to be addressed in the broader growth andpoverty reduction strategy, which the authori-ties could make public if they wished.

9

5We are not suggesting that the JSA should become a “negoti-ated” document with all of these stakeholders.

6The CAS details the partnership between the Bank and eachborrowing country. Usually covering a three-year period, it definesthe level and composition of assistance to be provided based onneeds, portfolio performance, and overall policy performance.

7Moreover, unlike the JSAs, the alternative vehicles are not au-tomatically made public.


• Rather than establish uniform “standards” forthe IMF’s role, expectations should be tailoredto country-specific circumstances, includingthe government’s wishes. These country-spe-cific “rules of the game” should be made pub-lic—perhaps in the form of an aide-mémoireattached to the JSA or the Article IV staff re-port—and could describe how IMF staff areexpected to participate in the broader policydebate, including what supporting analyticalwork they will undertake. This is also likely torequire a combination of: (i) more “standalone” missions, set apart from program nego-tiations, and (ii) enhancing the role of residentrepresentatives.

• There should be more systematic explorationsof country-specific macro-micro linkages—boththrough the IMF’s own analysis and research ac-tivities and by drawing more systematically onthe work of others.

• Article IV surveillance reports should be usedsystematically as a vehicle to convey the IMF’sown thinking and analysis on key issues, and thestaff should routinely seek the views of the au-thorities and others on the types of issues thatshould be addressed in surveillance.

• Assess systematically, as part of broader IMFsurveillance activities, obstacles to the achieve-ment of PRSP objectives originating in tradeand subsidy policies of main trading partners(drawing, to the maximum extent possible, oninputs from other agencies). The Global Moni-toring Report (World Bank and IMF, 2004a) hasalready made important steps in this direction.

(2) PRGF-related activities

• The rationale for IMF policy recommendationsand program design should be subjected tobroader scrutiny and debate. Possible steps inthis direction could include (i) facilitating widerdissemination and discussion of IMF analysisthat forms the basis of its policy advice (includ-ing, where appropriate, technical assistance re-ports); and (ii) encouraging greater openness to“independent/external voices” as inputs into pro-gram design, when agreeable to the authorities.8

• Clarify the approach to be taken by the IMF inthose cases where the PRS approach has addedsome value but has not yet produced an opera-tional road map or the necessary institutionalframework for implementation. Even in thesecases, there may be significant scope for openingup the policy space and incorporating evidenceon macro-micro linkages (including PSIA) moresystematically into program design, even thoughthe priorities for such activities will not be drivenby a coherent country-driven strategy.

• Clarify what the BWIs are trying to achievejointly through the streamlining of conditional-ity and how this fits with stronger domesticownership; establish a system for the monitor-ing of aggregate Bank-Fund conditionality atthe country level.

(3) Streamline IMF documentation and Boardscrutiny of PRS documents

Exactly how this streamlining should be done willdepend on what role the JSA plays (see Recommen-dation 3), but there seem to be two possible ap-proaches: (i) make the JSA the central assessmentinstrument, with operational meaning—in whichcase “comfort” letters and detailed analysis of thePRS strategy in other staff reports (Article IV orPRGF-related) would be unnecessary; or (ii) incor-porate the functions of the JSA into surveillanceand/or program documents. In any event, IMF Boarddiscussion of the PRSP itself seems unnecessary (al-though Directors should continue to receive PRSPsfor information). At the initiative-wide level, sepa-rate annual reports on Progress in Implementing thePRSP Approach and the Global Monitoring Reporton the Millennium Development Goals (MDGs) ap-pear duplicative and could be merged.

Recommendation 5. Strengthen prioritizationand accountability on what the IMF itself is sup-posed to deliver within the broader partnershipframework, built around the priorities emergingfrom the PRS process, and ensure that resourcesmatch commitments.

The IMF should tailor its involvement moreclosely to country needs, taking into account thecontributions from other partners. Because theseneeds vary widely and because resource constraintson the IMF will remain tight, a more systematic ap-proach is needed to (i) setting priorities on what theIMF itself will deliver; and (ii) ensuring appropriatecoordination of key inputs from other stakeholders.Prioritization and coordination could be improved asfollows:

(1) Generate, as part of the PRS process, specificpriority actions for the IMF to assist the country con-

10

8The Tanzania case is a good example and illustrates that achange in the “way of doing business” toward strengthening thecapacity of countries to engage in discussion on such policy ad-vice requires primarily a change in attitude and organizational ap-proach. We are not suggesting that the IMF has to provide the re-sources for any additional outside inputs, but bilateral donors maywish to do so. Indeed, in the Tanzania case the additional analysiswas financed by a bilateral donor.


cerned to reach its objectives, including for analyti-cal inputs and long-term capacity building. If theIMF is not well-suited to produce a particular input(e.g., some PSIA), but this input is judged critical forprogram design, a clear identification—stemmingfrom the PRS priorities—of who has agreed to dowhat and by when would strengthen incentives. Suchkey deliverables would be agreed with the govern-ment and made public. These priority actions wouldthen be the basis for evaluations of IMF perfor-mance, both at the country level and in aggregate.

(2) The IMF’s own budget decisions on allocationof administrative resources should be geared to thesepriorities. Linking realistic resource allocation deci-sions to a more explicit set of priorities should helpimprove decision making and make clearer to allwhat the IMF has committed to, and what it has not.However, the evaluation suggests that the ambitiousoriginal expectations of the IMF’s role in the PRSapproach would require more resources than havebeen provided so far.

(3) Experiment with broader “external reviews”of the PRS/PRGF process, monitoring in particularthe performance of donors and international finan-cial institutions (IFIs) in providing support, and notjust performance of the national authorities. Tanza-nia’s Independent Monitoring Group (IMG) is onepossible example of such an approach.9

Recommendation 6. The IMF should encouragea strengthening of the framework for establishingthe external resources envelope as part of the PRSapproach.

The present “architecture” of the PRS approachdoes not provide a clear framework for helpingcountries and donors decide what is an appropriatemedium-term resource envelope in which the macro-economic strategy should be derived. The evaluationhas not produced sufficient evidence for us to rec-ommend a particular “right” approach on this issue.However, we would not favor a greatly expandedrole for the IMF that would risk taking it beyond itscomparative advantage. Elements of a solutionwould include:

(1) The country itself, not the IMF or WorldBank, should eventually play the central role in elab-orating macroeconomic frameworks and catalyzingdonor support. The pace at which this transition canbe made will depend on specific capacity constraints

in each country, but country leadership seems essen-tial to “owning” the process.

(2) The IMF role would be to provide debt andmacroeconomic sustainability assessments and judg-ments on the policy framework, but it would not beresponsible for the “normative” judgment on appro-priate aid levels over the medium term.

(3) The tension between “ambition” and “realism”in determining the external resource envelope can per-haps be handled by presentation of alternative projec-tions (consistent with assumptions of stronger policyreforms as well as additional external financing). TheIMF should provide increased analytical support forsuch approaches when requested. But the choice toprepare alternative projections should remain with thecountry and not be a uniform requirement.

(4) Improving aid predictability is a wider problemthat the IMF cannot resolve on its own. Consistentwith the PRS approach, donors should be making andimplementing firmer medium-term commitments,linked where possible to country budget processeswith clear triggers for financing decisions. However,this is only feasible if the PRS process generates aneffective operational road map. The challenge is howto reconcile PRSP countries’ concern for aid pre-dictability with donors’ concern for aid effectiveness,which implies some performance-based selectivity.For the IMF, this requires finding a way—perhapsthrough a strengthened JSA—to provide signals to thedonor community on macroeconomic performancethat are sufficiently calibrated, and take account of thelonger-term framework of donor involvement, to be auseful input into selectivity decisions without provid-ing excessive “on-off” signals for financing. We re-turn to this issue below.

Some Reflections on the IMF’s Longer-Term Role in Low-Income Countries

Many of the issues raised in this evaluation haveto be addressed in the context of the larger issue ofthe IMF’s longer-term role in low-income countries.A discussion on this subject has already begun.10 It

11

9The IMG was appointed jointly by the Government of Tanza-nia and donors in 2002 to examine the factors that underminedthe impact of aid and its contribution to supporting Tanzania’sPRS. In particular, the IMG would monitor the policy dialogue,financial management and governance issues, predictability ofaid flows, and the nature and use of technical assistance.

10Since the summer of 2003, several discussions were held inthe IMF’s Executive Board on various aspects of the role of theFund in low-income countries. This discussion is still ongoing.The staff papers on which the Board discussions were based,along with the preliminary conclusions reached by the Board, areavailable on the IMF’s website (www.imf.org). The key papersare as follows (starting with the most recent): “The Fund’s Sup-port of Low-Income Member Countries: Considerations on In-struments and Financing” (February 24, 2004); “Fund Assistancefor Countries Facing Exogenous Shocks” (August 8, 2003);“Role of the Fund in Low-Income Member Countries Over theMedium-Term—Issues Paper for Discussion” (July 22, 2003).


is beyond the scope of this evaluation to make de-tailed recommendations on the strategic policychoices involved, but our investigation has high-lighted a number of unresolved issues—some ofwhich have not yet been aired sufficiently in theIMF’s own papers on the subject—that will requirethe Executive Board to take several critical deci-sions. We summarize here the key decisions that willbe needed and suggest a number of principles thatshould be borne in mind as they are taken.

The key decisions are:

(1) What should be the signaling role of the IMFin the longer term and how should it be provided?

The revealed preferences of major developmentpartners and of many low-income countries suggestthat if substantial donor support is to continue, anIMF signal of the quality of macroeconomic policywill be required. This certainly requires an IMF sig-nal that is credible and clear, but it may or may notrequire financing and conditionality. Decisions onthe signaling role should also distinguish between (i)the need for judgments on the overall macroeco-nomic framework and strategy (for which a JSA-type vehicle may be needed), and (ii) monitoring ofshorter-term performance where needed (a role tra-ditionally performed by IMF conditionality; see (3)below).

(2) What should be the IMF’s financing role inlow-income countries and what should be the crite-ria for “exit” from such financing?

The existing criteria for such financing are basedon the concept of “protracted balance of paymentsneed,” which is so vague as to be difficult to distin-guish from development financing in practice. As-sessments of a protracted balance of payments needare expected to consider not just the country’s over-all balance of payments position at the time of therequest for financing, but also “past, present andprospective external performance [indicators], suchas export and import performance, terms of trade de-velopment, access to capital markets, cost of debtservice, foreign reserves [and] medium term financ-ing gaps that the [country] might experience in itsgrowth oriented adjustment effort.” A present bal-ance of payments need is not required. These criteriado not provide effective guidance for when the IMFshould shift its support from traditional lendingarrangements to other forms of engagement and inpractice have justified continued financing.

In contrast, there has not been enough emphasis onthe implications of the PRS process itself for an “exit”strategy from IMF lending. Unbundling the variousnonlending functions that the IMF is expected to per-form under the PRS approach (see Recommendation4) makes clear that the IMF can remain very substan-

tially involved in the PRS process without providingbaseline financing over long periods.11

(3) What will be the nature of IMF conditionalityin the longer term as the PRS approach takes root?

This will require decisions on what is the ratio-nale of IMF conditionality in a longer-term, partner-ship-oriented framework where everyone recognizesthat donors are “in for the long haul” and there is astrong emphasis on country ownership.12 One possi-ble route would be to restructure IMF conditionalityto fit within a broader “partnership” approach tomonitoring and assessing progress in implementingthe PRS. For example, the IMF could join the collec-tive approaches to monitoring PRS progress that areevolving between those donors (including the WorldBank) providing direct budgetary support (e.g., thePAFs for Tanzania and Mozambique). However,even within such a framework, the IMF (and its part-ners in the PRS approach) might come to the conclu-sion that there are grounds to maintain a separatequantitative monitoring of the short-term evolutionof key macroeconomic policy variables—given thatproviding short-term feedback on key budget con-straints and macroeconomic sustainability is a cen-tral IMF contribution. But such monitoring need notnecessarily be tied to IMF lending, nor should con-tinuation of other flows be mechanically interruptedon the grounds of nonperformance of short-termconditionality.13

(4) What should be the IMF’s role in strengthen-ing the framework for establishing the external re-sources envelope that countries may count on?

12

11Present policies call for access to PRGF resources to declinewith successive three-year arrangements. However, policies thatcontemplate defining the appropriate levels of access for, say, thefourth and fifth rounds of three-year PRGF arrangements (i.e., 12or 15 years into the initiative) raise very uncomfortable questions.It is hard to see any justification for such protracted, albeit low-level, financial involvement except on the grounds that only tradi-tional programs carry sufficient credibility as signaling devices(either with the country or donors).

12The earlier IEO evaluation of the prolonged use of IMF re-sources suggests that traditional approaches to conditionality maylose their effectiveness in such situations, especially with regardto the types of reforms that take time and cannot easily be brokendown into specific short-term conditions. (See IEO, 2002, Chap-ter 5, especially paras. 35–37).

13This concern was raised by a number of I-NGOs in their in-puts to this evaluation. The Mozambique case study suggests apossible approach. Delays in completing reviews under thePRGF-supported programs flagged emerging concerns to donorsbut did not trigger an immediate interruption of financing bythose donors providing direct budgetary support. Instead, activeparticipation by IMF staff in a donors’ forum served as a mecha-nism for signaling what progress was being made in resolving theoutstanding issues. An agreement was eventually reached and theprogram review completed, without any interruption in thesebroader financing flows.


As noted in Recommendation 6, the evaluationevidence suggests no “right” answer to this ques-tion—in part because the IMF has not, until re-cently, encouraged much experimentation with dif-ferent approaches. However, elaboration of such aframework is not something the IMF can do on itsown; all development partners will need to be in-volved. Part of the agreed framework should be un-derstandings on the implications of IMF signalingon short-term macroeconomic performance forother financing flows.

We recommend that the following principles helpguide these decisions:

• The approach to the IMF’s role in the architec-ture of the PRS should build upon the specificcomparative advantage of the IMF. These includeanalysis and policy advice on macroeconomicand related structural policy issues, which arepreconditions for sustained pro-poor growth.They should incorporate a focus on managementof budget constraints and longer-term sustainabil-ity issues; technical assistance and capacity build-ing in these areas; and a relatively short-term fi-nancing role to help cushion against adverseshocks that might push the longer-term growthand poverty-reduction strategy off track.

• The IMF’s contribution in low-income countrieswill be only one part of a broader, partnership-based framework, and its modalities should bedesigned with this framework in mind. In otherwords, strong coordination elements rather than“stand alone” approaches need to be built infrom the outset, whether it be with respect toPSIA, conditionality, or decisions on overall ex-ternal financing.

• The choice of appropriate instruments to pro-vide signaling should be based on the needs oflow-income countries and their developmentpartners. However, the credibility of differentsignaling devices will depend partly upon the“architecture” chosen. At present, programs as-sociated with IMF lending arrangements are so

central to the IMF’s role in low-income coun-tries that nonprogram signaling tends to haveless credibility. This could change if a differentstructure, with new signaling and financing in-struments, were adopted.

• The modalities of the IMF’s role should bemindful of the incentives created and should bedesigned to encourage countries to press aheadwith their policy and process changes under thePRS approach. To give an example, one possi-ble way of integrating the IMF’s relativelyshort-term financing into the PRS approach thatmight be consistent with such a principle wouldbe for the PRGF not to provide financing whenthere is no immediate balance of payments need(nor to design somewhat arbitrary phaseoutschedules), but to incorporate provisions forgreater contingent ease of access to financingwhen needed because of adverse exogenousshocks—provided that the JSA (or its successor)has given a sufficiently positive signal on thecountry’s progress in implementing the PRS.

• Decisions on the intensity and length of IMF in-volvement in PRS countries and the way theIMF organizes its operations and allocate itsstaff resources to these countries need to bemade in an integrated manner, so as to ensureconsistency between commitments and availableresources.

To summarize, there can be no doubt on whetherthe IMF has a longer-term role to play in low-income countries. It clearly does, if only becausemacroeconomic stability is widely accepted as a pre-condition for sustainable pro-poor growth. However,it is important that the IMF’s role is kept within theinstitution’s comparative advantage. The challengeis to translate that basic approach into two sets ofchoices: first, on how to “embed” the IMF’s contri-bution most effectively into a broader partnershipframework and, second, what should be the intensityand scope of the IMF’s involvement in that broaderframework.

13

The IMF and the World Bank introduced thePoverty Reduction Strategy Paper (PRSP) ap-

proach in 1999 to strengthen their approach to provid-ing assistance to low-income countries, including bothnew financial assistance and debt relief under the en-hanced Heavily Indebted Poor Country (HIPC) Initia-tive. The new approach was accompanied by thetransformation of the Enhanced Structural AdjustmentFacility (ESAF)—the IMF’s concessional lendingwindow—into the Poverty Reduction and Growth Fa-cility (PRGF), with a view to giving a more centralrole to pro-poor growth considerations in the designof IMF-supported programs in low-income countries.

The purpose of this evaluation is to assess the ef-fectiveness of the Poverty Reduction Strategy (PRS)approach, especially as a framework for the role of theIMF in low-income countries and for the delivery ofIMF concessional assistance, the IMF’s own contribu-tion to the approach, and the PRGF.1 Since actualprogress toward the objective of poverty reduction canonly be assessed over a longer-term horizon, our eval-uation focuses primarily on intermediate outcomes,especially changes in policy processes and policies,not on final outcomes. The evaluation assesseswhether the approach is bringing about the fundamen-tal process changes expected of it—both in the coun-tries themselves and in the two Bretton Woods institu-tions (BWIs) and whether some course correctionsmight be needed at this stage. A parallel evaluation ofthe PRS process from the perspective of the WorldBank’s role has been undertaken by the OperationsEvaluation Department (OED) of the World Bank.2

Origin and Objectives of the PRS Approach and the PRGF

Adoption of the PRS framework and the conver-sion of the ESAF into the PRGF signaled a shift in ap-proach by the BWIs in their support for low-incomecountries. The PRS approach drew on key elementsidentified in earlier work, including the Comprehen-sive Development Framework (CDF) developed bythe World Bank, as well as on the debate that tookplace on strengthening the link between debt reliefand poverty reduction.3 It was also intended to ad-dress concerns identified by internal and externalevaluations of the ESAF and the related PolicyFramework Papers (PFPs). These reviews concludedthat PFPs had largely failed to reach their objectivesand highlighted a number of problems with ESAF-supported programs: (i) lack of national ownership;(ii) weaknesses in the analytical and empirical basesof the social policy content of programs; and (iii) in-sufficient attention to trade-offs involving policychoices that imply significantly different paths forgrowth and social welfare.4

Thus, the new approach was intended to strengthencountry ownership, enhance the poverty focus ofcountry programs, and provide for stronger collabora-tion between the BWIs and more broadly among de-

Introduction: Origin andObjectives of the PRSP and thePRGF and Scope of theEvaluation

CHAPTER

1

14

1In this report, we will use the acronym PRSP only when refer-ring specifically to the document itself. In all other cases, we willuse the term Poverty Reduction Strategy (PRS).

2In accord with the terms of reference for the evaluation, wehave not attempted to assess a number of important issues con-cerning the PRS where the primary responsibility lies with theWorld Bank. The terms of reference for the evaluation are avail-able on the IEO’s website (www.imf.org/ieo). The OED’s Ap-proach Paper is available at www.worldbank.org/oed.

3The framework for the approach was introduced in two back-ground documents for the September 1999 Annual Meetings ofthe World Bank and IMF (1999a and 1999b). A suggested ap-proach for the development of PRSPs and the role of the IMF andWorld Bank staffs were discussed in World Bank and IMF(1999c).

4For example, the External Evaluation of the ESAF (IMF, 1998,p. 36) noted that: “The predominant view—and many ministersand senior officials echoed it with some disappointment—is thatalthough initially the PFP process had held great promise, it hasbecome a rather routine process whereby the Fund brings uniformdrafts (with spaces to be filled in) from Washington, in which evenmatters of language and form are cast in colorless stone. Many se-nior officials expressed the view that the PFP has become so uni-form that it is difficult to distinguish one from the other.”

Chapter 1 • Introduction

velopment partners in supporting country efforts.Other intermediate objectives included greater publicaccountability and an improved setting of prioritiesand design of public actions.

The new approach focuses on developing a coun-try-driven process with broad-based participation, toevolve a comprehensive strategy for poverty reduc-tion based on a long-term perspective. Five underly-ing principles were enunciated that would guide theprocess in each country (Box 1.1). The process leadsto the development of a PRSP with three purposes:

• for the country, to lay out realistic but challeng-ing poverty objectives, along with the policiesneeded to achieve them;

• for the BWIs, to provide a suitable basis fortheir concessional lending; and

• for other development partners, to offer a key in-strument around which to organize their rela-tionship with low-income countries.

Initially, there were no specific standards laiddown that PRSPs were expected to meet. The ideawas that the more general the approach, the better itcould be tailored to different country needs, so thateven though low-income countries face tremen-dously diverse conditions, they would in principle allbenefit from the approach. Subsequently, in responseto countries’ requests for greater clarity, a number of“expectations” were outlined (e.g., in guidelines forJSAs to be prepared by IMF and World Bank staff)regarding the contents of PRSPs, but countries werefree to pick and choose, at least in theory (Box 1.2).

To underline the dynamic nature of the PRSprocess, it was emphasized that the production of aPRSP would not be a one-off exercise, but an itera-tive one, starting with an interim-PRSP, leading to afull PRSP, followed by annual PRSP progress re-ports and, after three years, a new PRSP, with resultsfrom implementation and monitoring feeding backinto formulation of the subsequent strategy papers.

For the IMF, the fact that PRGFs would be an-chored in a country-owned policy framework wasexpected to be a major improvement over the ESAF.Seven key features were identified that would distin-guish PRGF-supported programs from those sup-ported by the ESAF. (These are discussed in moredetail in Box 4.1.)

The policy papers introducing the new initiativesstressed (i) the need to be realistic about what couldbe achieved in the near term; (ii) that the degree ofprogress would depend on initial starting conditionsand the nature and content of PRSPs would varyfrom country to country—as would the participatoryprocess involved in their creation; and (iii) that theprocess would be a dynamic one—as over time bothcountries and donor institutions should learn by

doing. However, there was no doubt that the new ap-proach was intended to mark a significant change inthe IMF’s role and way of doing business in low-income countries.

Initial Experience with the New Approach

As of March 2004, 37 countries, almost half ofthe PRGF-eligible total of 77, had completed a fullPRSP (Table 1.1). A third of these had also producedtheir first PRSP progress reports, and 5 countries hadmore than one progress report. Of the full sample ofPRGF-eligible countries, about two-thirds have sofar had arrangements under the PRGF.

The initial experience was reviewed in reports onProgress in Implementation, which were initially pre-pared on a six-monthly basis, and since September2002, on an annual basis. Each of these reports pro-vided an opportunity to clarify expectations in someareas—with a strong focus on the contents of thedocuments to be produced by countries. In addition, amore comprehensive review of the PRS approach,discussed by the Boards of the IMF and World Bankin March 2002, incorporated extensive inputs fromexternal stakeholders.5 It also led to the identificationof a number of “good practices” expected to enhancethe effectiveness of the approach, directed not only tocountries engaged in the PRS approach but also tothe donor community and the BWIs.

All of these reports concluded that progress underthe approach was “encouraging” and judged it to be“on track,” although the criteria on which these judg-ments were based were not spelled out and progresswas not measured in a systematic way across coun-

15

Box 1.1. Underlying Principles of thePRSP Process

1. Country-driven involving broad-based participa-tion.

2. Results-oriented and focused on outcomes thatbenefit the poor.

3. Comprehensive in recognizing the multidimen-sional nature of poverty and in the proposed pol-icy response.

4. Partnership-oriented involving coordinated par-ticipation of development partners.

5. Based on a long-term perspective for poverty reduction.

5See IMF (2002b).

CHAPTER 1 • INTRODUCTION

tries, except in the sense of reporting increasingnumbers of Interim PRSPs (I-PRSPs), PRSPs, andPRSP Progress Reports (PRSP-PRs) completed. TheSeptember 2002 “Progress in Implementation” re-

port also made clear that “substantial challenges” re-mained in the following areas: building capacity,opening up the policy dialogue, aligning external as-sistance behind national strategies, integrating the

16

Box 1.2. Defining Characteristics of Poverty Reduction Strategy Papers

The JSA guidelines state an “expectation that, although the specific content of PRSPswill vary widely across countries, a PRSP will include four core elements:”

Core elements

1. Description of the participatory process.2. Poverty diagnosis.3. Targets indicators and monitoring systems.4. Priority public actions.

In addition, the JSA guidelines indicate a series of topics that PRSPs are expected toconsider. The most relevant to the IMF’s areas of responsibility are:1

Four pillars (of priority public actions)

1. Macroeconomic framework.2. Pro-growth structural and sectoral policies.3. Policies for social inclusion and equity.4. Governance and public sector management.

Analytical building blocks

1. Assessment of the impact of past policies on growth and poverty.2. Analyses of the key constraints to growth and poverty reduction.3. Analyses of the sources of growth.4. Poverty and social impact analysis.

Building blocks of the macroeconomic framework

1. Macroeconomic program (including growth projections and key fiscal choices).2. Financing plan.3. Prioritized action plan.4. Costing of the action plan.

1The guidelines note that “it is not expected that all PRSPs will address thoroughly all of thequestions listed.”

Table 1.1. Status of PRSPs and PRGF by Region, March 20041

First PRSP Second PRSP Third PRSP Number of Progress Progress Progress Countries Interim PRSP PRSP Report Report Report PRGF2

Africa 35 (12) 26 (10) 19 (12) 7 (7) 3 (3) 1 (1) 28 (12)Asia and Pacific 18 (2) 5 (2) 5 (2) 1 (1) 0 0 7 (2)Europe 2 (1) 2 (1) 1 (1) 1 (1) 0 0 2 (1)Middle East and Central Asia 13 (4) 8 (3) 8 (4) 1 (1) 1 (1) 0 9 (4)Western Hemisphere 9 (4) 4 (3) 4 (4) 2 (2) 1 (1) 0 5 (4)

Total 77 (23) 45 (19) 37 (23) 12 (12) 5 (5) 1 (1) 51 (23)

1Includes all (77) PRGF-eligible countries—the total has not changed since the late 1990s. Cases from the sample of 23 countries with full PRSPs as of end-2002 arein parentheses.

2Countries with an active PRGF-supported program as of March 2004.


PRS into budgetary priorities, and implementing thestrategies outlined in the PRSPs. Throughout thesereports, calls were made for flexibility in implemen-tation and realism on near-term expectations ofchanges.

There have also been numerous external reviews ofdifferent aspects of the PRS approach. These externalreviews reflect widely divergent expectations aboutwhat the new approach is meant to deliver, howquickly, and with what contributions from each groupof stakeholders, in particular the BWIs themselves. Anumber of criticisms directed at the IMF emerged andwe discuss them in our evaluation (see Box 1.3).

The Links Between Policies, Growth,and Poverty Reduction

Since the evaluation is about how the IMF canhelp countries improve their prospects for growthand poverty reduction, it is worth emphasizing at theoutset that knowledge of the links between policiesand growth remains limited and understanding of thelinks between policies and poverty reduction evenless so. While we do not propose to present a com-prehensive literature review, there appears to bebroad—albeit not universal—support for the follow-ing messages:

17

Box 1.3. Some Common External Criticisms of the IMF’s Role in the PRS1

In considering the messages emerging from thisevaluation, a review of the main external criticisms ofthe IMF’s role in the PRS/PRGF initiatives can behelpful.

Most external critics generally begin with a recogni-tion that the PRS process represents a significant stepforward, notably in opening up new spaces for policydialogue and giving various groups, including civil so-ciety representatives, greater access to policy debatespreviously closed to them.

Focusing on issues most directly relevant to the IMF,the main criticisms on process are:

• There has been too little broadening of the partici-patory debate on macroeconomic policy, althoughspecific country experience varies. Even withingovernments, the debate is generally too narrowlybased.

• The policy discussions and decision-makingprocesses are often not well-embedded in existingpolitical structures (e.g., role of parliaments is toolimited).

• Alternative policy options—especially ones thatdeviate from the so-called “Washington Consen-sus”—are rarely explored. Donor (and, in themacroeconomic area, IMF) priorities still drive theprocess too much.

• PSIA of key macroeconomic policy proposals arerarely undertaken and do not represent a significantex ante input into policy formulation.

• The linkage to the HIPC was partly responsible forrushed procedures that reduced the value added ofthe new approach.

On PRS content and the design of PRGF-supportedprograms, the main criticisms are:

• The PRGF still drives the PRSP on the macroeco-nomic framework and related policy issues.

• Program design is still insufficiently oriented to-ward poverty reduction. (However, different ob-servers emphasize different priorities; some focuson the need for programs to expand further pro-poor spending on key social sectors (health, educa-tion, and so on), while others criticize an insuffi-cient emphasis on strategies to improve incomes ofthe poor, as part of an alternative growth strategy.)

• Programs target too much reduction in fiscaldeficits and inflation, to below thresholds at whichthere is clear evidence that further macro-stabiliza-tion is good for the poor and growth.

• The IMF is still seeking to impose conditionalitythat is not derived from the country-driven PRS.

• The IMF is guilty of “aid pessimism.” There areusually two distinct strands to this accusation: (i)programs are designed around projected reductionsin aid flows, perhaps driven by an overall desire toreduce aid dependency. (ii) Macroeconomic frame-works in the PRSP/PRGF do not (but should) beginfrom a “needs-based” approach that takes as itsstarting point what external resources are needed tohelp countries progress toward achievement of theMDGs. Consequently, the IMF is failing to performits “catalytic” role of signaling to donors what aidwould be required to truly make a difference inreaching the MDGs.

This brief summary suggests that part—but not all—of the criticism of the IMF’s role reflects differences ofview about the underlying objectives. In particular, it isclear that a number of external observers—includingvarious I-NGOs—have a more radical view of the objectives of broad-based participation, the desirableapproach to catalyzing “need-based” aid, and the scopeof the IMF’s contribution to these activities.

1See, for example, Buira (2003); Coyle and Evans (2003aand 2003b); Evrard (2003); Gomes and Lawson (2003); Kil-lick (2002); Richmond and Ladd (2001); Oxfam (2003); andWorld Development Movement (2001).


• There appears to be a robust association be-tween reduction in absolute poverty and sus-tained, strong growth.

• Some policies and factors do seem to be associ-ated with growth and, with somewhat less cer-tainty, also with poverty reduction. These includemacroeconomic stability (albeit with differencesabout when stability is achieved—for example,at what threshold inflation ceases to be a prob-lem); investment in health and education; andopenness to international trade.6 However, it isimportant to avoid overly simplistic causal inter-pretations of statistical associations. For exam-ple, while there is evidence that successfuleconomies (in terms of both growth and povertyreduction) tend to be more open, this outcome isnot necessarily assured simply by pursuing mar-ket-oriented trade policies; it may require a com-bination of circumstances and preconditions inwhich trade policies are only one and not neces-sarily even the most important factor.

• Good institutions and a favorable environmentfor private-sector-led development linked to aproper system of incentives are generally favor-able for growth.7

• Country-specific analyses and policy choicesare essential for two types of reasons. First, opti-mal growth/distribution strategies aiming atpoverty reduction in a given time frame woulddiffer depending on initial conditions (e.g., start-ing levels of inequality) and on the compositionof growth.8 Second, to understand the linkagesbetween the policies discussed above andgrowth/poverty reduction requires a good under-standing of the macro-micro linkages in a par-ticular country context, including the nature ofcountry-specific institutional characteristics thatdetermine the efficiency of markets.9

These considerations suggest that policy designposes complex problems. Homegrown policy debates—and more country-specific analyticalwork as an input to these debates—have the poten-tial for improving the understanding of policy con-straints and macro-micro linkages in particularcountry circumstances.

Scope and Methodology of theEvaluation

A few important considerations influenced thescope of the evaluation, all dictated by a need tokeep it reasonably focused and consistent with theIEO’s comparative advantage:

• The evaluation concentrates on those dimen-sions of the PRS initiative that are directly rele-vant to the IMF’s mandate, the role of the IMFin the process, and the PRGF. Clearly though,the ultimate success or failure of the new ap-proach depends on a much broader set of fac-tors, including contributions from many partnersin addition to the IMF.

• The evaluation focuses on the experience ofcountries with full PRSPs, based on the assump-tion that this experience has the greatest poten-tial for generating lessons of importance for theIMF’s role in the future.

• At this relatively early stage, the evaluation can-not cast much light on long-term outcomes forgrowth and poverty though some information isavailable on short-term growth outcomes. Themain focus of the evaluation is on inputs (thePRS process and the way in which PRGF-sup-ported programs are formulated), outputs (i.e.,PRSP contents and PRGF-supported programdesign) and intermediate effects (i.e., what insti-tutional and policy changes has the new ap-proach brought about and how well is the PRSlinked to domestic budgetary and other countrydecision-making processes). It must be empha-sized that existing policy papers provide littleguidance as to how these changes are to be mea-sured. Therefore, in carrying out our evaluation,we had to define indicators and benchmarksourselves. This inevitably involves an elementof judgment, but it is indispensable if the evalua-tion is to be objective.

• We do not address issues associated with the fi-nancing of the PRGF Trust Fund.10

18

6See, for instance, Berg and Krueger (2003); Cashin and others(2001); Srinivasan (2001); Khan and Senhadji (2000); Winters(2000); and Fischer (1993).

7Rodrik and Subramanian (2003); Easterly (2003); IMF(2003d); and Frankel (2002).

8See, for example, Bourguignon (2003), who notes that, if theobjective is poverty reduction, changing the distribution of in-come may be of relatively greater significance for middle-incomeinegalitarian countries than for low-income egalitarian countries.

9See, for example, Lucas (2002, p. 72): “Simply advising a so-ciety to ‘follow the Korean model’ is a little like advising an as-piring basketball player to ‘follow the Michael Jordan model.’ Tomake use of someone else’s successful performance at any task,one needs to be able to break this performance down into its com-ponent parts so that one can see what each part contributes to thewhole, which aspects of this performance are imitable of these,and which are worth imitating.” The importance of country-spe-cific circumstances probably accounts for the limited robustnessof many cross-section regression findings on the links betweenpolicies and growth.

10For a discussion of these issues and various options for thefuture structure of the PRGF instrument, see IMF (2004a).


19

Not

e: T

he fi

gure

rep

rese

nts

the

logi

cal f

ram

ewor

k of

bot

h th

e PR

SP a

ppro

ach

and

the

IEO

eva

luat

ion

of t

he P

RSP

and

PR

GF.

As

can

be s

een

from

the

figu

re, w

e do

not

pro

pose

to

exam

ine

final

out

com

es in

det

ail.

The

eva

luat

ion

prop

oses

to

exam

ine

the

effe

ctiv

enes

s of

the

app

roac

h by

look

ing

at t

he o

bser

vabl

e ef

fect

s of

the

new

pol

icy

fram

ewor

k on

a r

ange

of v

aria

bles

rel

evan

t to

the

IMF

man

date

, and

acc

ount

ing

for

thes

e ef

fect

s by

ass

essi

ng (

i) w

heth

er t

he in

itiat

ive

is im

plem

ente

d ef

ficie

ntly

(bo

th b

y co

untr

y an

d as

far

as IM

F in

puts

are

con

cern

ed)

and

(ii)

whe

ther

its

desi

gn w

as r

elev

ant

to a

ddre

ss in

itial

con

ditio

ns.

1 Int

erm

edia

te e

ffect

s ar

e ex

pect

ed a

cros

s th

e w

hole

ran

ge o

f con

ditio

ns fa

ced

initi

ally.

Her

e w

e fo

cus

on t

he a

reas

rel

evan

t to

the

IMF

man

date

.

Fig

ure

1.1.

Log

ical

Fra

mew

ork

of t

he P

RS

P A

ppro

ach

and

IEO

Eva

luat

ion

of P

RS

Ps

and

PR

GF

Inpu

ts a

nd c

ontr

ibut

ions

from

oth

er d

evel

opm

ent

part

ners

and

mis

cella

neou

s ex

ogen

ous

fact

ors

In-c

ount

ry

deve

lopm

ents

Init

ial c

ond

itio

ns•

Pov

erty

/mac

roec

onom

ic

env

elop

e•

Kno

wle

dge

base

• P

olic

ymak

ing

proc

ess

• In

stitu

tiona

l cap

acity

• E

cono

mic

pol

icy

fr

amew

ork

• R

elat

ions

hips

am

ong

d

evel

opm

ent

part

ners

PR

SP

pro

cess

Part

icip

ator

y fo

rmul

atio

nI-P

RSP

PRSP

PRSP

-PR

Part

icip

ator

y im

plem

enta

tion

and

mon

itori

ng

Inte

rmed

iate

eff

ects

1

• M

acro

econ

omic

per

form

ance

• K

now

ledg

e of

mac

ro-m

icro

li

nkag

es•

Pol

icym

akin

g pr

oces

s an

d

inst

itutio

nal c

apac

ity in

mac

ro-

e

cono

mic

and

rel

ated

are

as•

Eco

nom

ic p

olic

y fr

amew

ork

• D

eter

min

atio

n of

ext

erna

l

res

ourc

es e

nvel

ope

F

inal

out

com

es•

Acc

eler

ated

pov

erty

re

duct

ion

• F

aste

r su

stai

nabl

e

grow

th•

MD

Gs

IMF

role

H

IPC

/BW

Ico

nditi

onal

ityJS

APR

GF

Eval

uatio

n cr

iteri

a

IMF

way

of d

oing

bus

ines

s

Rep

ort

stru

ctur

e C

hapt

er#

2

Cha

pter

1an

d le

sson

sC

hapt

er#

3C

hapt

er#

4C

hapt

er#

5C

hapt

er#

6

Rel

evan

ceEf

ficie

ncy

Impa

ct a

nd e

ffect

iven

ess


One way of thinking about the scope of the evalu-ation would be in the context of a logical frameworkdiagram (see White (1999) for an example of this ap-proach). Figure 1.1 gives a broad indication of thedifferent stages.

The evaluation mainly uses the following types ofevidence:

• In-depth country case studies. The IEO under-took six country case studies, four of whichwere done jointly with the OED. An additionalfour case studies were undertaken by the OEDalone (Table 1.2). The case studies were chosento reflect diverse regional experiences, eco-nomic performance, and stages of the PRSPprocess.

• Each of the case studies involved field visits toconsult with a broad range of national stake-holders, interviews with Fund and Bank staff,and a detailed review of both internal and publicIMF documents.11 The four country case studiesundertaken jointly with the OED involved jointfield visits and the preparation of a single coun-try report. The coverage of topics varies be-tween case studies, with IEO-alone cases re-

fraining from discussing Bank-specific issues,and OED-alone case studies refraining fromevaluating IMF performance. However, themethodology applied was similar in all cases,which allows us to draw upon, where appropri-ate, evidence from all of the case studies.

• Cross-country analysis of countries with “full”PRSPs and of countries with PRGF-supportedprograms. Most of the analysis focused on the23 countries that had completed their PRSPs byend-2002.12 This analysis involved both quanti-tative and qualitative comparisons across coun-tries. The latter inevitably involved some subjec-tive judgments, but wherever possible wesystemized these judgments by disaggregatingthem and then rating each issue according to acommon scale, drawing upon standard codifica-tions of what we would expect to see to warranta particular rating. The purpose was to provideas much uniformity as possible to the overallcross-country judgments.

• A survey of a broad range of stakeholders ineach of the case study countries, using a com-mon survey format (see Annex 1).

20

Table 1.2. Country Case Studies for Evaluation1

Number ofHuman Year of Number of Years of

Development Full PRSP Year of PRSP Progress ImplementationHIPC Case? Index Ranking2 Completion “New” PRGF Reports3 of PRSP3

Joint IEO/OEDMozambique Yes 170 2001 * 1 2.3Nicaragua Yes 118 2001 2002 2 2.3Tajikistan No 112 2002 2002 0 1.1Tanzania Yes 151 2000 2000 2 3.1

IEO onlyGuinea Yes 159 2002 2001 0 1.4Vietnam No 109 2002 2001 1 1.6

OED onlyAlbania No 92 2001 2002 1 1.6Cambodia No 130 2002 * 0 0.9Ethiopia Yes 168 2002 2001 1 1.3Mauritania Yes 152 2000 2003 2 2.9

*New arrangement approved under ESAF regime in 1999.1The IEO only and OED only case studies do not attempt to assess the role of the other parent institution, but do cover a number of issues of joint relevance.We

will draw upon the “OED only” case studies in this report wherever they contain relevant information.2According to UNDP (2002).3As of end-2003.

11In addition, a joint IEO/OED workshop, hosted by the UNEconomic Commission for Africa, was held in Addis Ababa onJanuary 13–14, 2004 at which government and civil society rep-resentatives from the case study countries as well as several othercountries discussed the case studies and the key messages emerg-ing from the evaluation.

12The end-2002 cut-off was chosen in light of the terms of ref-erence of the IEO, which call upon it not to interfere with ongo-ing operations. In addition to the case study countries, this sampleincludes Bolivia, Burkina Faso, The Gambia, Guyana, Honduras,the Kyrgyz Republic, Malawi, Niger, Rwanda, Senegal, Uganda,Yemen, and Zambia.


• A survey of IMF mission chiefs and residentrepresentatives involved in the PRS and PRGFprocesses.

• Evaluations of aspects of the PRS process by other groups as well as evidence submitteddirectly to the IEO in connection with thisevaluation.13

The rest of this report is organized as follows.Chapter 2 discusses how the PRS principles arebeing applied in practice and presents an assessmentof the contents of PRSPs. Chapter 3 discusses therole of JSAs. Chapter 4 discusses the PRGF and anumber of program design issues. Chapter 5 assessesthe IMF’s way of doing business. Chapter 6 consid-ers whether the new approach is making a differencevis-à-vis the intermediate and, where possible, finaloutcomes. Background details are provided in anumber of annexes. The detailed country case stud-ies will be available as separate documents.

21

13Evidence that was prepared specifically as inputs to the eval-uation included Oxfam (2004) and CIDSE/Caritas (2004). Otherevaluations that we have drawn upon are listed in the references.

22

CHAPTER

2 Implementation of the PRSApproach in Areas Relevant tothe IMF Mandate

Key Messages

The design of the PRSP approach incorporates critical but largely unrecog-nized internal tensions: the PRS is to be country owned, but BWI policy as-sessments still obviously play a major role in financing decisions; it is to pre-sent strategic country aspirations, but also provide a sufficiently detailedoperational guide for policymaking and external support; and it is to be coun-try driven but meet various BWI-driven process requirements. An understand-able desire to preserve flexibility has resulted in ambiguity about the interme-diate objectives and hence the benchmarks against which to measureprogress.

The result has been that, in practice, the underlying incentives focus too muchon procedures and specific documentation, and that there are considerable di-vergences in expectations about what the initiative is meant to deliver.

PRS processes have generally been country driven to a much greater extentthan previous approaches. However, BWI procedural requirements are widelyfelt to have too much influence. In a minority of countries, PRS processes arebeginning to become sufficiently entrenched in domestic institutions to be self-sustained. In most cases, however, the PRS process is a long way from beingone that adapts to unanticipated developments and where monitoring of im-plementation feeds back into policy design.

Participatory processes of varying nature and scope have been launched in allcountries. Despite ambiguity about what such processes can or should deliver,in most cases stakeholders viewed them as an improvement over past prac-tices. They have typically succeeded in improving poverty diagnostics andalso, to an uneven extent, in improving transparency, public accountability,and influencing choices on public expenditure priorities. Their role in the areaof macroeconomic and related structural issues has been marginal to date.

PRSP contents reflect value added over previous strategies in terms of com-prehensiveness, results orientation and long-term perspective, although sig-nificant scope for progress remains even in “good practice” cases. The ana-lytical bases of PRSPs are typically weak, most notably with regard to themacroeconomic framework and policies to promote growth.

Insufficient prioritization, inadequate costing, and a tendency to avoid contro-versial structural reform issues mean that most PRSPs do not yet provide anadequate framework for making strategic decisions on key trade-offs. An addi-tional consequence is that PRSPs generally do not provide an adequate basisto guide the design of IMF-supported programs.

In some cases, progress is being made over time in ameliorating these variousshortcomings, mainly in countries that have begun embedding the PRSprocess in domestic institutions.

Chapter 2 • Implementation of the PRS Approach

In this chapter, we examine how the PRS ap-proach unfolded compared to original expectations,looking first at the PRS process and then at the con-tents of the PRSP document. Before turning to theseassessments, we discuss briefly some internal ten-sions underlying the approach. The impact of thePRS approach on the broader policy formulationprocess and on policies actually implemented is dis-cussed in Chapter 6.

Tensions Underlying the PRS Approach

There are three potential tensions in the design ofthe approach. Reluctance to address explicitly thevarious trade-offs they involve has led to vaguenessin some important aspects of the design. First, thereis a tension between the principle that PRSPs shouldbe country driven, and the externally imposed re-quirement for “broad-based participation,” whichmust be associated with an underlying—but unex-pressed—judgment that existing political processesin the recipient countries are inadequate in somesense. This raises questions as to how the participa-tory process is supposed to mesh with existing do-mestic political processes and when these processeswould be judged sufficient in and of themselves.Most observers would agree that, as a practical mat-ter, policy processes in these countries would benefitfrom opening up to broader groups, especially thoserepresenting the disadvantaged, but there are obviousdangers in pushing participatory processes that sup-plant or bypass existing institutional structures. Rec-ognizing this problem, the policy papers do not spec-ify what broad participation should imply but onlythat the process should be managed by the country.1

Second, there is a tension between the two func-tions assigned to PRSPs, namely to present the strate-gic aspirations of the country with respect to povertyreduction and growth and to provide a sufficientlydetailed operational strategy to enable the assistanceprograms of the BWIs to be anchored in the PRSP.This tension is reflected in debates over the nature of

the PRSP as a document and its audience. Is it meantto be a broad “strategic vision” paper emphasizingaspirations and strategy in general terms, with thegeneral public of the country as primary audience oris it meant to be a more detailed operational roadmap, with government policymakers, BWIs, and bi-lateral donors as the primary audience? The tensionbecame evident as subsequent preparation of variousguidelines—beginning with those for Joint Staff As-sessments (JSA) of PRSPs—elaborated on the expec-tations and of what PRSPs should contain. The resulthas been a quite complicated “architecture” of expec-tations/requirements (see Box 1.1 in Chapter 1) and,as we shall see, an excessive focus on BWI-orientedprocedural aspects in actual implementation.

Third, there is a tension between the emphasis oncountry ownership of policies and the role of BWIpolicy assessments. This tension shows up in thequestion of how far the “policy space” should bebroadened. While the initiative is meant to encouragethe exploration of alternative country-driven policyoptions, there is still the basic requirement that theIMF (and the World Bank) must satisfy themselves onthe soundness of country policies and the implicit as-sumption that they can help improve outcomes by in-fluencing policy choices through (i) policy advice,drawing on good analysis and cross-country experi-ence; and (ii) conditionality designed to monitor mu-tually agreed outcomes. A further potential aim is toencourage internal decision making in favor of re-forms judged desirable by the BWIs. The latter aspectsits less comfortably with some notions of ownership,but the apparent conflict can be reconciled if there is asufficient widening of the policy space to allow gen-uine ownership consistent with mutual agreement.

These unresolved tensions have led to considerableflexibility in implementation and variation acrosscountries, which is desirable especially in the begin-ning. However, they made it harder to derive concretebenchmarks on what changes in processes the ap-proach was trying to achieve. They also created greatdivergences in expectations on what some parts of theapproach are meant to deliver, thereby exposing theIMF to criticism, especially from I-NGOs, that it isonly paying lip service to some aspects of the initia-tive (e.g., ownership and participation).

The PRS Process

We analyze the PRS process on the basis of therelevant key principles, focusing first on the extent towhich it is country driven and then on the implemen-tation of the principle of broad-based participation.2

23

1Ambiguity about precisely what is expected from countries re-flects the lack of a mandate for the BWIs to impose a specific po-litical “test.” Regarding the IMF, this reluctance derives in partfrom the principle under international law that an international or-ganization only has the powers specifically conferred upon it inthe relevant treaty. To the extent that the IMF’s expressed pur-poses (i.e., in Article I) are of a purely economic and financial na-ture, this has been interpreted to imply that the Fund should notinterfere in political issues within member countries. In addition,the IMF’s Articles of Agreement make reference to respecting“the domestic social and political policies of members” in its sur-veillance of members’ exchange rate policies (Article IV, Section3(b)).

2Implementation of the partnership principle, insofar as it re-lates to the IMF’s role, is discussed in Chapter 5.

CHAPTER 2 • IMPLEMENTATION OF THE PRS APPROACH

To what extent are PRS processes “country driven”?

Original policy papers do not define the notion of“country driven” but the spirit of the papers suggeststhat the definition has to do with how much controlnational stakeholders have over the structure, sched-ule, and outputs of the PRS process. We look at threetypes of evidence: (i) how the process was orga-nized; (ii) stakeholders’ own perceptions; and (iii) towhat extent the process became self-sustained be-yond the formulation of the initial strategy docu-ment. In making this assessment, one needs to bearin mind the situation prevailing prior to PRSPs,when BWI concessional lending was supposed to bebased on Policy Framework Papers (PFPs), which bythe late 1990s had become widely criticized as notbeing country driven.

PRSP formulation process

In most of the country case studies the drafting ofthe PRSP was done by nationals of the country.3 Inthree cases (Albania, Cambodia, and Nicaragua), thenationals in charge of drafting the PRSP were donor-funded consultants, which might have implied somedonor control over the agenda. Indeed, in those threecases, the authorities complained about pressuresfrom their international partners. By contrast, inMozambique, the authorities refused any direct“support” from donors. In other cases, the authori-ties received some support from their developmentpartners and indeed sought their feedback at variousjunctures in the process, but reported no complaintsabout feeling pressured by the latter.

Initially, the timing of the PRS process generallydid not take into account domestic political cycles andplanning processes but was driven by the timetable ofthe HIPC Initiative or, in non-HIPC countries, by theprocedural requirements of the PRGF.4 In two cases(Albania and Ethiopia) the PRSP happened to come ata very opportune time—in the sense that the countrywas ready and willing to engage in the formulation ofa comprehensive poverty reduction and growth strat-egy—but this was fortuitous. In Vietnam, the PRSPfollowed the adoption of a new socioeconomic devel-opment plan, which provided the basis for several ele-ments of the PRSP, although their priorities were not

well aligned. In several other cases, the timing of thePRSP proved to be a source of difficulties. InNicaragua, there was a change of government shortlyafter the completion of the PRSP, bringing into officea team that did not endorse key aspects of the strategy.Lack of ownership by the incoming government criti-cally diminished the relevance of the initial process.However, there was eventually a broader debate onkey aspects of the development strategy (Box 2.1). Bycontrast, in Albania, many stakeholders interviewedreported that the PRSP provided an element of stabil-ity through a protracted political crisis and fourchanges of government. In Tanzania, the PRS processovertook the formulation of a homegrown external as-sistance strategy and led to it being temporarily setaside. Formulation of the latter strategy resumed later,although its purpose was modified somewhat to re-duce redundancy with the PRSP.

Several countries in our sample had a well institu-tionalized planning tradition at the onset of the PRSprocess but the latter generally unfolded through dif-ferent channels than these planning exercises, with-out integration of the outputs (Mozambique beingone exception). The relationship between these tra-ditional plans and the PRSP was unclear. In Cambo-dia and Vietnam, for instance, the PRSP was claimedto be the action plan operationalizing the NationalDevelopment Plan. In practice, however, the plansare the more important strategic documents in theeyes of the authorities.

The role of parliaments in approving the PRSPwas uneven: of the 29 PRSPs produced up to July2003, the final draft of the document was formallysent to parliaments in only 13 cases, but was not al-ways debated—let alone approved. In only 3 of thecases was the PRSP reviewed by a parliamentarycommittee.5

Stakeholders’ perceptions

In all country case studies, the authorities con-firmed that the key initial driver for embarking on aPRSP process was the fact that it was a necessarycondition for getting access to debt relief under theHIPC Initiative and/or to concessional lending fromthe IMF and the World Bank. However, in some ofthe cases (Albania, Ethiopia, Guinea, Mauritania,and Mozambique) the authorities rapidly came toperceive the exercise as an opportunity rather thanan imposition. Tanzania also adapted the approach toits own needs, even though it was initially viewed ascrowding out similar domestic initiatives.

24

3In most cases, the bulk of the drafting was done in English,with only final outputs translated in the local language. In Guineaand Vietnam, original drafts were in the official language. InTajikistan, the I-PRSP was translated into Russian and served as abasis for some of the subsequent consultations.

4Access to PRGF resources (whether approval of a newarrangement or completion of a review) is conditional upon acountry’s issuance of an I-PRSP, PRSP preparation status report,PRSP, or PRSP Progress Report in the previous 12 months.

5Among the 10 country case studies, the final PRSP was for-mally approved by parliament only in Mauritania and Tajikistan.A more detailed account of the role of parliaments in the PRSPprocess is given in Eberlei and Henn (2003).


The IEO/OED survey reveals interesting differ-ences across groups on the extent to which the PRSPwas driven by national stakeholders—from a clearlypositive assessment by government stakeholders, toa negative one by I-NGOs and a neutral one bydonors and local civil society. In 3 countries out of10 (Mozambique, Nicaragua, and Tanzania), the av-erage assessment across stakeholder groups wasnegative, whereas it was positive on average in theother cases.

Implementation and sustainability of the PRS process

A key test of the extent to which a PRS process iscountry driven is how it evolves beyond the formula-tion of the first full PRSP, in particular, whether thePRSP is being implemented and monitored, andwhether it becomes as originally intended, a “livingdocument,” periodically modified in the light of ex-perience. A more demanding test is whether theprocess is sustained once there are no financial in-centives directly attached to it. We examine each ofthese issues in turn.

As far as implementation is concerned, the casestudies indicate that there was frequently a drop in

momentum immediately after the finalization of thePRSP, but implementation of some aspects pro-ceeded relatively soon in most cases.6 However, tothe extent that a PRSP incorporated preexistingdonor-funded program and reform strategies, imple-menting these aspects does not represent a particu-larly noteworthy development. Moreover, the verycomprehensiveness of PRSPs and frequent lack ofspecificity and prioritization of their policy orienta-tions also made it relatively easy to relate any newpolicy undertaking of the government to the PRS(see Box 2.2 on Guinea for illustration).

More generally, any assessment of implementa-tion depends critically on the quality of monitoringand reporting systems. Reflecting the initiative’s em-phasis on results orientation, most PRSPs envisagedambitious monitoring systems, based on a largenumber of indicators. Actual institutional arrange-ments for monitoring PRS implementation havebeen set up in many of the case studies, sometimesbuilding on existing structures (Albania, Ethiopia,and Mozambique), in other cases involving newly

25

Box 2.1. Change of Government and Modification of PRSP Strategy:The Case of Nicaragua1

The Nicaragua case provides an interesting exampleof how “country ownership” interacted with a changein government. In the fall of 2002, the new Bolaños ad-ministration sought to modify the strategic priorities in-troduced by the previous government with the launchof a draft new development strategy. This initiativesought to change public expenditure composition inline with the new government’s vision on how to fostergrowth and employment creation. In essence, the de-bate revolved around the “growth pillar” of the PRSPand the new government’s greater emphasis on invest-ment in areas that they viewed as having greater growthpotential, but which were not always where the poorwere found.

The government’s initiative, however, was markedby tensions in the process. The launching of the newstrategy coincided with the agreement on a new PRGF-supported program and a new World Bank Country As-sistance Strategy (CAS)—both of which were linked tothe strategy in the earlier PRSP. The government feltconstrained in its ability to pursue a revised strategy,while donors had expected the PRSP to bring a more

lasting policy framework, and were concerned that thenew strategy had not been subject to sufficient partici-patory debate. Interestingly, the intention to modify thegrowth pillar was not a significant topic of debate dur-ing the electoral campaign, which raises obvious ques-tions about the degree of integration of the PRS processwith existing political mechanisms.

As a result of the mixed reception, the governmentdecided to strengthen the original draft and broaden theconsultation process. A planned Consultative Groupmeeting was postponed, and a consultation process ona shorter, revised document was set up. The authoritiespresented a revised growth strategy document duringthe Consultative Group meeting in October 2003.Moreover, the government indicated its intention tohold a workshop in early 2004 with stakeholders to dis-cuss how to modify the PRSP based on the new strat-egy. Although still ongoing and too early to assess theoutcome, recent developments point to changes in theright direction, including wider consultation and thebroadening of policy dialogue.

This experience illustrates some of the potential ten-sions between different principles of the PRSP frame-work, notably between ownership—and the provisionof greater policy space to the government—and otherobjectives such as partnership and participation.

1The background case study on Nicaragua provides a morecomprehensive discussion.

6This is true even in cases where implementation was initiallydisrupted by political developments (e.g., Albania and Nicaragua)or by interruptions in aid flows (e.g., Guinea).


dedicated units (Cambodia, Nicaragua, Tajikistan,and Tanzania). In most countries, these institutionalarrangements have yet to produce usable data, re-flecting in part acute capacity constraints and diffi-culties in prioritizing indicators.7

The main evidence on which to base an assessmentof PRS implementation are the PRSP Progress Re-ports (PRSP-PRs) that countries are required to pre-pare on an annual basis in order to have continued ac-cess to PRGF resources. The guidelines for JSAs ofPRSP-PRs note that “with a view to minimizing theadministrative burden on countries, it is expected thatthe reporting on progress could be integrated withinregular government processes and be presented asshort summary documents. The exact timing of theannual report is flexible, and could also change overtime in order to become consistent with national

reporting and decision making processes.” However,PRSP-PRs issued to date suggest that, in practice,countries have not made use of the flexibility pro-vided, nor in most cases have they integrated thepreparation of PRSP-PRs with their own budgetprocess. For instance, in Mozambique, the authoritieswish to align substantive updates and revisions of thePRSP to the domestic political cycle, and they find itunfortunate that the production of a PRSP-PR followsa cycle driven by external procedural requirements.

As of February 2004, 12 countries had issued atleast one PRSP-PR.8 Many PRSP-PRs report goodprogress in the implementation of relevant structuralreforms, often in the areas of public expendituremanagement, decentralization and privatization, and

26

Box 2.2. A Case of Observing PRS Principles Without Achieving Its Objectives:The Guinean PRSP

The Guinean PRSP, prepared by senior civil servantsfollowing a broad-based participatory process, wasgenerally well regarded by the BWIs as a document, inthe sense that it makes a comprehensive diagnosis ofpoverty, and presents a well-articulated strategy thattakes a long-term perspective, is reasonably results ori-ented, and reflects some prioritization efforts.1

As a process, however, the PRS did not outlive theformulation of the PRSP itself. Dissemination of thePRSP within the country has been very restricted, sothat there is little awareness—let alone ownership—ofits contents. The participatory infrastructure set up forthe formulation of the PRSP has been largely dormantsince its completion, with the exception of two the-matic groups, one focused on culture and communica-tion, the other on evaluation and monitoring. The lat-ter, however, only has a consultative role, and twoyears after the completion of the PRSP there are nowell-defined transparent arrangements for the moni-toring of PRS implementation. Mechanisms to updatethe PRSP are also lacking, and—with some excep-tions in the year following the completion of thePRSP—the policy formulation process is driven byother considerations.

The PRSP had little impact as a strategic guide (es-pecially in the macroeconomic and related areas), forthe following reasons: (i) The macroeconomic frame-work outlined in the PRSP was superseded by events(including exogenous shocks, deviations from policycommitments, and withdrawal of external assistance)even before the PRSP was finalized. It was not subse-quently updated to reflect changes in the external envi-ronment; nor did it incorporate contingency scenariosor broad orientations indicating how trade-offs shouldbe resolved in the event of a changed environment. (ii)The PRSP was insufficiently prioritized, in the sensethat the priority actions envisaged exceed both the ca-pacity of the country to deliver and the resources avail-able. In some areas, there is also a disconnect betweenthe priorities outlined in the PRSP (e.g., governance)and the government’s actual agenda. (iii) The PRSP ini-tially had some links with the medium-term expendi-ture framework (MTEF) and, through it, the budget, butthe budget actually implemented reflected only looselyPRSP priorities. (iv) The PRSP failed to discuss in anyoperational way structural reforms considered key foran improved macroeconomic performance, such asstate-owned enterprise reform.

For reasons largely unrelated to the contents of thePRSP, the donor community has remained unwilling toprovide financial support commensurate with thePRSP’s needs and assumptions. The PRGF-supportedprogram approved in 2001 went off track a few monthsafter the PRSP was reviewed by the Boards of theBWIs, and the PRSP has not proved a useful guide todiscussions between IMF staff and the authorities onways to bring it back on track.

1The participatory process was broad in the sense that it in-cluded direct popular consultation at the grassroots level andinvolved representatives of civil society and donors in the-matic working groups and occasional workshops. However, itdid not lead to a meaningful policy debate in most areas, espe-cially with respect to the macroeconomic framework and re-lated structural reforms.

7Issues related to monitoring of poverty and related indicatorsare discussed in greater depth in the parallel OED evaluation.

8Out of the 23 countries with over a year of experience under aPRSP. The 12 countries are Albania, Burkina Faso, Ethiopia,Honduras, Malawi, Mauritania, Mozambique, Nicaragua, Niger,Tanzania, Uganda, and Vietnam. Five of these countries have is-sued a second PRSP-PR and two have issued a third.


in the setting up of working institutional arrange-ments to monitor PRS implementation. Only onecountry (Nicaragua) reported progress against an-nual quantitative targets.

The extent to which PRSPs are “living docu-ments” varies considerably. At one extreme, the in-stitutional arrangements that produced the PRSPhave largely dried up (e.g., Guinea). At the other endof the spectrum, Tanzania has set up a process thatallowed it to fill gradually the major gaps in its ini-tial PRSP while also adapting it to changes in themacroeconomic environment. Mozambique has alsomade progress in institutionalizing the process. Inbetween, most of the countries that have preparedPRSP-PRs have used it as an opportunity to updatethe initial strategy, often adding substantively to theoriginal paper with regard to diagnostics (e.g., as theresults of large-scale surveys became available) orwith respect to specific sectoral strategies (e.g., inMauritania, Nicaragua, and Vietnam).

To summarize, the extent to which PRS processeshave been country driven varies substantially. In allcases, the primary driver was initially access to debtrelief and concessional assistance, but in a number ofcases country ownership appears to have strengthenedover time. The extent to which the process has takenroot, and brought about changes, also varies acrosssectors, being typically slimmest with respect to the

formulation of macroeconomic policies. Among thefour country case studies that had reached their com-pletion point under the HIPC Initiative at the time ofwriting, two (Mozambique and Tanzania) seem tohave established processes now sufficiently en-trenched that they are likely to be self-sustained.

Excessive documentation requirements

Another important message on PRS processesemerging from our evaluation—and one that wasemphasized by both country authorities and IMFstaff—is the excessive paperwork burden (see Box2.3). Indeed, paperwork seems to have increased,rather than diminished, over time.

How effective has “broad-based participation”been in PRS processes?

What does “broad based participation” mean?

As noted above, the original policy papers weredeliberately not very precise on what is meant bybroad-based participation and there are a number ofdifferent concepts of participation in the literature.McGee and Norton (2000), for example, propose aladder of degrees of participation consisting of (i) in-formation dissemination; (ii) consultation (involving

27

Box 2.3.Too Much Paperwork?

The focus of the PRSP initiative on document production imposes a heavy burden onboth country authorities and IMF staff. As the list below shows, both are asked to pro-duce a series of papers with partially overlapping purposes and issues coverage. For theauthorities, these paperwork requirements coexist with those of “homegrown” processes(e.g., budget documents, medium-term expenditure frameworks, and developmentplans).

Documents relevant to the PRS/PRGF initiative that are periodically produced by:

Country authorities IMF staff

I-PRSP PRGF staff reportPRSP Article IV staff report2

PRSP Progress Report Joint Staff Assessment (JSA) of PRSP3

Preparation Status Report1 JSA of PRSP Progress Report3

LOI/MEFP Comfort/assessment letter1

HIPC-related reports (three types)2, 3

Country Strategy Paper1

Ex post assessment1

Progress in Implementation Report3, 4

Global Monitoring Report3, 4

1Document to be produced only under specific circumstances, that is, not systematically required.2Periodicity and other rules regulating the production of these documents are independent of the PRSP/PRGF initia-

tive, but they are nonetheless highly relevant and have a high potential for overlap.3Produced jointly with the World Bank.4These reports are initiative-wide, but build on country-specific data. Their main authors are not the country teams

responsible for the production of other reports listed.


a two-way flow of information); (iii) collaborativedecision making; and (iv) initiation and control bystakeholders.9 The degree of participation neededwithin this hierarchy depends upon the objectivessought. The matrix in Table 2.1 identifies differentlevels of objectives and indicates the level of partici-pation relevant for each objective while also indicat-ing whether the objective is explicitly mentioned inIMF policy documents. A review of IMF policy pa-pers suggests that only objectives listed under A andB in the matrix are expected intermediate outcomesof the PRS/PRGF initiative. “Empowerment” re-ceives more emphasis in external commentaries, butis not included among the explicit objectives of theinitiative. Consistent with the concern not to imposea political test on countries, the JSA guidelines callfor the staff to describe, not assess, the participatoryprocess—with any assessment left to the ExecutiveBoards, although the criteria on which such an as-sessment would be made are nowhere defined ex-plicitly. In practice, the IMF Board rarely makessuch an assessment. In the absence of specific stan-dards, governments typically understood participa-tion as information sharing or, at best consultation,with some variations across topics (see Annex 5 fordetails on the case study countries).10

What has participation achieved inmacroeconomic and related areas?

Clearly, the appropriate or even feasible degreeof participation will vary according to country cir-cumstances (in particular, their “starting point” interms of openness of the policy debate and literacyof the population), but also depending on the topicat stake. Thus, the yardstick of success is at least asmuch progress from the starting point as distancefrom the ultimate objective. With this in mind, weassess the impact of participation on policy issueswhere the IMF has primary responsibility amongthe BWIs.

Did participatory processes contribute to a bet-ter design of poverty reduction strategies? Mostcase studies suggest that the participatory processwas instrumental in reaching a good diagnosis ofpoverty and in choosing related indicators to be in-cluded in the PRSP, by bringing out previously un-deremphasized dimensions such as vulnerabilityand other nonincome aspects of poverty. Beyondthis admittedly important element, the impact ofthe participatory process on final strategies wasgenerally not large, at least in areas of primary con-cern to the IMF.11 It contributed to determiningbroad priorities in the composition of public expen-ditures in a number of cases and influenced policiesin a few other areas (e.g., user fees for primary

28

9Other authors, while retaining similar categories, prefer defin-ing participation as a continuum of approaches (e.g., Robb andScott, 2001). The PRSP Sourcebook describes participation as“the process through which stakeholders influence and share con-trol over priority setting, policy-making, resource allocations andaccess to public goods and services.”

10The idea of specific “standards” for participation appears tohave been rejected during discussions in the early stages of the ini-tiative, for two reasons. First, as in other areas, it was felt that im-posing norms across countries would run against the initiative’semphasis on country-driven processes; second, that such cross-country guidelines or standards would be meaningless because of

Table 2.1. Possible Categories of Participation and Their Objectives

Is Objective Explicit (Minimum) Degree of ParticipationPossible Objectives of Participation in IMF Policy Documents? Required to Achieve Objective

A. Better design of poverty reduction/growth strategies YesA1. From improved diagnostics Information exchangeA2. From richer policy debate (considering broader Consultation

range of alternatives)

B. Improved policy implementation YesB1. From enhanced accountability Information exchangeB2. From enhanced ownership /broader consensus Collaborative decision making

C. Empowering disadvantaged groups NoC1. By fostering a sense of inclusion Information exchangeC2. By strengthening the voice and influence of the Collaborative decision making

poor in the domestic balance of power

the wide differences found between country contexts and politicalsystems (a notion that is supported in the literature on participa-tion). However, it is hard to see how at least clarifying what was ex-pected from the participatory process (along the lines of the “lad-der” discussed above) would have imposed more limits on theaccommodation of country specificities than the current approach.

11This is reinforced by the broader sample. Only about one-fifth of the 23-country sample of PRSPs note any impact of theparticipatory process on policies.


schools) in several countries. In three of the casestudies (Guinea, Mozambique, and Nicaragua),stakeholders noted that the prominence of gover-nance in the strategy was a result of the participa-tory process. Beyond these specific examples, thereis little evidence of a substantive impact of partici-patory processes on the macroeconomic and relatedstructural policy choices embedded in PRSPs.12 Ina number of cases, some important but controver-sial structural policy issues did not surface in thebroader debate around the PRSP (see the section“Contents of PRSPs in Macro-Relevant Areas.”).Consistent with these findings, our survey of PRSPstakeholders revealed that stakeholder groups out-side of government disagree that alternatives werefully explored, and also generally disagree that thefinal document was modified to accommodatesome of their viewpoints.

What was the impact of PRS participatoryprocesses on accountability and ownership? In mostof the cases studied, stakeholders felt that the partic-ipatory process leading to PRSP formulation repre-sented a significant improvement over past practices,in particular by enhancing the transparency of gov-ernment policies and access to government represen-tatives. It also fostered a sense of inclusion that wasnew for many stakeholders. This effect was oftenshort-lived, reflecting the ad hoc nature of theprocess, but it raised expectations that could, insome cases, be a driving force behind more sus-tained improvements in accountability going for-ward. Indeed, a number of civil society stakeholderswe interviewed stressed that, viewed in a dynamiccontext, the process had opened up new windows toinfluence policy formulation and to provide feed-back on implementation, which they viewed as amajor value added even if the results had been moreevident in some areas (e.g., some sectoral policies)than others. The case studies also suggest that theprocess created strong ownership of the strategy atleast within the relatively narrow circle of those di-rectly in charge of managing the PRSP formulationprocess. Ownership at the broader government levelsubsequently strengthened in many cases as imple-mentation proceeded, at least in “priority” ministries(e.g., Albania, Ethiopia, Guinea, and Tanzania), butit remained elusive in other parts of government (es-pecially at the local levels) and in large parts of civilsociety.13

What were the main obstacles to more effective participation?

In most cases, the debate was not framed in a waythat was conducive to effective participation in thesense that participants were generally not madeaware of key data and analytical underpinnings andof the trade-offs at stake, nor were they explicitlyasked to prioritize their preferences. Moreover, theformat of consultations limited participants’ abilityto provide substantive inputs, in particular: lack ofaccess to information before the meetings and of follow-up afterward; limited lead time to react todrafts submitted (especially in Cambodia andGuinea); implicit constraints to open expression ofviews by stakeholders (Cambodia, Guinea, Maurita-nia, and Vietnam); and, often, the advanced stage atwhich inputs were sought, implying little or noscope for discussion of alternative options and trade-offs.14 Furthermore, in most cases, some key stake-holder groups (e.g., parliamentarians, the poor, andthe business sector) did not take part in a meaningfulway (see Annex 5 for details).

Participatory processes typically did not addressall building blocks of the PRS in the same depth,with policy areas of primary concern to the IMF re-ceiving the least attention. Poverty diagnostics wereusually extensively covered, but macroeconomicpolicy and structural reforms were typically side-stepped or not discussed in any depth, with the im-portant exception of the composition of public ex-penditures (Box 2.4). Still less participation andpolicy debate appears to have occurred in countriesfacing immediate macroeconomic stabilization chal-lenges. Indeed, there was generally not even “infor-mation sharing” on some sensitive policy issues(e.g., bank restructuring and its fiscal consequencesin Mozambique and Nicaragua). Many of these is-sues would be difficult to address substantively in“town-hall” type meetings, but except in a few casesthere was little experimentation with other ap-proaches to broadening the debate on macroeco-nomic issues (e.g., through involvement of local re-search groups, and more systematic discussion inrelevant parliamentary committees). Countrieswhere steps had been taken to establish more institu-tionalized processes did better in this respect (al-though as the Tanzania case suggests, even herethere were limits to civil society involvement. Wewill return to this in Box 4.2). Not surprisingly,

29

12In several cases, a further broadening of the debate and of thepolicy space did occur over time, but the outcome was not neces-sarily reflected in the PRSP itself. This is discussed further inChapter 6.

13This latter finding broadly echoes the conclusion reached,among others, by Booth (2003).

14Similar problems are reported in external reviews of thePRSP experience in Bolivia, Cameroon, Ethiopia, Honduras,Mozambique, and Nepal (see, for instance, Stewart and Wang,2003). While in most cases the pressure imposed by the scheduleof the HIPC Initiative was reportedly an important factor, it can-not be an explanation for other cases.


therefore, most nongovernmental stakeholders sur-veyed disagreed that the macroeconomic frameworkof the PRS was derived in a participatory manner.

The participatory process related to the formula-tion of the PRSP was conducted as an ad hoc exer-cise in most cases, not as one that would need to besustained overtime. As a result, an initial drop-off inthe level of participation generally occurred follow-ing completion of the PRSP. Efforts to disseminatethe strategy to the general public once adopted havebeen limited.15 In some cases (e.g., Albania, Mauri-tania, Mozambique, and Tanzania), this was fol-

lowed by a gradual pickup at the monitoring and im-plementation stage, often as governments startedpreparing a PRSP-PR. In those cases, the develop-ment of participation was a dynamic one—graduallydeepening over time as governments learned the ad-vantages of such processes and various civil societygroups began to identify various “entry points”—inparticular the budget process—where they couldhave an influence on the decision-making process.In most other cases, however, the tapering off of par-ticipation has continued.16

In sum, the objectives requiring only informationexchange (such as improved diagnostics, enhancedaccountability, and greater sense of inclusion) were

30

Box 2.4.Was Macroeconomic Policy Discussed in the Participatory Process? Insights from Case Studies1

Tanzania. Initial consultation involved some discus-sion of expenditure priorities but not of the broadermacroeconomic framework or trade-offs involved. Thesituation changed dramatically afterward, with keymacroeconomic issues now being debated in openforum in the Public Expenditure Review (PER) work-ing groups. NGO involvement in those debates remainslimited, however, at least in the macroeconomic area.

Mauritania. Limited involvement of NGOs in the ac-tivities of the working groups dealing with macroeco-nomic issues, due to lack of interest and capacity.

Mozambique. Macroeconomic policy not on theagenda of public consultations, but some issues sur-faced nonetheless from NGO inputs, such as trade pro-tection, industrial subsidies, and access to credit.

Guinea. One of six multi-stakeholder thematic groupsset up to do the groundwork of PRSP formulation wasdedicated to macroeconomic policy and growth sectors.Its participants reported having regular and active dis-cussions throughout the PRSP formulation process.However, by their own account, these discussions werenot very substantive. In particular, the overall macro-economic framework in which the PRSP would be em-bedded was not discussed, nor were policy alternatives

or trade-offs. The main output of these meetings was tospell out broad policy objectives such as increasing do-mestic revenue mobilization, keeping inflation low, andenhancing the availability of micro credit.

Vietnam. Macroeconomic policy and some relatedstructural reforms were not put on the agenda of con-sultations by the organizers of the process (notablythe World Bank and I-NGOs, with the approval of the authorities) partly because they were notviewed as priority issues given prevailing macroeco-nomic stability.

Tajikistan. One of nine multi-stakeholder workinggroups set up to do the groundwork of PRSP formula-tion was dedicated to “macroeconomic management”and there was some discussion of macroeconomic is-sues in the consultation process. In particular, thepoverty assessment report prepared by the NGO um-brella group does contain a number of recommenda-tions related to macroeconomic policy, focused ontransparency in public finances, tax policy, exchangerate management, and the fight against inflation. Theseinputs suggest there is scope for broadening participa-tion in macroeconomic policy discussions.

Ethiopia. Discussions of the macroeconomic frame-work were generally viewed by stakeholders as insuffi-ciently participatory, although they broadly acknowl-edged that the framework represented a step forwardfrom earlier approaches.

Cambodia. The macroeconomic policy sections of thePRSP were not subject to any significant consultationwith civil society representatives.

1Countries are listed in chronological order of completionof their PRSP (more recent last). Here, we define macroeco-nomic policy as including key medium-term macroeconomicassumptions, fiscal policy, the objectives of monetary policy,exchange rate, and related structural reforms (including re-form of state-owned enterprise (SOE), financial sector re-form, and trade policy).

15As of July 2001, of the 36 PRSPs and I-PRSPs documentsproduced by non-Anglophone countries, only 25 existed in an of-ficial language version and out of these 25, 17 were not publishedby the member country or by the IMF. Among the case studies,Mauritania and Tanzania are noteworthy exceptions. Resourceconstraints appear to have been an important factor in at leastsome cases.

16Across the 10 country case studies, civil society stakeholderssurveyed generally disagreed that the government had continuedto engage them in the PRS process beyond formulation. Therewas no apparent correlation between these perceptions and thematurity of the process.


often met to some extent. However, progress with re-spect to the objectives requiring a deeper form ofparticipation—that is, richer policy debate and en-hanced country ownership, strengthening the voiceof the poor, and so on—was much less and was alsouneven across countries and policy areas, with theleast impact on issues where the IMF is more di-rectly concerned. Some degree of institutionalizationof the process seems essential to ensure continuedstakeholder involvement and capacity building overtime.17 However, this also raises difficult questionson how to integrate the ad hoc PRS process with tra-ditional democratic institutions, which, no matterhow flawed initially, should take over the responsi-bility for the participatory process after a transitionphase. In this connection, the very limited involve-ment of parliaments in PRS processes to date re-mains a cause for concern.

Contents of PRSPs in Macro-Relevant Areas

We now turn to an assessment of the contents ofthe PRSPs in the areas most relevant to the IMF man-date based on a desk review of the 23 PRSPs and ad-ditional insights from the 10 IEO/OED case studiesand external evidence.18 We assess PRSP contentsagainst four sets of criteria: first, compliance with theunderlying PRS principles (see Box 1.2); second,whether the contents are so uniform that they implyconformity to BWI-driven policy approaches; third,the expectations and suggested characteristics estab-lished by the JSA guidelines; and finally, suitability asa basis for an IMF-supported program.

Compliance with underlying PRS principles

Comprehensiveness

The principle of comprehensiveness applies bothto the definition of poverty, and the policy agenda en-visaged to tackle it. Our evaluation suggests that theprinciple has been implemented with greater successin the former area. The survey of stakeholders in casestudy countries indicates broad agreement thatPRSPs make an adequate diagnosis of both the natureand causes of poverty. The case studies also revealedthat stakeholders typically felt that the PRS approachhad added value in this respect. JSAs also considerthis area to be one of PRSP’s main strengths in overtwo-thirds of cases.

Our assessment is much less positive with regard tothe comprehensiveness of the policy agenda embed-ded in PRSPs. While virtually all PRSPs (in the 23-country sample) discuss growth, human capital devel-opment/social services delivery, and governance,these areas are generally not covered in the samedepth. In most countries, significant efforts had al-ready gone into defining strategies in the areas ofhealth and education prior to the PRSP, resulting inmore fully fleshed out strategies in these areas. Therewas typically much less emphasis on the broader ob-stacles to growth or on strategies in sectors such asagriculture or infrastructure. In some cases (e.g.,Mozambique, Nicaragua, and Tanzania), strategies inthe latter areas were completed post-PRSP. Strategiesfor private sector development were frequently under-developed. Fox (2003) notes that even though virtu-ally all PRSPs assign a key role to the private sector asan engine of growth, less than a fifth outline concretecommitments in that regard (among the case studies,Vietnam is a noteworthy exception). Similarly, tradeliberalization is an issue that, although of significancefor growth and poverty, was discussed in some detailonly in a minority of PRSPs.

Another problem revealed by the case studies isthat the various components of the strategy are oftenpoorly integrated, in the sense that the linkages andsynergies between them are not fully discussed andthat they do not necessarily fit in the macroeconomicframework of the PRSP. For example, in the minor-ity of cases where it is spelled out, the total of expen-ditures required to implement all the parts of thestrategy often exceed by far the available resourceenvelope (the Cambodia PRSP offers one of thestarkest illustrations). Furthermore, in most cases,comprehensiveness was achieved at the expense ofrealism and operational relevance owing to a failureto prioritize the policy agenda (see next section).

Results orientation

Virtually all PRSPs discuss monitoring indicatorsand/or arrangements. Nevertheless, this area is identi-fied by JSAs as a key weakness in 60 percent ofPRSPs. A frequent problem revealed by our case stud-ies is that the indicators and monitoring arrangementscontemplated in PRSPs far exceed the administrativecapacity to collect and analyze the underlying data(e.g., in Cambodia, Ethiopia, Guinea, and Vietnam)—a process that is very resource intensive and, realisti-cally, will take many years to institutionalize in mostcountries. Furthermore, indicators are typically betterdefined for health and education strategies than inother areas, where inputs or process-based indicatorstend to dominate, or indicators are simply lacking(e.g., in the area of structural reforms or governance).In some cases, there is no straightforward relation be-

31

17See Eberlei (2002) for a discussion of the defining character-istics of a fully institutionalized participatory process.

18Our focus here is on the PRSPs as documents. The impact ofthese documents on policymaking is discussed in Chapter 6.


tween the indicators monitored and the targets of thePRSP.

The most important obstacle to a greater resultsorientation of PRSPs is the still tenuous nature oflinks to the budget and medium-term expenditureframeworks (MTEF). Over half of the full sample ofPRSPs lacked such links initially; among our casestudies, only 4 countries out of 10 foresaw links intheir PRSP, at least for expenditures in designated“priority sectors” (Albania, Ethiopia, Guinea, andMozambique). In other cases (e.g., Tanzania), suchlinks were successfully developed ex post. Wherelinks between the PRSP and the MTEF existed onpaper from the start, their effectiveness was weak-ened by the fact that the MTEF was itself onlyloosely connected with the budget actually imple-mented (e.g., in Guinea).19 In Vietnam, efforts to es-tablish a MTEF are only just beginning. Theseweaknesses reflect a combination of limited costingand prioritization in PRSPs and the poor state ofpublic expenditure management generally—an issuewe return to in Chapter 4.20

Despite these difficulties, interviews conducted aspart of the case studies suggest that bringing somedegree of results orientation to poverty reduction/de-velopment strategies has been one of the main valueadded of the PRSP approach.

Long-term perspective21

Our survey of PRSP stakeholders suggests a gen-eral agreement (in 8 countries out of 10) that PRSPsprovide an adequate road map to long-term goals, al-though I-NGOs and donors tend to be more skepticalthan domestic civil society and government stake-holders. However, our evaluation suggests a morenegative assessment. The linkages between medium-term, operational targets and longer-term ones aregenerally not well specified. In particular, mostPRSPs do not provide an effective guide to how thestrategy will need to be modified if the macroeco-nomic framework is disrupted, a decision that de-pends on how critical trade-offs are handled. We re-turn to this issue later.

Are PRSP contents so uniform as to suggestimplicit BWI-driven policy standards?

While the JSA guidelines and the PRSP Source-book are explicitly not prescriptive, outside critics

have argued that there are implicit standards thatcountries feel obliged to abide by to ensure that theBWIs endorse their PRSPs.22 We tested this hypoth-esis by reviewing the contents of the 23 PRSPs com-pleted by end-2002. A high degree of uniformitycould suggest the existence of implicit standardsconstraining the extent to which PRSPs can truly becountry driven.

The desk review of 23 PRSPs does suggest somedegree of uniformity, but not to such an extent as toprove the existence of implicit standards.23

• All PRSPs contemplate measures to increasedomestic revenue mobilization, although themeans envisaged vary by country: some men-tion side-by-side general increases in revenue,improvements in tax administration, broadeningthe tax base, and reforming tax legislation,while others choose to emphasize one or severalof these avenues.

• A little over half the sample also considers mea-sures on the expenditure side: in two cases, ex-penditure cuts; in other cases rationalization ofexpenditures or improvements in public expen-diture management, alone or in combination.

• As regards key structural measures defined at ahigh level of generality, common themes doemerge: almost two-thirds of PRSPs envisagesome form of privatization, and a little under halfgive some emphasis to financial sector reform(liberalization) and decentralization. The vastmajority of PRSPs also put some emphasis onimproving public sector governance. The cover-age of all these areas in PRSPs may well reflectencouragements to that effect found in the JSAguidelines and the PRSP Sourcebook. Likewise,coverage of so-called cross-cutting issues, suchas gender, environmental sustainability, andHIV-AIDS, was gradually mainstreamed aftertheir absence in the first generation of PRSPs at-tracted repeated criticism from the Boards of theBWIs and was also mentioned in several of thestaff reports on Progress in Implementation. But

32

19The approach taken in Uganda—to publish these documentsas backing for the annual budget—seems to be the best one.

20Bevan (2004) provides a more detailed elaboration of theseissues.

21Other aspects of this issue, including linkages between thePRSPs and the MDGs, are addressed in the OED evaluation.

22Strictly speaking, the IMF and the World Bank do not “ap-prove” the PRSP. However, for countries to receive concessionallending (and, where relevant, debt relief) from them, the Boardsof both institutions have to determine that the PRSP constitutes acredible poverty reduction strategy and a sound basis for theirlending operations. In practice, most external stakeholders eitherare unaware of this subtle distinction, or regard it as irrelevant ormisleading, in the sense that the positive signal from the BWIs iscritical for external assistance in support of PRS implementationto materialize.

23The desk review involved recording in a matrix (not reportedhere) the following two elements for each PRSP: (i) key fiscalmeasures; and (ii) key structural policy measures in macro-rele-vant areas.


the policy choices made under these headings arecountry specific.

• PRSPs vary with respect to some fundamentalpolicy choices. For instance, while a majorityenvisages a reduction of the role of governmentin the economy, about a third do not. In the areaof trade policy, about a third of PRSPs envisagepolicies that directly contradict the policy ad-vice given in the PRSP Sourcebook.24

There is also a widespread perception that theprominence given to governance in many PRSPsowed much to donor pressures. However, the casestudies indicate that this high profile also accuratelyreflected the views expressed by many stakeholdersduring the PRS process.

While this evidence on content does not suggestan externally driven blueprint, perhaps of greaterconcern is what is not in the PRSPs. The case studiessuggest that a number of issues that might be politi-cally sensitive or where the authorities and the BWIsmay have diverging views are given only a cursorytreatment. Trade policy and state enterprise reformswere common examples; another area skirted overwas issues associated with bank restructuring (e.g.,in Nicaragua), although in this case there are practi-cal reasons—avoidance of bank runs—why broaderpublic debate would have been difficult. In anyevent, the fact that these issues were not dealt withadequately in the PRSP often led to “downstream”problems when the same issues arose in the contextof PRGF negotiations. (We will trace through theimpact of the policy debate on specific policy out-comes in Chapter 6.)

Conformity to expectations set out in JSA guidelines

Analytical basis

According to the JSA guidelines, PRSPs are ex-pected to contain four key analytical elements: an as-sessment of the impact of past policies on growthand poverty reduction; an analysis of the key con-straints to growth and poverty reduction; an analysisof the sources of growth; and poverty and social im-pact analysis of the key policies contemplated.

Most PRSPs are rather weak in their coverage ofthese aspects. No PRSP in the 23-country samplethat we evaluated has all four analyses, and thelargest group (43 percent of the cases) provides onlyone of the four elements. Two PRSPs (Burkina Faso

and the Kyrgyz Republic) that were accepted as pro-viding a sound basis for concessional lending hadnone of these analytical elements. Of the four com-ponents, PSIA (loosely defined) is the least fre-quently found (in only 2 cases out of 23). Assess-ments of past policies and analyses of the sources ofgrowth were the most frequent—undertaken in abouthalf of all cases—but in a somewhat cursory way inmost cases. Analyses of the key constraints togrowth and poverty reduction are present to someextent in only 30 percent of PRSPs, with typicallymore emphasis on obstacles to poverty reductionthan to growth.

It is important to recognize that these types ofanalyses are very demanding in terms of data avail-ability and technical capacity and it is perhaps notsurprising that the analytical basis is an area ofweakness in most initial PRSPs. However, thisweakness may limit the validity and relevance of thestrategy adopted. Encouragingly, PRSP Progress Re-ports suggest that in several cases (e.g., Ethiopia,Nicaragua, and Tanzania) work is under way to rem-edy these knowledge gaps, but most PRS processeshave not generated prioritized action plans on whatthe BWIs themselves should be doing to help fillthese gaps.

Prioritization, costing, and financing

JSAs identify costing, financing, and prioritiza-tion as one of the main weaknesses of PRSPs in two-thirds of the 23 cases. Costing of action plans is fre-quently attempted, but is usually partial andrudimentary.25 Over a third of PRSPs lack a financ-ing plan and/or a prioritized action plan.

The case studies reinforce these messages. Whileall PRSPs contain a policy matrix outlining elementsof an action plan in various areas, these are fre-quently little more than wish lists resulting from theaggregation—without prioritization or costing—ofthe preferences expressed by various stakeholdergroups in the participatory process (e.g., in Albaniaand Cambodia). Clearly, with practically no knowl-edge of the respective poverty impact of variousmeasures and only imprecise estimates of their costs,it is difficult to achieve a meaningful prioritization ofthe action plan. In some cases, efforts have beenmade to assign degrees of priority to the various ob-jectives of the PRSP, but the large number of “high-

33

24See Ladd (2003). This review also concluded that mostPRSPs skirted many of the fundamental strategic choices involv-ing trade policy.

25Bevan (2004) distinguishes three levels of PRSP “costing”—corresponding to rising standards: (i) wish lists of aspirations ac-companied by tentative expenditure estimates where possible; (ii) aggregative and extrapolative expenditure estimates; and (iii) activity-based budgeting, based on unit costs. He concludes bystressing the very preliminary nature of what has been achieved inPRSPs against these standards, even in the strongest cases.


priority” objectives limits the practical value of thiseffort (e.g., in Guinea), and the actions considered toachieve each of these objectives are not, themselves,prioritized. Often, prioritization and costing aremore advanced in “priority sectors” (typicallyhealth, education, and infrastructure) than in otherareas. PRSP Progress Reports (for the whole sampleof PRSPs) suggest that little progress is made in im-proving the prioritization and costing of PRSP ac-tion plans in the years following the completion ofthe full PRSP—Mozambique being a “good prac-tice” exception.

Diagnosis of implementation constraints

Most PRSPs fall short of expectations in dis-cussing risks and implementation constraints. Lessthan one-fifth of cases provide such a discussion, eventhough our case studies indicate that administrativecapacity constraints are severe and constitute a seriousrisk to implementation. Other risks, such as disrup-tions in aid flows or exogenous shocks, are also rarelydiscussed in detail, even though most countries expe-rience them with some frequency. This assessment isechoed by our survey of PRSP stakeholders: in 6countries out of 10, a majority of stakeholders sur-veyed do not agree that the targets and plans outlinedin the PRSP are realistic, with I-NGOs and donorsbeing the most skeptical groups.

Suitability as basis for IMF-supportedprograms

As noted at the beginning of this chapter, the PRSinitiative has established relatively few concretebenchmarks, in terms of intermediate objectives forchanges in processes, against which progress of theinitiative could be systematically assessed. SincePRGF-supported programs are supposed to be em-bedded in the PRSPs, one potential benchmark ofparticular importance to the IMF suggests itself: doPRSPs provide an effective strategic guide for poli-cymaking in macroeconomic and related areas thatwould provide a suitable basis for IMF-supportedprograms? Existing IMF policy documents do notdefine explicitly which, among all of the expected orsuggested components of a PRSP, are critical forthem to constitute a sound basis for a PRGF-sup-ported program. Our case studies suggest that thefollowing characteristics are essential if PRSPs areto play this role:

• A macroeconomic framework (i.e., growth pro-jections, revenue and expenditure projections,and balance of payments) that is realistic, inter-nally consistent, and incorporates reasonablecost estimates for identified priority actions. It

should also identify how policies will be ad-justed in response to unanticipated shocks;26

• A clearly articulated growth strategy, mindful ofthe trade-offs likely to arise and of their implica-tions for poverty reduction.

• Prioritized public policy actions in macro-rele-vant areas, in particular related to the majorstructural policies that have been an issue in thepast.

Very few PRSPs had satisfactory macroeconomicframeworks. In only 4 out of the 10 case studies (Al-bania, Vietnam, Mauritania, and Tanzania) didPRSPs present a realistic macroeconomic frame-work. In the case of the latter two countries, this wasobtained by explicitly adopting in the PRSP thesame framework as in the pre-existing PRGF-sup-ported program. In the 6 other cases, the originalmacroeconomic framework was either supersededby events by the time the PRSP was discussed by theBoard (Tajikistan, Guinea, Mozambique, andNicaragua) or unrealistic (Ethiopia) or incomplete(Cambodia). In these cases, there was an under-standing that annual PRSP Progress Reports wouldprovide an opportunity for revisions of the macro-economic framework.

In none of the 10 country case studies were alter-native scenarios explored in the PRSP, and only inCambodia did the PRSP consider, to a limited ex-tent, the downside risks to the framework and theirmacroeconomic implications.27 In most cases therewere no contingency plans identified or even broadorientations outlined to indicate how the frameworkmight be adjusted in response to adverse or favor-able shocks.28 Interestingly, these weaknesses werenoted candidly in the respective JSAs, but did notprevent the staffs from reaching the unqualified con-clusion that the PRSPs constituted sound bases forIMF and World Bank concessional lending. (We re-turn to this issue in Chapter 3.)

34

26The idea is not for a PRS to offer detailed contingency plansattempting to address every possible type of shock that mightarise, as experience shows that this is unrealistic and ineffective.Rather, the PRSP should set key strategic orientations that wouldprovide guidance in solving the trade-offs that arise when policyadjustments are needed.

27Among the 23 countries that completed a PRSP by end-2002,only three (Niger, Rwanda, and Senegal) presented fully fleshedout alternative scenarios. Madagascar is another more recent example.

28The medium-term macroeconomic framework of the TajikistanPRSP, for instance, was judged by staff to be conservative andtherefore the JSA recommended to the authorities that they shouldflesh out a higher case scenario. In Guinea and Vietnam, the PRSPoutlined broadly what additional expenditures would be undertakenin the event of higher resource availability. In the case of Guinea,the baseline scenario itself was optimistic, so the value added of aneven more optimistic alternative scenario was limited.


Most PRSPs also fell short of the requirement toarticulate a comprehensive growth strategy and struc-tural reform priorities. Either the policy agenda insome or all of the relevant areas lacked detail (Cam-bodia, Guinea, Nicaragua, and Tanzania), or it wasinsufficiently prioritized (e.g., Ethiopia and Mozam-bique), or both (Albania and Tajikistan). Only inVietnam was the policy matrix found to be suitablydetailed and prioritized, although even in this casesome key issues (e.g., the pace and content of stateenterprise reform and related program conditionality)resurfaced during program negotiations, reflectingdisagreements between the authorities and IMF staffon the relative importance of key objectives.

Thus, none of the PRSPs for the 10 case studieswould have met fully our suggested benchmark ofproviding a sound basis for design of a PRGF-sup-

ported program. See Box 2.5 for some possible rea-sons why. The survey of staff involved inPRSP/PRGF countries confirms these findings.Asked to assess various factors hindering the abilityto align PRGF-supported programs on PRSPs, re-spondents almost universally responded that policiesdiscussed in PRSPs are too imprecise to be an opera-tional guide. Many also considered the lack of real-ism of macroeconomic frameworks as a key impedi-ment to such alignment.

There is, however, some evidence of progressover time. A majority of PRSP-PRs indicateprogress in the definition of the macroeconomicframework (especially in Albania, Mauritania, andNicaragua). Progress in articulating an operationalgrowth strategy and structural reform plans hasbeen less frequent. Among the case study countries,

35

Box 2.5.Why Is It Proving So Difficult for the PRS Approach to Provide an Effective Operational Road Map on Macroeconomic Policy?

The evaluation suggests several reasons why the PRSapproach has generally not been successful in using thenew way of doing business implied by the PRS princi-ples to generate a strategic guide to policymaking onmacroeconomic issues.

1. Problems of transition. Many countries alreadyhad IMF-supported programs before their PRSP was fi-nalized, and in these cases the macroeconomic strategyand framework was generally taken from the programand incorporated into the PRSP. Over time, morePRSPs are preceding the PRGF-supported programs,which would be the normal sequence (see Chapter 5).

2. Capacity weaknesses, especially in costing, publicexpenditure management, and budgetary processes.These weaknesses can be remedied over time, andsome progress has been made but expectations need tobe based on realistic assessments of each country’s ini-tial capacity in these areas. Major improvements willalmost certainly take significant time.

3. Insufficient change in the IMF’s way of doingbusiness. The IMF’s internal procedures have adapted,but there continues to be substantial differences be-tween “average” and “best” practice. Moreover, theleast change—in terms of exploring ways of openingup the policy debate to consider alternative options—has occurred in those circumstances where the IMF’srole is likely to be most critical (e.g., in “difficult”cases where immediate macro-stabilization concernsare at the forefront). This suggests that additional ef-forts may be needed to apply the PRS principles insuch cases, even when the PRSP itself is imperfect. Wediscuss these issues further in Chapter 5.

4. Reluctance to address some controversial butmacroeconomically critical issues in the PRS process.Part of the solution to this and the preceding issue is tohave clearer “rules of the game” on the nature of thepolicy debate.

5. The nature of some macroeconomic policy issuesmakes it difficult to generate an effective, lasting opera-tional guide. Most low-income countries face espe-cially volatile environments that complicate longer-term macroeconomic planning.1 Consequently, detailedquantitative macroeconomic frameworks typically havea short shelf life. This means that discussions on howmacroeconomic policies should respond to such shockswill inevitably take place at a much higher frequencythan can be managed by the broader-based processesmeant to drive the overall PRS. To be an effective oper-ational guide, the broader strategy would need to incor-porate guidance—in effect, contingency plans—forhow to handle trade-offs in such circumstances. Butthere are deep political economy reasons why govern-ments are reluctant in practice to spell out such contin-gency plans in advance.2 The frequent result is that de-cisions on such matters are left within the traditionalnarrowly based framework of program negotiations.

6. The operational role of the macroeconomic frame-work in the PRS is ambiguous. In particular, it is un-clear how the tension between “needs based” and “real-ism” aspects are to be resolved (see Chapter 5).

1For example, according to IMF (2004a) between 1992 and2001, the average size of a negative commodity price shock(i.e., the decline in real price from the preceding year) was 22percent for low-income countries and 16 percent for other de-veloping countries. Such shocks also occurred with greater fre-quency in low-income countries. Low-income countries alsosuffer from more frequent and more costly natural disasters.

2This is illustrated by the fact that many discussions withinthe IMF on different aspects of program design call for a bet-ter specification of contingency plans as a means of dealingwith uncertainty, but that such exhortation is rarely heeded inpractice.


only Tanzania and, to some extent, Mozambiqueseem to have made substantive progress in that direction.

To conclude, the weaknesses identified in the firstround of PRSPs were probably unavoidable giventhe learning process involved in their formulation.Moreover, although most still fall well short of ex-pectation, there is a wide recognition that they aregenerally a significant improvement over previousBWI-fostered approaches (notably the PFP). Thegreatest changes in both processes and content so farhave been with regard to various sectoral policies

(including health and education) and public expendi-tures. Changes in processes have been less in areasof macroeconomic policy formulation, and this is reflected in the contents of most PRSPs. “Goodpractice” cases—especially those where the PRSprocesses have begun to be embedded in domesticinstitutional arrangements—indicate continuedprogress, but in most cases the approach has not yetgenerated the sort of country-driven strategic frame-works in these policy areas that would help guideshorter-term decision making or provide an effectivebasis for the design of PRGF-supported programs.

36

An important part of the “architecture” of thePRS process is the JSA, by the staffs of both the IMFand the World Bank, which evaluates the strengthsand weaknesses of a country’s PRSP and considerswhether the PRSP provides a sound basis for con-cessional assistance from the BWIs and for HIPCdebt relief. JSAs are submitted to the ExecutiveBoards of the two institutions along with countries’PRSPs and are always made public. In principle,they can serve three purposes:

• An internal governance function: providing theBoards with a basis on which to decide whetherthe country’s strategy deserves concessionalsupport, which may include debt relief.

• A feedback and outreach function: informingthe government and other domestic stakeholdersof staff views on the strengths and weaknessesof the PRS.

• A partnership function: signaling to the donorcommunity whether the BWIs view a particularPRS as worthy of support, what areas needstrengthening, and so on.

The jointness of the assessment primarily re-flected the perception that it was not sensible to dis-sociate the assessment of the macroeconomic com-ponents of the PRS from that of the developmentalcomponents, as they would succeed or fail together.

Joint Staff Assessments

CHAPTER

3

37

Key Messages

On balance, Joint Staff Assessments (JSAs) do not perform adequately themany tasks expected of them.

JSAs generally cover most of the ground suggested by the guidelines, but theclarity, candor, and comprehensiveness of the assessment are uneven, withscope for improvement even in “good practice” cases. This partly reflectsshortcomings in the guidelines themselves and a built-in bias to reach a posi-tive conclusion given the “yes-no” character of the underlying judgment onsuitability as a basis for concessional financing.

Their value added to the IMF as an internal governance tool is unclear, sincedecisions on access to IMF concessional financing, on the nature and scope ofnonfinancial assistance, and on program design are made in separate contextswithout being guided by the JSA.

JSAs of PRSP Progress Reports add little value to the Progress Reports themselves.

The JSA does provide useful feedback to countries about their PRS, but lack of awareness of the JSA among nongovernment stakeholders reduces its effectiveness.

JSAs have not contributed significantly to fostering coordination among de-velopment partners, reflecting in part shortcomings in coverage of issues, andin part the limited relevance of JSAs in the eyes of many donors, who wouldlike a greater input into the assessments.

CHAPTER 3 • JOINT STAFF ASSESSMENTS

It was expected to improve Bank-Fund collaborationby ensuring that their staff reached a common per-spective on the country.

Interestingly, when asked what they thought wasthe primary value added of the JSAs, IMF staff sur-veyed emphasized not the internal governance func-tion, but the provision of feedback to the authoritiesand, to a lesser extent, of an independent expert as-sessment of the PRSP to interested parties.1 More-over, the increase in the amount of paperwork ex-pected from IMF staff (Box 2.3) suggests that theJSA did not make any of the preexisting staff assess-ments redundant. In the remainder of this section,we discuss the contents of JSAs, assess how effec-tive they have been in meeting their different pur-poses, and review possible factors underlying thisperformance.

Contents of JSAs

The guidelines for JSAs of PRSPs indicate thatthe assessments should succinctly answer key ques-tions about each of the four core elements of thePRSP (i.e., a description of the participatory process;poverty diagnosis; targets, indicators, and monitor-ing systems; and priority public actions), givinggreatest weight to the priority public actions and tothe arrangements for monitoring and evaluation ofPRSP implementation, and paying due regard to thecountry’s starting point. We assess how the JSAshandle the key questions asked by the guidelines inthe areas relevant to the IMF’s mandate.2

Participatory process

While in general the guidelines call for JSAs to as-sess rather than describe or repeat the contents ofPRSPs, in the case of participatory processes theguidelines note that “the Executive Boards have in-structed the staffs to describe, but not evaluate, theparticipatory process.” In practice, however, 22 JSAsout of 23 do assess the participatory process, and 16of those identify the participatory process as one ofthe main strengths of the PRSP, while four mention itas one of the weaknesses.3 In about half of the cases,

these assessments are not backed by an analysis ofthe participatory process and its impact on owner-ship. Several of the key questions asked by the guide-lines in relation to the participatory process are fre-quently not addressed. Furthermore, based on ourcase studies, our overall judgment is that the JSAstend to present participation in a more favorable lightthan warranted by circumstances (except in Cambo-dia, where the assessment is quite candid and in Tan-zania, where the JSA refrained from assessing theparticipatory process). In particular, JSAs often over-state the extent and significance of civil society par-ticipation (e.g., in Mozambique, Nicaragua, andTajikistan) and the degree to which the participatoryexercises were country driven (e.g., Vietnam), whileunderstating (and sometimes entirely overlooking)constraints to effective participation (e.g., in Guinea).A key reason why JSAs tend to present an insuffi-ciently critical assessment of the participatoryprocess is that (in keeping with the guidelines) theyemphasize who was consulted and how without giv-ing much information on what was discussed orwhether the discussions were meaningful.

Targets and monitoring

On average, JSAs provide a good description ofPRSP targets and monitoring arrangements, alongwith some assessment of their adequacy. However,there are wide variations: the realism of targets andmonitoring arrangements and the suitability of theindicators retained are generally assessed. By con-trast, the transparency and participatory nature ofmonitoring arrangements are discussed in just overhalf of the cases, and the feedback of monitoringinto policymaking is discussed only in a minority ofcases. This pattern is common to both early and re-cent JSAs.

Priority action plans

JSAs’ treatment of priority actions plans was ex-amined focusing on three areas that are particularlyrelevant to the IMF’s mandate: the macroeconomicframework; key fiscal choices; and financing plans.In all three areas, JSAs offer a broadly satisfactoryassessment, although with room for improvement.

• Almost all JSAs discuss the soundness and real-ism of macroeconomic frameworks. However,only half discuss their robustness to shocks, andonly a minority of JSAs assess whether keytrade-offs are discussed in the PRSP.

38

1Reflecting the ambivalent nature of the JSA, a relatively highpercentage of respondents chose the “don’t know” option whenanswering questions about its value added.

2This analysis is based on a desk review of JSAs issued throughJune 2003, with complements drawn from the case studies. Thedesk review involved using a standard assessment matrix to as-sess each JSA vis-à-vis the criteria outlined in the JSA guidelines.The coding scheme underlying this assessment and average rat-ings under each criterion are presented in Annex 6.

3We considered that JSAs assessed the participatory processesrather than just described them when a qualitative value judgment

was expressed and/or when the participatory process was listedamong either the main strengths or the main weaknesses of thePRSP.

Chapter 3 • Joint Staff Assessments

• With respect to key fiscal choices, the three di-mensions highlighted by the guidelines—inter-nal consistency; quality of costing estimates;and administrative capacity to deliver (includingquality of PEM system)—are tackled in mostcases.

• The assessment of financing plans is generallythe weakest. A majority of JSAs discuss the re-alism of financing plans, but their sustainabilityis rarely addressed, nor is the existence of con-tingency plans.

The desk review results suggest that some learn-ing over time has taken place with respect to the ap-praisal of macroeconomic frameworks, but not in theother areas.

Other issues

Surprisingly, JSAs rarely discuss “gaps” in PRSPs,in the sense of critical policy issues that are not ad-dressed. The analysis of key implementation risks isalso limited. All JSAs identify at least a couple ofsuch risks but most do not suggest remedial actions.Furthermore, the case studies (e.g., Guinea, Maurita-nia, and Mozambique) suggest that JSAs are some-times insufficiently candid in their assessment of theseriousness of the risks to implementation, especiallywith respect to the implications of weak administra-tive capacity and of inadequate links between the PRSand normal government processes.

JSA guidelines require the staff to “take into con-sideration the country’s starting point.” Ideally, JSAsshould give a sense of what initial conditions were inkey areas and, based on that, of what could realisti-cally be expected from the PRSP process, so that thePRSP would be assessed against explicit country-specific benchmarks. In practice, JSAs have gener-ally not adopted such an approach. In about one-third of cases, there is a perfunctory discussion ofstarting conditions. In others, starting conditions arediscussed adequately, but the implications for thePRSP are not clearly drawn out.

JSAs of PRSP Progress Reports

According to the guidelines for JSAs of PRSP-PRs, these reports “must make an overall assessmentfor the Executive Boards as to whether progress inimplementation has been satisfactory, and whetheror not the strategy presented in a PRSP remains asound basis for concessional assistance from theFund and the Bank.” We assessed the JSAs of 12PRSP-PRs issued through the first quarter of 2004.

Surprisingly, the key questions suggested by theguidelines for staff to consider in drawing their as-sessment are narrowly focused on the contents of

Progress Reports, rather than on actual progress inimplementation. Reflecting the guidelines, JSAsfocus on the PRSP-PR documents, and in manycases the latter are appraised exclusively on the basisof information reported by the authorities in the doc-ument itself.4 A significant proportion of JSAs limitthemselves to summarizing the Progress Report andassessing its strengths and weaknesses without at-tempting to remedy gaps in coverage (e.g., in assess-ing progress against the targets set in the PRSP it-self), by bringing in information from otheravailable sources, or making an independent assess-ment of progress. Only a small minority attempt toappraise progress in implementation of the strategyagainst the objectives of the initiative, such asbroader ownership, better design of policies and pri-ority setting, and the extent to which the PRSP is as-suming a central role in the domestic policymakingprocess. Among the case studies, the JSAs for Tan-zania and Ethiopia are examples of “good practices”in that regard.

Value Added of the JSA

Performance of the JSA as an internalgovernance tool

The usefulness of the JSA as an internal gover-nance tool is limited by the fact that JSAs are con-strained to reach a binary (yes/no) conclusion onwhether the strategies presented in PRSPs constitutea sound basis for concessional lending by the BWIs,with the use of standardized language.5 All JSAs todate have provided a “yes” signal and this conclu-sion has systematically been endorsed by theBoards—even when the PRSP did not in fact pro-vide a satisfactory framework in which to anchor aPRGF-supported program. In these cases, the JSA

39

4Such exclusive reliance on the document produced by the au-thorities to assess progress in implementation is bound to giverise to candor issues (see, for instance, the somewhat rosy treat-ment of progress under the governance pillar in the JSA ofMozambique’s PRSP-PR).

5The use of a standardized conclusion was decided in the earlydays of the PRSP initiative, at a time when it was tightly linked tothe implementation of the HIPC Initiative. In that context, the useof the standard sentence (“The staffs of the World Bank and theIMF consider that this PRSP presents a credible poverty reductionstrategy and provides a sound basis for IDA and Fund concessionalassistance”) was meant to avoid language that could have given theimpression that the BWIs cleared or endorsed countries’ PRSPs, asthis would have run counter to the new emphasis on country owner-ship. While the use of standardized language made sense in thecontext of the HIPC Initiative, where the amount and schedule ofassistance are entirely determined on the basis of objective criteria,the same cannot be said of other BWI lending decisions. Neverthe-less, the use of a standardized conclusion was confirmed in theguidelines for JSAs of PRSP Progress Reports.

CHAPTER 3 • JOINT STAFF ASSESSMENTS

typically noted serious weaknesses in the PRSP,along with encouragements to amend the PRS in anumber of ways, but nevertheless reached a positiveconclusion. Furthermore, what constitutes a soundbasis for BWI concessional lending (or, more criti-cally, what fails to constitute a sound basis) is notdefined anywhere.

JSAs have not contributed in any substantive wayto the determination of the amounts of BWI assis-tance and detailed design of the programs.6 AllJSAs identify a (varied) number of strengths andweaknesses in PRSPs, but not in a way that wouldallow the reader to get a sense of the overallstrength of the PRSP, that is, by delivering a gradu-ated assessment on the basis of explicit bench-marks. For example, our review of JSAs shows thatthey listed more strengths than weaknesses for only12 of the 23 PRSPs appraised up to end-2002. Six-teen PRSPs were considered “sound bases” eventhough they had among their main weaknesses atleast two of the following—seemingly critical—di-mensions: (i) costing, financing, and prioritization;(ii) public expenditure management; and (iii) imple-mentation and monitoring. Five PRSPs were con-sidered deficient in all three respects and yet theJSAs concluded that they provided a suitable basisfor concessional lending.

JSAs typically do not discuss what the BWIs planto do, both to support the country’s strategy and tohelp remedy any weaknesses (nor are they requiredto do so explicitly by the guidelines). In particular,they do not underline the implications of perceivedweaknesses for the extent to which their own opera-tions may be aligned on the PRSP, nor the priorityactions that the BWIs themselves should take. TheCAS is intended to fulfill this purpose in the WorldBank. However, there is no clear equivalent in theIMF. As a result, the path from the JSA to a PRGFarrangement is often not clear, and many key issuesare left to be resolved in the context of PRGF-relatednegotiations.7

With regard to IMF–World Bank collaboration,the main value added of the JSA reported by IMFstaff was to foster the expression of a comprehen-sive, unified World Bank perspective in its areas ofcompetence. Nevertheless, there is also a viewamong IMF staff that the need to reach a compro-mise with Bank staff on issues of shared interest isan important contributing factor to the lack of candorand clarity of the assessments conveyed in the JSA.Some additional findings on the JSA emerging fromthe OED review are summarized in Box 3.1.

As a result of these various limitations, the mainvalue added of the JSA from the perspective of inter-nal governance seems to be of an informational na-ture, that is, by summarizing for Board members themain features of the PRSPs prepared by membercountries. Even in this role, however, JSAs faceshortcomings because of their limited scope to focuson the dimensions of keenest interest to each BWI.For instance, the commentary on participatoryprocesses found in JSAs typically does not discusswhether macroeconomic policies and related issuesat the core of IMF concerns were addressed, let

40

6The guidelines for JSAs of PRSPs note that these elements aredetermined through the World Bank’s CAS and arrangementsunder the PRGF, but that “the JSA contributes to these determina-tions through its assessment of the strengths and weaknesses ofthe strategy.” At the World Bank, the determination of lendinglevels continues to rely primarily on Country Policy and Institu-tional Assessment (CPIA) ratings (which do not incorporate in-formation on status of the PRS process), while at the IMF accessunder PRGF arrangements takes into account, in a nontransparentmanner, a variety of criteria (such as balance of payments need,strength of adjustment program, and track record with the IMF)that do not include PRS-related factors.

7This is in contrast with the process adopted in selected countries(including Mozambique and Tanzania) by donors who provide gen-eral budget support on the basis of a mutually agreed performanceassessment framework (PAF) that builds heavily on the PRSP itselfand forms the basis for annual Joint Partner Reviews (JPRs). For

Box 3.1. Key Findings of the Review ofJSAs by the World Bank’s OED

As background to its evaluation of the PRSPprocess, the OED conducted a review of JSAs, as-sessing how well each JSA identified the strengthsand weaknesses of a PRSP and signaled directionsfor improvements, using a four-point scale where“3” (out of “4”) indicates satisfactory treatment.This assessment was made for 11 issues, 7 of whichare not listed as such in the JSA guidelines but weredeemed relevant based on the case studies and onthe record of discussions of the World Bank’s Exec-utive Board. The average overall rating given byOED to the 28 JSAs reviewed was 2.7, with signs oftentative improvement over time in some respects.The topics most satisfactorily handled in a majorityof JSAs were endogenous and exogenous risks,structural and sectoral policies, and poverty diag-noses. The least well covered topics were capacityconstraints, private sector participation in PRSP for-mulation, and donor partnerships issues.

The OED review also concurs with the IEO as-sessment of limited usefulness of the JSA to localstakeholders and the overly optimistic or incompletetreatment of the value and contribution of the partic-ipatory process.

the IMF, there is no such straightforward mapping between PRGFconditionality, the PRSP policy matrix, and the JSA.

Annex 4

91

Table A4.1 (concluded)

Average Rating of ESAF and PRGF BenchmarkBriefing Papers and Highest Scoring Briefs1

_________________________________________ESAF PRGF____________________ ____________________

Percentage Percentage Comments on Changes in Policy Approach Average of briefs Average of briefs as Indicated by Comparisons of ESAF- and rating of scoring 3 rating of scoring 3

Issues for Assessment PRGF-Related Benchmark Briefing Papers brief (1–4) and above brief (1–4) and above

7. Does the brief identify policy issues The importance of PSIAs is highlighted by n.a. n.a. 2.6 57where poverty and social impact many briefs. However, inputs during analysis (PSIA) inputs would be implementation were assumed to come necessary/useful? (Explain precisely from the government and other agencies,what was proposed/done on PSIA in while methodological and capacity comments section.) constraints were not discussed.

B. Review departments’ comments

8. Is the need for more “policy space” for The review process did not put much 1.2 4 1.9 22homegrown options recognized in the emphasis on creating policy space for review process? (The focus of this homegrown options—by, for example,question is not on the magnitude of encouraging a more participatory approach adjustment but on how prescriptive to policy formulation. This was then review comments were.) mirrored in the briefs’ relatively inflexible

approach to policy formulation, as indicated above.

9. Are poverty issues discussed or their Review departments insisted on raising the 1.8 26 2.5 52absence highlighted in review depart- profile for poverty and social expenditure ments’ comments? issues, including PSIA, in the majority of

briefs, explaining to a large extent the increased emphasis on poverty and related issues seen in the PRGF briefs.

10. Did review departments press for Compared to the ESAF briefs, PRGF briefs 1.4 13 2.4 48more or less conditionality (prior were less inclined to press for more actions, performance criteria, conditionality. However, this is mostly true structural benchmarks) in the policy for countries that had established credibility areas specified by the mission? with the IMF (via a successful ESAF program,

for example). For “early stabilizers” conditionality did not decline by much.

1The ranking scheme is based on the degree of consistency with the PRSP/PRGF approach: 1 = Highly inconsistent; 2 = Inconsistent; 3 = Consistent; 4 = Highly con-sistent (see Table A4.2 for the criteria used in the ranking).

2Including governance (financial transparency and anticorruption), trade liberalization, and debt sustainability.3Including privatization and SOE reform, public sector reform, private sector development, capacity building, agricultural sector and land reform, forestry sector

policy, industrial sector reform, and other reforms.

ANNEX 4

92

Table A4.2. Coding Scheme for the Cross-Country Task on Briefing Papers and Department Reviews1

Numbers in Parentheses Refer to Assessment Reached

Section 1. Briefing paper

A. Issues and timing(i) Key policy issues to be addressed by the Examples of policy issues (but add specifics in “comments” section):

mission (specify in each case, with a focus on Fiscal policymajor issues). Trade liberalization

PrivatizationCivil service reformPublic expenditure reformMonetary policyFinancial sector reformGovernance

(ii) Was PRGF launched before full PRSP? Yes/No; provide date of full PRSP.

B. Contents of briefs(i) Does the brief discuss the links between the (1) No linkage (mere mention of poverty not enough).

PRSP and the PRGF objectives? (2) Discussion of poverty issues and references to PRSP but no indication of how thetwo are linked.

(3) Some discussions of PRSP strategy and objectives and of links with PRGF proposals,but not comprehensive.

(4) Comprehensive discussion of key objectives and strategy of PRSP and of how pro-posed PRGF program will be integrated with them.

(ii) In resolving the issues identified in A above (1) Brief leaves no policy space (i.e., sets objectives and specifies policies).does the brief leave room for discussion, (2) Brief leaves room for discussion of a narrow range of policy alternatives.based on domestic policy debate, on the (3) Brief leaves room for discussion of a broad range of policy alternatives.specific policies to be adopted to achieve the (4) Policy options not restricted in discussions of objectives and analyses of key trade-offs.objectives of the poverty reduction strategy? (i.e., are a range of possible policy options considered in key areas?) (specify for each policy issue).

(iii) Are alternative macroeconomic frameworks (1) No discussion of alternative frameworks.and the trade-offs between them considered? (2) Limited discussion of an alternative macroeconomic framework (but no discussion of

potential trade-offs).(3) Alternative macroeconomic framework and trade-offs discussed but implications for

PRSP objectives not analyzed in depth.(4) Alternative macroeconomic frameworks, potential trade-offs between them, and

their implication for overall PRSP objectives clearly analyzed.

(iv) Does the brief discuss how the program will (1) No discussion of potential shocks or of how program would adapt to them.protect key objectives in the event of un- (2) Risk of shocks is discussed but no discussion of trade-offs between adjustment and anticipated negative shocks? financing or how to protect key objectives.

(3) Potential shocks are identified and mix of financing and adjustment discussed, but nosignificant discussion of implication for key objectives and how to protect them.

(4) Potential shocks are identified; trade-offs between adjustment and financing clearlyanalyzed; and potential strategy for preserving key objectives is set out.

(v) Does the brief allow for flexibility to use (1) Brief proposes a specific fiscal deficit target, with no flexibility for use of additional additional concessional external financing, if concessional financing.available? (2) Some limited flexibility to accommodate additional inflows.

(3) Proposed program allows for significant flexibility in accommodating additional fi-nancing, but no systematic discussion of implications for key PRSP objectives.

(4) Brief discusses implications of different levels of external financing for achievement ofPRSP objectives, and lays out an explicit strategy on how the program will accommo-date additional inflows.

(vi) Is a participatory process in resolving key (1) Not discussed.policy issues considered? (2) Consultations with donors, NGOs etc., by Fund staff are suggested but no indication

of how the results will be incorporated into program design.(3) Discusses a broad strategy for seeking views of all key stakeholders on the major

program design issues.(4) Clear recognition that program design issues should draw upon a participatory

process included in the PRSP and sets out a clear strategy for Fund participations insuch a debate.

(vii) Does the brief identify policy issues where (1) No discussion of PSIA.poverty and social impact analysis (PSIA) (2) Identifies broad policy issues where poverty/social impact may be significant, but no inputs would be necessary/useful? (Explain discussion of actual impact or of how PSIA will be brought to bear on these issues.precisely what was proposed/done on PSIA (3) Areas where PSIA is needed are identified, but no comprehensive strategy for use of in comments section.) PSIA is set out.

(4) PSIA undertaken, even if in limited manner, before policy decisions are taken.

Annex 4

93


Numbers in Parentheses Refer to Assessment Reached

Section 2. Department Reviews2

(i) Is the need for more “policy space” for home- (1) Review comments remain highly prescriptive in content.grown options recognized in the review (2) Reviews suggest some limited alternative policy options, but no recognition that process? (The focus of this question is not on authorities should be given more “policy space.”the magnitude of adjustment but on how (3) Significant range of alternative policy options suggested.prescriptive review comments were.) (4) Explicit recognition in review comments that approach should be one of identifying

broad objectives and helping authorities implement homegrown options to achievethese objectives.

(ii) Are poverty issues discussed or their absence (1) Not discussed.highlighted in review department comments? (2) Poverty issues mentioned but discussion shallow.

(3) Review comments on macro and related policies are explicitly linked to the broaderpoverty objectives.

(4) Staff urged to be more proactive in the poverty reduction strategy (e.g., referencesto “key features of PRGF-supported program”).

(iii) Did departments press for more or less (1) More conditionality demanded, with no indication of priority or streamlining.conditionality (prior actions, performance (2) Recognition that streamlining (to the Fund’s core areas) required, but wide criteria, structural benchmarks) in each of the conditionality pressed for in core areas and requests that Bank strengthen its areas specified in A? (specify)3 conditionality in other areas.

(3) Strong emphasis on streamlining conditionality, but no link to ownership or consid-eration of aggregate level of Bank-Fund conditionality.

(4) Review departments recognize that ownership and streamlined conditionality arelinked, and that any conditionality should be closely associated with PRSP coreobjectives. Aggregate level of Fund-Bank conditionality explicitly considered.

1The grading scheme had the following scale: 1 = Highly inconsistent; 2 = Inconsistent; 3 = Consistent; 4 = Highly consistent.2Coverage of review department comments will be mainly on the basis of PDR, but comments from other departments (such as PDR, FAD, MFD, and RES) where

substantive program design issues are raised are also noted.3“Not applicable” could be used here to indicate that conditionality was not discussed at all.

Many of the countries concerned had no traditionof participatory policy formulation (e.g., Albania,Cambodia, and Tajikistan). Others did have a partici-patory tradition or experience, but a separate partici-patory process was launched nonetheless, in order tocomply with HIPC/BWI conditionality (e.g., Mozam-bique and Vietnam).

The participatory process was generally orga-nized by the authorities according to modalities oftheir own choosing. However, in practice, donors—including the World Bank, but not the IMF (whichrarely intervened in discussions on modalities of theparticipatory process) frequently had a strong influ-ence on the process, primarily because they fundedit, but sometimes also because the authorities them-selves did not show much interest, at least initially(e.g., Nicaragua and Vietnam), and/or did not knowhow to proceed, especially in the absence of orga-nized representation of civil society (e.g., Albania,Tajikistan, and Vietnam).

Who Participated?

In all the country case studies, good efforts weremade to involve government stakeholders beyondthe department coordinating the process, as well asthe donor community. The involvement of civil soci-ety was generally more patchy. It is to be expectedthat country choices vary with respect to the involve-ment of specific interest groups (such as religiousorganizations and indigenous groups) according totheir relevance in each country. However, the follow-ing crosscutting issues surfaced from both our casestudies and broader evidence from outside sources:

(i) Parliaments were generally not significantlyinvolved.1

(ii) Criteria for selecting civil society representa-tives lacked transparency and/or a clear ratio-nale, giving rise to criticism of poor represen-tativity and pro-government political bias ofthe stakeholders chosen (especially in Guineaand Mauritania).

(iii) The involvement of the business sector wasoften unsatisfactory, either because they werenot called in to participate in the participatoryprocess (in Tajikistan) or because they felt thatthe modalities of their involvement did notallow them any substantive role (in Albania,Mauritania, and Nicaragua), leading to theirwithdrawal. In some cases, however (Guineaand Ethiopia), private sector representativeswere relatively satisfied with both their involve-ment and their contribution to the process.

(iv) Involvement of stakeholders at the decentral-ized level was sometimes neglected (e.g., inMauritania).

(v) Also left out of the process in most cases werethe poor themselves. As a result, some com-mentators have characterized participatoryprocesses as opening up policy space only tothe benefit of a technocratic elite expert in de-velopment issues, but little qualified to speakfor the poor.2

These limits are echoed by the results of the sur-vey of PRSP stakeholders, with government stake-holders and donors broadly agreeing that their re-spective interest groups were adequately consulted,and civil society in mild disagreement with thatview.

How Was Participation Organized?

In most cases, several modalities of consultationwere used in combination. The most commonly used

Modalities of ParticipatoryProcesses: How Broad-Based HasParticipation Been?

94

ANNEX

5

1In Ethiopia and Nicaragua, the authorities view parliamentari-ans as key stakeholders, but in the latter case opposition partiesrefused to engage for political reasons. In Guinea, Tanzania, andVietnam, parliaments were given an opportunity to review thefinal document. This finding is consistent with those of externalliterature (see, for instance, Stewart and Wang, 2003). 2See, for instance, Bretton Woods Project (2003).

Annex 5

95

Tabl

e A

5.1.

Sum

mar

y E

valu

atio

n o

f the

Typ

es o

f Par

tici

pati

on

in M

acro

eco

nom

ic a

nd R

elat

ed A

reas

1

PRSP

Des

ign

Stag

e2PR

SP Im

plem

enta

tion

Stag

e3__

____

____

____

____

____

____

____

____

____

____

____

____

____

___

____

____

____

____

____

____

____

____

____

____

____

____

____

____

_In

itiat

ion

and

Initi

atio

n an

dIn

form

atio

nJo

int

deci

sion

cont

rol b

yIn

form

atio

nJo

int

deci

sion

cont

rol b

yC

ount

rysh

arin

g?C

onsu

ltatio

n?m

akin

g?st

akeh

olde

rs?

shar

ing?

Con

sulta

tion?

mak

ing?

stak

ehol

ders

?

IEO

cas

e st

udie

sG

uine

aPa

rtia

lYe

sN

oN

oLi

mite

dN

oN

oN

oM

ozam

biqu

eYe

s4Ye

sN

oN

oYe

sYe

sN

oN

oN

icar

agua

Yes4

Yes

No

No

Yes

Part

ial

No

No

Tajik

ista

nYe

sYe

sN

o—bu

t de

taile

dN

oPa

rtia

lLi

mite

d so

far

No

No

reco

mm

enda

tions

by

civ

il so

ciet

yTa

nzan

iaYe

sYe

sN

oN

oYe

sYe

sPa

rtia

l—so

me

No

inst

itutio

naliz

edfe

edba

ckV

ietn

amN

o5N

oN

oN

o5N

oN

o

OED

-onl

y ca

se s

tudi

es6

Alb

ania

Part

ial

Yes

No

No

Part

ial

No

No

No

Cam

bodi

aPa

rtia

lPa

rtia

lN

oN

on.

a.n.

a.n.

a.N

oEt

hiop

iaYe

sYe

sN

oN

oYe

sPa

rtia

l7Pa

rtia

l7N

oM

auri

tani

aPa

rtia

ln.

a.N

oN

on.

a.n.

a.n.

a.N

o

1 Thi

s ta

ble

sum

mar

izes

the

IEO

ass

essm

ent

of t

he e

xten

t to

whi

ch d

iffer

ent

type

s of

par

ticip

atio

n w

ere

impl

emen

ted

in e

ach

of t

he c

ase

stud

ies.

It d

raw

s up

on t

he d

etai

led

disc

ussi

on o

f the

par

ticip

ator

y pr

oces

s de

-sc

ribe

d in

the

indi

vidu

al c

ount

ry r

epor

ts,b

ut u

ses

the

cate

gori

es o

f pos

sibl

e in

tens

ities

of p

artic

ipat

ion

disc

usse

d in

McG

ee a

nd N

orto

n (2

000)

.In

each

cas

e,th

e as

sess

men

t is

of th

e ex

tent

of p

artic

ipat

ion

in m

acro

econ

omic

and

rela

ted

area

s (in

clud

ing

the

com

posit

ion

of b

udge

tary

exp

endi

ture

s).A

sses

smen

ts o

f the

inte

nsity

of p

artic

ipat

ion

wou

ld p

roba

bly

be h

ighe

r in

som

e so

cial

sec

tor

polic

y ar

eas

(e.g

.,he

alth

and

edu

catio

n).

2 McG

ee a

nd N

orto

n (2

000)

pro

pose

a m

ore

deta

iled

disa

ggre

gatio

n of

the

sta

ges

of t

he “

desi

gn”

cycl

e in

to “

Dia

gnos

is,”

“For

mul

atio

n,”

and

“App

rova

l.” H

owev

er,t

he in

tens

ity o

f the

par

ticip

ator

y pr

oces

s do

es n

ot s

eem

to h

ave

vari

ed w

idel

y be

twee

n su

ch p

oten

tial s

ubca

tego

ries

,so

we

have

use

d a

sim

pler

cla

ssifi

catio

n.3 I

nclu

des

mon

itori

ng a

nd e

valu

atio

n.4 I

n th

e ca

se o

f Moz

ambi

que

and

Nic

arag

ua,i

nfor

mat

ion

abou

t th

e fis

cal c

onse

quen

ces

of b

anki

ng c

rise

s—w

hich

had

impo

rtan

t im

plic

atio

ns fo

r m

acro

econ

omic

pol

icy

desi

gn a

nd k

ey t

rade

-offs

—w

ere

not

shar

ed w

ith a

broa

der

grou

p of

sta

keho

lder

s.5 C

onsu

ltatio

n on

fisc

al/m

onet

ary

polic

y w

as li

mite

d in

Vie

tnam

,but

thi

s se

ems

to h

ave

part

ly r

efle

cted

a m

utua

l agr

eem

ent

that

the

mor

e pr

essi

ng is

sues

wer

e el

sew

here

.Deb

ate

on k

ey is

sues

suc

h as

pac

e of

sta

te e

n-te

rpri

se r

efor

m w

as s

igni

fican

t,bu

t to

ok p

lace

with

in t

he t

radi

tiona

l fra

mew

ork

of t

he p

arty

.6 W

e ha

ve d

raw

n up

on t

he d

escr

iptio

ns in

the

four

cou

ntry

cas

e st

udie

s un

dert

aken

by

the

OED

alo

ne t

o m

ake

an a

sses

smen

t of

the

deg

ree

of p

artic

ipat

ion

in t

he m

acro

econ

omic

pol

icy

area

.How

ever

,the

sum

mar

yju

dgm

ents

pro

vide

d he

re a

re t

he r

espo

nsib

ility

of t

he IE

O a

lone

.In

som

e in

stan

ces,

ther

e is

not

eno

ugh

info

rmat

ion

in t

he c

ase

stud

y to

mak

e su

ch a

judg

men

t an

d th

ese

have

bee

n in

dica

ted

as “

n.a.”

7 In

Ethi

opia

,the

pri

vate

sec

tor

influ

ence

d PR

SP im

plem

enta

tion

via

busi

ness

org

aniz

atio

ns b

ut t

here

was

no

inst

itutio

naliz

ed d

ialo

gue

mec

hani

sm fo

r ci

vil s

ocie

ty.

ANNEX 5

format was large attendance workshops and confer-ences held at the national and, in most cases, also atthe regional level, with large variations in stake-holder coverage and number of workshops orga-nized. In several countries, a lot of background workfor the PRSP—and some of the material subject tobroader consultation—was prepared in workinggroups involving at least two of the three main stake-holder groups (i.e., government, civil society, anddonors). While in Guinea and Mauritania thesegroups were ad hoc and ceased to meet after the for-mulation of the PRSP, a permanent structure hasbeen put in place in Tanzania, building on the preex-isting PER working groups. In Nicaragua, such agroup was recently constituted on an ad hoc basis, toexplore tax reform issues.

Several countries made attempts to hold directconsultations with NGOs and grassroots organiza-tions, but these efforts were often impaired by thelack of organization of the sector. However, where

civil society already had a form of institutional rep-resentation (e.g., miscellaneous government spon-sored NGO umbrella groups in Guinea, CONPESin Nicaragua, and mass organizations in Vietnam)consultation of these institutions formed a core partof the participatory process. In others, such CSOumbrella organizations were established for the oc-casion of the PRSP, either by the government (e.g.,in Albania) or independently of it, with donor sup-port (e.g., in Ethiopia, Nicaragua, and Tajikistan).In many of these cases, these (or other) civil societygroups managed to prepare substantive inputs intothe PRSP. Such inputs, while rarely called for, werewelcomed in most cases (though not always takeninto consideration). However, in Nicaragua, it wasdissatisfaction with the official participatoryprocess that led certain groups (in particular anNGO umbrella group and a group of municipali-ties) to develop their own alternative version of thePRSP.

96

The IEO undertook a desk review of the 28 JSAsof full PRSPs issued through May 2003, involving asystematic analysis of their contents in areas of in-terest to the evaluation team. The table below pre-

sents the criteria used in assigning quality ratings toselected aspects of JSAs, as well as the average andmedian ratings obtained by JSAs in our sample foreach of these aspects.

Qualitative Appraisal of JointStaff Assessments

97

ANNEX

6

JSA Assessment Matrix: Coding Scheme Mean Median

1: Highly unsatisfactory 2: Unsatisfactory 3: Satisfactory 4: Highly satisfactory

Analysis of Risks to Successful PRSP Implementation 3.18 31. Risks to PRSP implementation not discussed2. Discussion of risks is limited3. Risks adequately discussed but no remedial actions suggested4. Risks exhaustively discussed along with remedial actions

Clear and Candid Assessment of the Following Key Areas:A. Ownership and Participation 2.43 21. Little or no description of participatory process and no discussion of ownership at all2. Incomplete discussion of country ownership and participation3. Good description of participatory process and discussion of ownership4. Extensive description of country ownership and participation and its impact on the

content of the strategy

B.Targets, Indicators, and Monitoring 3.36 3.5Criteria: (i) Realism, (ii) Consistency with priorities, (iii) Transparency/

Participatory methods for monitoring the PRSP, and (iv) Feedback into policy decisions

1. Partial description without assessment2. Good description but no assessment3. Good description and some assessment (1 or 2 criteria met)4. Full description and good assessment (3 or 4 criteria met)

C. Priority Public Actions

C1. Macroeconomic Framework 3.04 3Criteria: (i) Soundness, (ii) Realism, (iii) Trade-offs, and (iv) Robustness1. No assessment at all2. Assessment of soundness only3. Assessment of soundness and some qualitative assessment4. Assessment covers soundness, realism, robustness, and discussion of trade-offs

C2. Fiscal Choices 3.14 3Criteria: (i) Internal consistency, (ii) Quality of data and cost estimates, and

(iii) Administrative capacity to deliver1. No discussion at all2. Assessment covers only one criterion3. Assessment covers two criteria4. Assessment covers all three criteria

ANNEX 6

98

JSA Assessment Matrix: Coding Scheme Mean Median

C3. Financing Plan 2.79 3Criteria: (i) Realism, (ii) Sustainability, (iii) Alignment, and (iv) Contingency

plans1. No discussion at all2. Assessment covers only one criteria3. Assessment covers two criteria4. Assessment covers three or four criteria

Due Consideration of Country Situation 2.81 31. No discussion of initial conditions2. Limited discussion of domestic context with respect to the PRSP3. Discussion of the country situation is adequate, but weakly linked to PRSP4. Country situation is well discussed as well as its links to the PRSP

This annex presents details of regression analysesthat underlay some of the results discussed in Chap-ter 4. The regressions were based on data from 88new arrangements approved by the IMF under eitherthe ESAF or the PRGF during 1995–2003; 47 of thearrangements were approved as ESAFs and the restas PRGFs.

Methodology

In order to examine the relative importance ofvarious factors in the setting of targets for fiscal ad-justment in ESAF- and PRGF-supported programs,we estimated a regression model of the size of thetargeted change in the government balance exclud-ing grants, over the first and first two years of theprogram—that is, changes between T–1 and T, andbetween T–1 and T+1, respectively. We used the fol-lowing as regressors: (i) the initial level of the gov-ernment balance [gbal (T–1)]; (ii) the targetedchange in the external current account balance[∆cab]; (iii) the initial level of grants in the govern-ment budget [grants (T–1)]; (iv) the projectedchange in the level of grants [∆grants]; and (v) pro-jected growth [growth].1

We postulated a two-way relationship betweentargeted fiscal adjustment and targeted external cur-

rent account adjustment, and employed two-stageleast squares (with robust standard errors) to accountfor the endogeneity of the targeted change in the ex-ternal current account.

We replicated the regressions using outturn datain place of targeted/projected data in order to com-pare the determinants of targeted and actual fiscaladjustment.

Results

We obtained broadly similar results for targetedfiscal adjustment under ESAFs and PRGFs over atwo-year horizon: the targeted magnitude of fiscaladjustment was inversely related to the initial levelof the fiscal balance and to the projected change inthe level of grants (top half of Table A7.1). Specifi-cally, over the two-year horizon, the targeted fiscaladjustment reflected a near-halving of the initial fis-cal deficit and a full adjustment to projected changein the availability of grants.

The estimations using outturn data indicated sig-nificant influence of the initial fiscal balance andchange in the level of grants as determinants of ac-tual fiscal adjustment, but only for the initial pro-gram year of ESAFs (bottom half of Table A7.1).The limited number of observations for outturnsunder the PRGF suggest caution in interpreting theresults for the PRGF, but the influence of the initiallevel of fiscal balance is no longer significant.

Determinants of Targeted andActual Fiscal Adjustment

99

ANNEX

7

1All variables were measured as ratios to GDP, except growth,which was measured in percent a year.

ANNEX 7

100

Table A7.1. Determinants of Targeted and Actual Fiscal Adjustment in ESAFs and PRGFs

ESAFs PRGFs________________________ ________________________Horizon Horizon Horizon Horizon

T T+1 T T+1

Dependent variable: targeted fiscal adjustmentConstant –1.60* –2.6 0.56 2.53gbal (T–1) –0.25** –0.46*** –0.29 –0.52**

∆cab 0.33 0.32* 0.52 0.22∆grants –0.43** –0.98** –1.04* –1.11***

grants (T–1) –0.16 –0.51** –0.43 0.65growth (T) 0.19 –0.08 –0.08 0.48growth (T+1) 0.47 –0.96

N 44 43 40 39

Dependent variable: actual fiscal adjustmentConstant –1.6** –1.97 –5.15 –6.37gbal (T–1) –0.32** –0.14 –0.60 –0.31∆cab –0.13 –0.36 1.02 0.89∆grants –0.46** –0.52 1.23 0.32grants (T–1) –0.27 –0.11 –0.20 0.46growth (T) 0.13 –0.06 0.14 0.14growth (T+1) 0.29 0.29*

N 45 45 32 23

Note: The asterisks denote levels of statistical significance: * = 10 percent, ** = 5 percent, and *** = 1 percent.

Composition of TargetedChanges in Expenditures andExternal Financing in IMF-Supported Programs

ANNEX

8

The Composition of ProgrammedFiscal Adjustment

As discussed in Chapter 4, PRGF-supported pro-grams target, on average, greater increases in rev-enues and smaller expenditure cuts than their ESAFcounterparts. These results are reinforced by lookingat the distribution of targeted changes in expendituresin ESAF- and PRGF-supported programs. FigureA8.1 shows that whereas most ESAF arrangementsproject expenditure reductions (Panel A), a majorityof PRGF-supported programs accommodate higherexpenditures (Panel B).

Another way of studying the evolution of pro-jected revenues and expenditures in PRGF-sup-ported programs is to regress the programmedchange in expenditures (or revenues) on a number ofvariables that might be expected to affect these tar-gets (i.e., previous expenditure/revenue levels, GDPgrowth, targeted fiscal adjustment, etc.) and includea dummy variable to measure the impact of PRGF-supported programs. The dummy variable shouldcapture changes in programmed expenditures (or

revenues) due to the introduction of the PRGF, con-trolling for other factors that might be expected toinfluence the setting of expenditure (or revenue) tar-gets. The results of this exercise suggest that, hold-ing all else constant, PRGF-supported programstarget total expenditure levels about 1.6 percent ofGDP higher than their ESAF counterparts. Simi-larly, controlling for all else, PRGF-supported pro-grams target a similar increase in revenues.

The Availability of External Financing

Figure A8.2 shows that there is a great degree ofvariability in the projected amounts of net externalpublic sector financing in both ESAF- and PRGF-supported programs. However, the center of the dis-tribution shifts substantially to the right in the caseof PRGF-supported programs. On average, whilePRGF-supported programs project increases in theavailability of external budget financing in the firsttwo program years, they project a strong decline inthe third year.

101

ANNEX 8

102

Figure A8.1. Distribution of Programs According to the Targeted Change in Expenditures

Perc

ent

shar

e of

pro

gram

s

Change in expenditures(In percent)

30

20

10

0–15 –10 –5 0 5

A. ESAF

Perc

ent

shar

e of

pro

gram

s

Change in expenditures(In percent)

30

20

10

0–20 –10 0 10 20

B. PRGF

Figure A8.2. Distribution of Projected Changes in the Availability of Grants and Concessional Loans in ESAF- and PRGF-Supported Programs

Perc

ent

shar

e of

pro

gram

s

Change in net external financing(In percent of GDP)

20

10

0–15 –10 –5 0 5

A. ESAF

Perc

ent

shar

e of

pro

gram

s

Change in net external financing(In percent of GDP)

20

10

0–20 –10 0 10

B. PRGF

Changes in Domestic PolicyProcesses and Institutions:Evidence from Case Studies

ANNEX

9

103

Table A9.1. What Happened to Domestic Policy Formation Under the PRSP Approach? SomeExamples from the Case Studies

Did the Debate Lead to a Broadened Policy Space, e.g.,

Did a Broader-Based Policy Considering Additional/Country and Policy Issue Debate Take Place on the Issue? Alternative Options? Outcome

Guinea1. State enterprise reform No. No. Still no country-owned strategy that could

be supported by the IFIs.2. Governance Yes, initially under the pressure of Yes. Debate led to large range Comprehensive strategy defined,

donor conditionality. Debate of solutions being considered, emphasizing decentralization and capacity started prior to PRSP process but taking a very broad view of building as a solution.Very limited progress was substantially amplified by it. governance problems. in implementation. Governance problems

remain pervasive, at all levels of govern-ment. Local CSOs supported BWIconditionality in this area.

Mozambique1. Petroleum tax increase No broad-based debate but the The PSIA considered two main The PSIA concluded that “the aggregate

IMF and the authorities agreed to scenarios on updating the short-term impact of a rise in fuel tax on await the outcome of a PSIA on specific tax that had not been poverty is modest.”the impact of the measure before adjusted for five years: (i) in-its implementation. crease the tax in line with The tax on petroleum products was

exchange rate movements (a increased by an average of 62.5 percent in 100 percent increase in the May 2003.tax); and (ii) increase the tax in line with domestic inflation (a 65 percent increase in the tax).

2. Rationalization of tax No broad-based debate; the No. The government initially complained to the incentive and exemptions measures were recommendations IMF that the recommendations were

of an IMF technical assistance “unhelpful” but was subsequently mission. persuaded to implement some of the

measures.Nicaragua1. Growth strategy Not initially, but consultation later Initially, the debate was not New growth strategy endorsed at

on pushed by donor community opened up, as the new govern- Consultative Group meeting and work in (see Box 2.1). ment chose not to “rock the progress to determine how to modify

boat” (and delay progress on PRSP.HIPC debt relief) even though it disagreed with the growth strategy in the PRSP.Subsequently, however, a broader debate did begin.

2. Tax reform Debate took place in the context Yes, IMF showed greater Tax reform approved. Passage of reform in of a technical commission flexibility to reform design, assembly eased by IMF flexibility to design comprising representatives of provided expected net revenue considerations.various domestic stakeholders. was maintained.

3. Bank restructuring No—but proposed approach was No—although some debate Strategy for bank restructuring was“owned” by key government did take place ex post on the determined by the authorities.officials. choices made on “burden-

sharing.”

ANNEX 9

104


Did the Debate Lead to a Broadened Policy Space, e.g.,

Did a Broader-Based Policy by Considering Additional/Country and Policy Issue Debate Take Place on the Issue? Alternative Options? Outcome

Tajikistan1. Gas tariff reform and No broad-based debate. No. Provision was made in the budget for some

compensation to poor compensation to poor households.households However, the initial compensation scheme

was ill-designed. It would have been betterto have undertaken a proper PSIA beforeimplementation of the tariff increase.

2. Limit on external No broad-based debate. No. Although not part of the The authorities went along with the limit,borrowing to finance formal conditionality in the albeit reluctantly, because they thought it PIP PRGF-supported program, it constrained their use of available

is generally understood to be concessional loans for growth-promoting an important condition for public investments.completing program reviews.

Tanzania1. Macroeconomic pro- Yes—but not as part of initial Yes. A serious debate on fiscal Substantial changes to program design

gram design PRSP. Debate took place later and monetary policy design were eventually made (see Box 4.2).within domestic framework ensued, involving government,(PER system). donors, and the IMF. Civil

society largely uninvolved.

2. Trade policy No. The bulk of trade-related Trade issues were mentioned Trade policy remains amenable to reforms were undertaken before with very little specificity in the protectionist pressure from local the launch of the PRGF and PRSP PRSP, although the importance producers. IMF staff have argued against in 2000. Subsequently, no trade of ensuring that the poor such measures, but PRGF-supported policy debate that was distinct shared in the benefits of programs have not included any trade-from the broader macroeconomic globalization was stressed. In related conditionality.issues has emerged. the absence of a broad

domestic discussion—even the broadly participatory PER process has given it less attention than other issues—no alternative policy approaches were discussed.

3. Fiscal federalism Although under implementation The IMF adopted a cautious Capacities at the local level have remained for close to a decade, local self- stance on fiscal decentral- low, affecting the ability to monitor priority governance received renewed ization, warning that capacities expenditures. The government has impetus under the PRS process— at the local level were still embarked on the training of local-level seen as a key point of departure very weak, and that expanding staff, with the assistance of local NGOs.for poverty eradication. Financial the process would have decentralization and the need to implications for the integrity create local capacity for the of the central budget. The IMF management of public funds were did not press the government natural consequences. to abandon its plans, but did

warn against rapid implement-ation and pushed for introduc-tion at local levels of a more transparent and efficient financial management system.

Vietnam1. State enterprise reform The PRSP represented an “action No, but the policy debate and Bank and Fund came to accept the

plan” for the implementation of resulting options were clearly authorities’ approach to SOE reform, albeit the five-year socioeconomic country-driven. not until after pressing a privatization/development plan that was equitization agenda at the start of the approved at the Party Congresses PRGF-supported program and in the in which various “mass organiza- context of the first PRSC.tions” were represented. Meaning-ful discussion affecting policy formulation did not take place in the context of the PRSP participa-tory exercise.

Annex 9

105

Table A9.2. Institutional Changes: Evidence from Case Studies

Country Macroeconomic Policymaking PRSP Links to Budget Monitoring and Evaluation

Guinea Analysis and discussion of macro- There are no systematic links PRSP itself did not provide much detail on economic conditions and policies are between the PRSP policy matrix, institutional arrangements for Monitoring andconducted within a very narrow circle sectoral strategies, expenditure Evaluation. But a group was set up in October around the Minister of Finance and plans of line ministries, and regional 2002 to: (i) determine set of indicators to beGovernor of the central bank. poverty reduction strategies. monitored; (ii) arrange for their production if

Nevertheless, budget allocations to not already available; and (iii) propose institu-designated priority sectors have been tional arrangements for publication of indica-relatively protected from ad hoc cuts. tors and for providing feedback from monitor-

ing to updating and improving the PRSP.

Mozambique There continues to be little public Integration of the PARPA (Mozambi- Targets vary considerably in scope and preci-discussion of macroeconomic policy que’s PRSP) into the government’s sion. In a few areas targets are yet to be spec-issues. Donors providing general planning, budgeting, and reporting ified (e.g., anticorruption).budget support have formed a macro- processes is under way but at a economic working group that slow pace. Weak institutional capacity is adversely discusses macroeconomic policy affecting the quality of monitoring.issues, but membership is limited to An annually updated MTEF is the the donor community. main instrument for translating A Poverty Observatory has been establish-

PARPA priorities into budgetary ed to review PARPA implementation allocations. But PEM system needs annually, with a focus on drawing lessons and strengthening to ensure that budget- making recommendations for improving ed funds reach spending units and implementation.that funds are spent as budgeted.

Role of nongovernment actors in PARPAmonitoring not yet well defined.

Nicaragua No significant movement toward Little linkage between the PRSP and The PRSP included a comprehensive set of opening up debate on the macro- the budget process. targets and indicators for monitoring the economic framework and alternative impact of policies on the poor. Links betweenpolicy trade-offs. Recently, there Poor PEM system is a major hindrance goals and intermediate indicators are clearlyseems to be some movement in the to implementation of the PRSP. set out in the social area, but there are nocontext of the new development plan. intermediate indicators for the economic/

Fragmentation of policymaking has production sectors.generated problems for coordination and control over public finances. A comprehensive system for monitoring and

evaluation of PRSP implementation is being setup (SECEP, SINASIP).

Tajikistan There was some public discussion of There is a disconnect between the Five of nine main poverty reduction targets aremacroeconomic policy issues during PRSP, the PIP, and the government linked to the MDGs.the formulation of the PRSP process. budget process. In particular, the But macroeconomic policy formulation PRSP and the PIP appear to have The PRSP policy matrix contains some 200 on an ongoing basis seems to revolve different priorities (the former heavily indicators, most of which refer to processesaround discussions between IMF staff tilted to social sectors and the latter and policy measures, rather than to outcomes.and a handful of government and to infrastructure). Work is under way to develop a more focusedcentral bank officials. set of intermediate indicators.

Initial effort at an MTEF was based on only three sectors. Efforts are A PRSP Implementation and Monitoring Unit under way to broaden its coverage. has been established in the Presidency to coor-

dinate monitoring and evaluation activities.

Tanzania Macroeconomic policies now more The PRS has provided a poverty- The PRS process has substantially enhancedfrequently and openly discussed in focused policy framework for budget- national processes for poverty monitoring, al-broadly inclusive forums, but little ing that was lacking before. But though there are some inconsistencies in theimpact of nongovernment actors on spending within sectors is not always follow-through of actual PRS targets.policy choices. well targeted.

Establishing a realistic mechanism for There is a major gap between PRS monitoringimplementation at the subnational and feedback to policy formulation.level remains a major challenge.

IFMS is a major instrument for improving PEM and accountability.

ANNEX 9

106


Country Macroeconomic Policymaking PRSP Links to Budget Monitoring and Evaluation

Vietnam Limited public discussion of macro- The CPRGS (Vietnam’s PRSP) is Institutionalization of CPRGS monitoring andeconomic policy issues. linked to the socioeconomic develop- evaluation of implementation is at an early

ment plans—the SEDPs (products of stage. An interministerial steering committeeVietnam’s indigenous planning pro- has been established, headed by the Deputy cess). However, the CPRGS falls Prime Minister. Each ministry has responsibility short of being an “action plan;” in for monitoring a specified set of indicators some important areas (e.g., SOE re- and to report back to the steering committee.form) the strategy is not sufficiently operational. The CPRGS contains an extensive system of

136 indicators to monitor economic develop-There is no fully costed and priori- ments and poverty reduction at both program tized MTEF. and aggregate levels.

At end-2003, there was not yet in place aframework of indicators for monitoring thelinkages between policy measures and out-comes. Also, for about one-third of theindicators, information was either not availableor of poor quality.

ActionAid, 2002, “Inclusive Circles Lost in Exclusive Cy-cles,” Contribution to the Comprehensive Reviews ofthe Poverty Reduction Strategy Papers by the WorldBank and the IMF, January (London: ActionAid).

Adam, Christopher, and David Bevan, 2001, “PRGFStocktaking Exercise on Behalf of DFID,” November(Oxford: Oxford University).

Baqir, Reza, Rodney Ramcharan, and Ratna Sahay, 2003,“The Consistency of IMF Programs,” paper presentedat the IMF Research Workshop on MacroeconomicChallenges in Low-Income Countries (Washington:International Monetary Fund).

Berg, Andrew, and Anne O. Krueger, 2003, “Trade,Growth, and Poverty: A Selective Survey,” IMFWorking Paper No. 03/30 (Washington: InternationalMonetary Fund).

Bevan, David, 2000, “Public Expenditure Review: Tanza-nia 1999/2000—Fiscal Policy and the Macroeco-nomic Context,” Department of Economics (Oxford:Oxford University).

———, 2001, “Tanzania Public Expenditure Review:2000/2001—The Fiscal Deficit and the Sustainabilityof Fiscal Policy,” Department of Economics (Oxford:Oxford University).

———, 2004, “Background Paper on Public ExpenditureManagement (PEM) Linkages to the PRSP Process,”Operations Evaluation Department, March (Washing-ton: World Bank).

Bird, Graham, 2004, “Growth, Poverty and the IMF,”Journal of International Development, Vol. 16, Issue4 (May), pp. 621–36.

———, and Thomas D. Willet, 2003, “IMF Conditional-ity, Implementation and the Political Economy ofOwnership,” Surrey Centre for International Eco-nomic Studies (Surrey: University of Surrey).

Booth, David, 2001, “PRSP Processes in Eight AfricanCountries: Initial Impacts and Potential for Institu-tionalisation,” paper presented at the WIDER Devel-opment Conference on Debt Relief (Helsinki: WorldInstitute for Development Economics Research).

———, 2003, “Introduction and Overview,” DevelopmentPolicy Review, Vol. 21, No. 2, pp. 131–59.

———, and Henry Lucas, 2002, “Good Practice in theDevelopment of PRSP Indicators and MonitoringSystems,” Working Paper No. 172 (London: OverseasDevelopment Institute).

Bourguignon, François, 2003, “The Poverty-Growth-Inequality Triangle,” paper presented at a Confer-ence on Poverty, Inequality and Growth (Paris:

Agence Française de Développement/EU Develop-ment Network).

Branson, William, and Nagy Hanna, 2000, “Ownershipand Conditionality,” OED Working Paper Series No.8 (Washington: World Bank).

Bretton Woods Project, 2003, “PRSPs: Political Space atWhose Expense?” Global Policy Forum, September(Washington: Bretton Woods Project).

Buira, Ariel, ed., 2003, Challenges to the World Bank andIMF: Developing Country Perspectives (London: An-them Press).

Bulír̆, Ales̆, and Javier A. Hamann, 2001, “How Volatileand Unpredictable Are Aid Flows, and What Are thePolicy Implications?” IMF Working Paper No.01/167 (Washington: International Monetary Fund).

CAFOD, Oxfam International, World Vision, ChristianAid, Bretton Woods Project, EURODAD, Save theChildren, WaterAid, 2003, “Where Is the Impact?”Joint submission for debt sustainability review, Au-gust (Brussels: European Network for Debt and De-velopment).

Cardoso, Eliana, 1992, “Inflation and Poverty,” NBERWorking Paper No. 4006 (Cambridge, Massachusetts:National Bureau of Economic Research).

Cashin, Paul, Paolo Mauro, Catherine A. Pattillo, andRatna Sahay, 2001, “Macroeconomic Policies andPoverty Reduction: Stylized Facts and an Overviewof Research,” IMF Working Paper No. 01/135 (Wash-ington: International Monetary Fund).

Chu, Ke-young, Hamid Davoodi, and Sanjeev Gupta,2000, “Income Distribution and Tax and GovernmentSocial Spending Policies in Developing Countries,”IMF Working Paper No. 00/62 (Washington: Interna-tional Monetary Fund).

CIDSE/Caritas, 2004, “PRSP: Are the World Bank andIMF Delivering on Promises?” Submission to IEOEvaluation of PRSP/PRGF (Dublin: Trocaire).

Cling, Jean-Pierre, Mireille Razafindrakoto, and FrançoisRoubaudk eds., 2003, New International Poverty Reduction Strategies, Routledge Studies in Devel-opment Economics No. 35 (London, New York:Routledge).

Coyle, Erin, and Alison Evans, 2003a, “Experience withPRSP in Transition Countries,” PRSP Monitoring andSynthesis Project (London: U.K. Department for In-ternational Development and Overseas DevelopmentInstitute).

———, 2003b, “Experience of PRSPs in Asia,” PRSPMonitoring and Synthesis Project (London: U.K.

Bibliography

107

BIBLIOGRAPHY

Department for International Development and Over-seas Development Institute).

Craig, David, and Doug Porter, 2003, “Poverty ReductionStrategy Papers: A New Convergence,” World Devel-opment, Vol. 31, No. 1, pp. 53–69.

Dahou, Karim, Tarik Dahou, and Eric Hazard, 2001, “Par-ticipation and Consistency in Strategic Frameworksfor Poverty Reduction: A Comparative Study of Sene-gal and Burkina Faso” (Washington: Oxfam America;and Paris: ENDA Tiers Monde).

Department for International Development, 2002, “Sub-mission to the World Bank and the IMF Review of thePRSPs” (London: U.K. Department for InternationalDevelopment).

———, 2003, “Learning the Lessons on PSIA: A Syn-thesis of Experience from the DFID Pilot Studies,”PRSP Briefing Note 8, Monitoring and SynthesisProject (London: U.K. Department for InternationalDevelopment).

Diamond, Jack, Matt Davies, Pokar Khemani, and Feri-doun Sarraf, 2003a, “Assessment of Technical Assis-tance in PEM Area in Sub-Saharan Francophonecountries” (unpublished; Washington: InternationalMonetary Fund).

———, 2003b, “Public Expenditure Management Reformin Anglophone African Countries: An Assessment ofFAD TA” (unpublished; Washington: InternationalMonetary Fund).

Drazen, Allan, 2001, “Conditionality and Ownership inIMF Lending: A Political Economy Approach,” paperpresented at the Second Annual IMF Research Con-ference (Washington: International Monetary Fund).Also, IMF Staff Papers, Vol. 49 (2002), Special Issue,pp. 36–67.

Easterly, William, 2000, “The Effect of International Mon-etary Fund and World Bank Programs on Poverty,”Policy Research Working Paper No. 2517 (Washing-ton: World Bank).

———, 2003, “Tropics, Germs, and Crops: How Endow-ments Influence Economic Development,” Journal ofMonetary Economics, Vol. 50 (January), pp. 3–39.

Eberlei, Walter, 2002, “Institutionalisation of Participationin PRS processes,” available via the Internet athttp://www.uneca/prsp/docs/Eberlei_paper.htm.

———, and Heike Henn, 2003, “Parliaments in Sub-Saharan Africa: Actors in Poverty Reduction?”(Bonn: Deutsche Gessellschaft für TechnischeZusammenarbeit).

European Commission and World Bank, 2003, “Joint Noteon Supporting the PRSP Process in Africa,” paperprepared by European Commission and World BankStaff (Brussels: European Commission; Washington:World Bank).

EURODAD, 2003a, “Streamlining of Structural Condi-tionality—What Has Happened?” (Brussels: Euro-pean Network for Debt and Development).

———, 2003b, “Brief Review of the Chapter on theMacroeconomic Issues in the PRSP Sourcebook,”submission to the IEO (Brussels: European Networkfor Debt and Development).

Evrard, Jean-Pol, 2003, “Évaluation du processus DSRPFMI-Banque mondiale,” Contribution de CaritasFrance (Paris: Caritas France).

Fischer, Stanley, 1993, “The Role of Macroeconomic Fac-tors in Growth,” NBER Working Paper No. 4565(Cambridge, Massachusetts: National Bureau of Eco-nomic Research).

Foster, Mick, Adrian Fozzard, Felix Naschold, and TimConway, 2002, “How, When and Why Does PovertyGet Budget Priority: Poverty Reduction Strategy andPublic Expenditure Reform in Five African Coun-tries,” Working Paper No. 168 (London: Overseas De-velopment Institute).

Fox, James, 2003, “Poverty Reduction Strategy Papers:Review of Private Sector Participation” (Washington:USAID).

Frankel, Jeffrey, 2002, “Promoting Better National Institu-tions: The Role of the IMF,” paper presented at theIMF Third Annual Research Conference, November(Washington: International Monetary Fund). Also, IMFStaff Papers, Vol. 50 (2003), Special Issue, pp. 21–30.

General Accounting Office (GAO), 2001, InternationalMonetary Fund—Few Changes Evident in Design ofNew Lending Program for Poor Countries, GAO-01-581 (Washington: United States, General AccountingOffice).

Ghosh, Atish, and Steven Phillips, 1998, “Warning: Infla-tion May Be Harmful to Your Growth,” IMF Staff Pa-pers, Vol. 45 (December), pp. 672–710.

Goldstein, Morris, 2001, “IMF Structural Conditionality:How Much Is Too Much?” revision of paper presentedat a 2000 NBER Conference on “Economic and Fi-nancial Crises in Emerging Market Economies”(Washington: Institute for International Economics).

Gomes, Rafael, and Max Lawson, 2003, “Pro-PoorMacroeconomic Policies Require Poverty and SocialImpact Analysis,” Economic Policy EmpowermentProgram, May (Brussels: European Network on Debtand Development).

Grusky, Sara, 2000, “The IMF and World Bank Initiate aNew Reform Package—The Poverty Reduction Strat-egy Papers: An Initial NGO Assessment,” Debt & De-velopment Dossier, No. 3, April (Washington: Breadfor the World Institute).

Gunter, Bernhard G., 2002, “Towards Poverty ReducingMacroeconomic Policies—Strategy Paper,” paperprepared for the New Rules for Global Finance Coali-tion (Washington: New Rules for Global Finance).

Gupta, Sanjeev, Mark Plant, Benedict Clements, ThomasDorsey, Emanuele Baldacci, Gabriela Inchauste,Shamsuddin Tareq, and Nita Thacker, 2002, “Is thePRGF Living Up to Expectations? An Assessment ofProgram Design,” IMF Occasional Paper No. 216(Washington: International Monetary Fund).

Hajro, Zlata, and Joseph P. Joyce, 2004, “A True Test: DoIMF Programs Hurt the Poor?” Working Paper No. 04-01, Department of of Economics (Wellesley,Massachusetts: Wellesley College).

Hewitt, Adrian, and Ian Gillson, 2003, “A Review of theTrade and Poverty Content in PRSPs and Loan-Related Documents,” Commissioned Report (London:Christian Aid).

Inchauste, Gabriela, 2002, “Poverty and Social ImpactAnalysis in PRGF-Supported Programs,” IMF PolicyDiscussion Paper No. 02/11 (Washington: Interna-tional Monetary Fund).

108

Bibliography

International Labor Organization (ILO), 2002, PRSPs: AnAssessment of the ILO’s Experience (Geneva: ILOCommittee on Employment and Social Policy).

Independent Evaluation Office (IEO), 2002, Evaluation ofProlonged Use of IMF Resources (Washington: Inter-national Monetary Fund).

———, 2003, Fiscal Adjustment in IMF-Supported Pro-grams (Washington: International Monetary Fund).

International Monetary Fund (IMF), 1998, External Eval-uation of the ESAF (Washington: International Mone-tary Fund)

———, 1999a, “Poverty Reduction and Growth Facility—Operational Issues,” SM/99/293, December (Wash-ington: International Monetary Fund).

———, 1999b, “Concluding Remarks by the Chairman;Poverty Reduction Strategy Papers: Operational Is-sues and Poverty Reduction and Growth Facility—Operational Issues,” BUFF/99/156, Executive BoardMeeting 99/135, December (Washington: Interna-tional Monetary Fund).

———, 2000a, “Key Features of IMF PRGF-SupportedPrograms,” SM/00/193, August (Washington: Interna-tional Monetary Fund).

———, 2000b, “Key Features of IMF PRGF-SupportedPrograms—Review Checklist,” October (Washing-ton: International Monetary Fund).

———, 2000c, “Communication Guide,” External Rela-tions Department (Washington: International Mone-tary Fund).

———, 2001a, Code of Good Practice on Fiscal Trans-parency, March, available via the Internet athttp://www.imf.org/external/np/fad/trans/code.htm.

———, 2001b, “Revised Guidelines on Poverty-RelatedWork,” August (Washington: International MonetaryFund).

———, 2002a, “Actions to Strengthen the Tracking ofPoverty-Reducing Public Spending in Heavily In-debted Countries,” SM/02/30, January (Washington:International Monetary Fund).

———, 2002b, “Synopses of External Comments andContributions on the Joint IMF/World Bank Staff Re-view of the PRSP Approach,” January (Washington:International Monetary Fund).

———, 2002c, “Review of the Poverty Reduction andGrowth Facility—Issues and Options,” SM/02/51,Revision 1 (Washington: International MonetaryFund).

———, 2002d, “Summing up by the Acting Chair; Re-view of the Poverty Reduction and Growth Facility—Issues and Options,” BUFF/02/40, Executive BoardMeeting 02/24, March (Washington: InternationalMonetary Fund).

———, 2002e, World Economic Outlook, September(Washington: International Monetary Fund).

———, 2003a, “Update on Implementation of ActionPlans to Strengthen Capacity to Track Poverty-Reducing Public Spending,” March (Washington: In-ternational Monetary Fund).

———, 2003b, IMF Macroeconomic Research on Low In-come Countries, report prepared by Fiscal Affairs, Pol-icy Development and Review, and Research Depart-ments (Washington: International Monetary Fund).

———, 2003c, “Aligning the PRGF and the PRSP Ap-proach—Issues and Options,” SM/03/04, March(Washington: International Monetary Fund).

———, 2003d, World Economic Outlook, April (Washing-ton: International Monetary Fund).

———, 2003e, “Concluding Remarks by Acting Chair;Aligning the Poverty Reduction and Growth Facility(PRGF) and the Poverty Reduction Strategy (PRSP)Approach—Issues and Options,” BUFF/03/55, Exec-utive Board Seminar 03/2, April (Washington: Inter-national Monetary Fund).

———, 2003f, “The Role of the Fund in Low-IncomeCountries over the Medium-Term—Issues Paper forDiscussion,” paper prepared by the Staff of the PolicyDevelopment and Review Department, July (Wash-ington: International Monetary Fund).

———, 2003g, “Concluding Remarks by the Chair onRole of the Fund in Low-Income Member Countriesover the Medium Term and Fund Assistance forCountries Facing Exogenous Shocks,” BUFF/03/164,Executive Board Meeting 03/82, August (Washing-ton: International Monetary Fund).

———, 2003h, “Concluding Remarks by the Chair;Poverty Reduction Strategy Papers—Progress in Im-plementation,” BUFF/03/165, Executive Board Meet-ing, September (Washington: International MonetaryFund).

———, 2003i, “Guide for Staff Relations with CSOs,”Guidance Note (Washington: International MonetaryFund).

———, 2003j, “The Acting Chair’s Summing Up—Re-view of Fund’s External Communications Strategy,”BUFF/03/32, Executive Board Meeting (Washington:International Monetary Fund).

———, 2004a, “The Fund’s Support of Low-IncomeMember Countries—Considerations on Instrumentsand Financing,” SM/04/53, February (Washington:International Monetary Fund).

———, 2004b, “Debt Sustainability in Low-IncomeCountries—Proposal for an Operational Frameworkand Policy Implications,” SM/04/27, February (Wash-ington: International Monetary Fund).

———, 2004c, “Report of the Task Force on the Organiza-tion and Management of the African Department,”February (Washington: International Monetary Fund).

———, 2004d, “Technical Assistance Evaluation Pro-gram—Findings of Evaluations and Updated Pro-gram,” SM/04/67, March (Washington: InternationalMonetary Fund).

Johnson, John H., and Sulaiman S. Wasty, 1993, “Bor-rower Ownership of Adjustment Programs and thePolitical Economy of Reform,” World Bank Discus-sion Paper No. 199 (Washington: World Bank).

Khan, Mohsin S., and Malcolm, D. Knight, 1981, “Frame-work for Stabilization Programs in Developing Coun-tries; A Formal Framework,” IMF Staff Papers, Vol.28 (March), pp. 1–53.

———, 1985, “Fund-Supported Adjustment Programs andEconomic Growth,” IMF Occasional Paper No. 41(Washington: International Monetary Fund).

Khan, Mohsin S., and Abdelhak S. Senhadji, 2000,“Threshold Effects in the Relationship Between Infla-

109

BIBLIOGRAPHY

tion and Growth,” IMF Working Paper No. 00/110(Washington: International Monetary Fund).

Killick, Tony, 2002, “The ‘Streamlining’ of IMF Condi-tionality: Aspirations, Reality and Repercussions,”Report for U.K. Department for International Devel-opment, April (London: Overseas Development Institute).

Knoke, Irene, and Pedro Morazan, 2002, “PRSP: Beyondthe Theory: Practical Experiences and Positions of In-volved Civil Society Organizations” (Washington:Bread for the World).

Ladd, Paul, 2003, “Too Hot to Handle? The Absence ofTrade Policy from PRSPs” (London: Christian Aid).

Levinsohn, Jim, 2003, “The World Bank’s Poverty Reduc-tion Strategy Paper Approach: Good Marketing orGood Policy?” G-24 Discussion Paper Series No. 21(Geneva: United Nations Conference on Trade andDevelopment).

Lucas, Robert E., 2002, Lectures in Economic Growth(Cambridge, Massachusetts: Harvard UniversityPress).

Malaluan, Jenina J.C., and Shalmali Guttal, 2002, “Struc-tural Adjustment in the Name of the Poor: The PRSPExperience in the Lao PDR, Cambodia and Vietnam,”Report, Focus on the Global South, January(Bangkok: Chulalongkorn University).

McGee, Rosemary, and Andy Norton, 2000, “Participationin Poverty Reduction Strategies: A Synthesis of Experi-ences with Participatory Approaches to Policy Design,Implementation and Monitoring,” IDS Working Paper,December (Brighton: IDS, University of Sussex).

McGee, Rosemary, Josh Levene, and Alexandra Hughes,2002, “Assessing Participation in PRSPs: A Desk-based Synthesis of Experience in Sub-SaharanAfrica,” Research Report No. 52 (Brighton: IDS, Uni-versity of Sussex).

Molenaers, Nadia, and Robrecht Renard, 2002, “Strength-ening Civil Society from the Outside? Donor-DrivenConsultation and Participation Processes in PovertyReduction Strategies (PRSP): the Bolivian Case,”IDPM/UA Discussion Paper 2002-5 (Antwerp: Insti-tute of Development Policy and Management, Uni-versity of Antwerp).

Munoz, Sonia, and Stanley Sang-Wook Cho, 2003, “So-cial Impact of Tax Reform: The Case of Ethiopia,”IMF Working Paper No. 03/232 (Washington: Inter-national Monetary Fund).

Nordic Countries, 2003, “Review of Nordic Monitoring ofthe World Bank and the IMF Support to the PRSPProcess,” March (available on EURODAD website:http://www.eurodad.org).

Operations Evaluation Department (OED), 2003, “DebtRelief for the Poorest—An OED Review of the HIPCInitiative,” Evaluation Report (Washington: WorldBank).

———, 2004, “OED Review of the Poverty ReductionStrategy (PRS) Process” (Washington: World Bank).

Oxfam, 2003, “The IMF and the Millennium Goals—Fail-ing to Deliver for Low Income Countries,” September(Washington: Oxfam International).

———, 2004, “From ‘Donorship’ to Ownership?” OxfamBriefing Paper, January (London: Oxfam).

Posner, Richard, A., 2003, Law, Pragmatism and Democ-racy (Cambridge, Massachusetts: Harvard UniversityPress).

Richmond, Jennie, and Paul Ladd, 2001, “Ignoring the Ex-perts: Poor Peoples’ Exclusion from Poverty Reduc-tion Strategies,” Christian Aid Policy Briefing, Octo-ber (London: Christian Aid).

Robb, Caroline M., 2003, “Poverty and the Social ImpactAnalysis—Linking Macroeconomic Policies toPoverty Outcomes: Summary of Early Experiences,”IMF Working Paper No. 03/43 (Washington: Interna-tional Monetary Fund).

———, and Alison M. Scott, 2001, “Reviewing SomeEarly Poverty Reduction Strategy Papers in Africa,”IMF Policy Discussion Paper No. 01/5 (Washington:International Monetary Fund).

Rodrik, Dani, 2003, “Growth Strategies,” paper, John F.Kennedy School of Government, September (Cam-bridge, Massachusetts: Harvard University).

———, and Arvind Subramanian, 2003, “The Primary ofInstitutions: and What This Does and Does NotMean,” Finance & Development, Vol. 40, No. 2, pp.31–34.

Srinivasan, T.N., 2001, “Growth and Poverty Alleviation:Lessons from Development Experience,” ADB Insti-tute Research Paper No. 17 (Tokyo: Asian Develop-ment Bank Institute).

Stewart, Frances, and Michael Wang, 2003, “Do PRSPsEmpower Poor Countries and Disempower the WorldBank, or Is It the Other Way Round?” Working PaperNo. 108 (Oxford: Queen Elizabeth House).

United Nations, 2002, “Draft Guidelines: A Human RightsApproach to Poverty Reduction Strategies” (NewYork: United Nations); available via the Internet athttp://www.unhchr.ch/development/povertyfinal.html.

United Nations Conference on Trade and Development(UNCTAD), 2003, Back to Basics: Market Access Is-sues in the Doha Agenda (Geneva: UNCTAD).

United Nations Development Program (UNDP), 2002,Human Development Report (New York: United Na-tions).

———, 2003, Evaluation of UNDP’s Role in the PRSPProcess (New York: UNDP Evaluation Office).

White, Howard, 1999, Dollars, Dialogue and Develop-ment, SIDA Evaluation Report (Stockholm: SwedishAgency for International Development Cooperation).

Whaites, Alan, Kelly Currah, Haidy Ear-Dupuy, and BrettParris, 2003, “Promoting Growth for Poverty Reduc-tion: The Role of IMF Lending Advice,” Policy Paper(London: World Vision International).

Wilks, Alex, and Fabien Lefrançois, 2002, “Blinding withScience or Encouraging Debate?—How World BankAnalysis Determines PRSP Policies” (Washington:Bretton Woods Project and World Vision).

Winters, L. Alan, 2000, “Trade, Trade Policy and Poverty:What Are the Links?,” CEPR Discussion Paper No.2382 (London: Centre for Economic Policy Re-search).

———, Neil McCulloch, and Andrew McKay, 2004,“Trade Liberalization and Poverty: The Evidence So Far,” Journal of Economic Literature, Vol. 42,pp. 72–115.

110

Bibliography

World Bank and International Monetary Fund, 1999a,“HIPC: Strengthening the Link Between Debt Reliefand Poverty Reduction,” background document, An-nual Meetings of the World Bank and the IMF, Sep-tember (Washington: World Bank and InternationalMonetary Fund).

———, 1999b, “Building Poverty Reduction Strategies inDeveloping Countries,” background document, An-nual Meetings of the World Bank and the IMF, Sep-tember (Washington: World Bank and InternationalMonetary Fund).

———, 1999c, “Poverty Reduction Strategy Papers—Op-erational Issues,” December (Washington: WorldBank and International Monetary Fund).

———, 2000a, The PRSP Sourcebook (Washington:World Bank and International Monetary Fund); alsoavailable via the Internet at http://www/worldbank.org/poverty/strategies/sourcetoc.htm.

———, 2000b, “PRSPs—Progress in Implementation Re-port,” September (Washington: World Bank and Inter-national Monetary Fund).

———, 2001a, “PRSPs—Progress in Implementation Re-port,” April (Washington: World Bank and Interna-tional Monetary Fund).

———, 2001b, “Guidelines for the Joint Staff Assessmentof a PRSP,” April (Washington: World Bank and In-ternational Monetary Fund).

———, 2001c, “IMF–World Bank Collaboration on Pro-gram Conditionality,” July (Washington: World Bankand International Monetary Fund).

———, 2002a, “External Comments and Contributions onthe Joint Bank/Fund Staff Review of the PRSP Ap-proach,” Volume I: Bilateral Agencies and Multilat-eral Institutions (Washington: World Bank and Inter-national Monetary Fund).

———, 2002b, “External Comments and Contributions onthe Joint Bank/Fund Staff Review of the PRSP Ap-proach,” Volume II: Civil Society Organizations andIndividual Contributions (Washington: World Bankand International Monetary Fund).

———, 2002c, “Operationalizing Bank-Fund Collabora-tion in Country Programs and Conditionality,” Guid-

ance Note, April (Washington: World Bank and Inter-national Monetary Fund).

———, 2002d, “Guidelines for JSAs of PRSP ProgressReport,” June (Washington: World Bank and Interna-tional Monetary Fund).

———, 2002e, “Market Access for Development CountryExports—Selected Issues,” September (Washington:World Bank and International Monetary Fund).

———, 2003a, “Achieving the MDGs and Related Out-comes—A Framework for Monitoring Policies andActions,” Paper Prepared by the staffs of the Fundand the World Bank for Consideration by the Com-mittee of the Whole for the Development Commit-tee, EB/CW/DC/03/3, February (Washington: WorldBank and International Monetary Fund).

———, 2003b, “A User’s Guide to Poverty and Social Im-pact Analysis,” (Washington: World Bank PovertyReduction Group and Social Development Depart-ments); available via the Internet at (http://www.worldbank.org/psia).

———, 2003c, “Poverty Reduction Strategy Papers—Progress in Implementation Report,” September(Washington: World Bank and International Mone-tary Fund).

———, 2004a, Global Monitoring Report on Policies andActions to Achieve MDGs and Related Outcomes,March (Washington: World Bank and InternationalMonetary Fund).

———, 2004b, “Strengthening IMF–World Bank Collab-oration on Country Programs and Conditionality—Progress Report,” March, EBM/04/26 (Washington:World Bank and International Monetary Fund).

World Development Movement, 2001, “Policies to Roll-Back the State and Privatise? PRSPs Investigated,” ADebt Policy Report (Washington: World DevelopmentMovement).

Zuckerman, Elaine, and Ashley Garrett, 2003, “DoPoverty Reduction Strategy Papers (PRSPs) AddressGender? A Gender Audit of 2002 PRSPs,” GenderAction; available via the Internet at http://www.gender.action.org.

111

Statement by the Managing DirectorIMF Staff Response

IEO Comments on Management/Staff Responses

Summing Up by the Acting Chair

The Independent Evaluation Office is to be com-mended for its wide-ranging and informative reporton the Poverty Reduction Strategy Paper (PRSP) ap-proach and the Poverty Reduction and Growth Facil-ity (PRGF). The report has generated considerabledebate within the Fund and provides useful materialfor our continuing assessment of effective modes ofengagement with low-income countries through thePRSP approach and the PRGF.

Relatedly, earlier this year, Ms. Krueger formedan interdepartmental Committee on Low-IncomeCountry Work (CLICW) to strengthen the Fund’swork on low-income countries. This committee,chaired by her and consisting of Deputy ManagingDirectors and Department Directors (AFR, FAD,MCD, PDR, RES, WHD), will coordinate furtherwork by staff on these issues, with the aim of prepar-ing concrete proposals for Board consideration at alater date.

I have asked staff to prepare a statement present-ing our views on the IEO report’s recommendationsand indicating how we envisage taking up these rec-ommendations subject to the conclusions of theBoard discussion. While the staff statement outlinessome broad options for modifying the PRSP ap-proach and PRGF, the forthcoming annual Bank-Fund PRSP Progress in Implementation Report willgive specifics of these options. Following theseBoard discussions, staff will come back to theBoard with cost estimates of any new initiatives,possible avenues for savings on other Fund outputs,and the appropriate pace of phasing in of any new initiatives.

I look forward to the Board discussion, whichwill provide an opportunity to explore the optionsfor adapting the Fund’s policies and procedures withrespect to the PRSP approach and PRGF, bearing inmind resource costs.

STATEMENT BY THE MANAGING DIRECTOR ON THE

INDEPENDENT EVALUATION OFFICE REPORT ON THE EVALUATION OF

POVERTY REDUCTION STRATEGY PAPERS AND THE

POVERTY REDUCTION AND GROWTH FACILITY

Executive Board MeetingJuly 21, 2004

115

I. Introduction

1. The Independent Evaluation Office (IEO) re-port offers a trenchant analysis of the Poverty Re-duction Strategy (PRS) approach. The report con-cludes that the PRS approach has had some successin improving country ownership, enhancing partici-pation, and providing better quality strategies.Nonetheless, its achievements so far fall short of itspotential to deliver credible, effective, and country-owned strategies for growth and poverty reduction.Staff broadly support the IEO’s recommendations,some of which were identified in previous assess-ments by the staff. However, staff differ with theIEO on how best to manage the tensions inherent inthe PRSP approach.

2. The IEO report represents a welcome opportu-nity to reflect on the accomplishments and chal-lenges of the PRSP/PRGF approach. It has gener-ated substantial discussion within the Fund staffabout the PRSP as a framework for Fund interactionwith low-income countries. In view of the impor-tance and complexity of the questions raised by theIEO, further reflection on these issues will be neces-sary. Given the joint nature of the PRS approach, theBank and the Fund will need to cooperate closely inmoving the initiative forward. The Bank’s Opera-tions Evaluation Department (OED) has completed aparallel review of the PRSP process and the Bank’srole. The OED report, and the response of Bankmanagement, will be discussed by the Committee onDevelopment Effectiveness on July 19, 2004. Thenext opportunity for the Board to consider related is-sues will be in September 2004 in the context of itsupcoming discussion of the annual Bank-FundPRSP Progress in Implementation Report.

3. One overarching implication of the analysisand various recommendations put forward in theIEO report is that staff resources allocated to low-income work are insufficient for the strategic ap-proach endorsed by the Board. Staff agree that the

ability of the Fund to play a more effective role inlow-income countries depends on its staff beingable to contribute constructively to the analyticaldiscussion and policy debate in the PRSP process.Any resultant increased demands on staff time willthen need to be balanced by more overall resourcesor a reduction in other activities. Alternatively, theFund’s role could be adapted to the resource con-straint. At this stage, the IEO’s recommendationsare too diffuse to provide estimates of their resourceimplications, although their resource costs are likelyto be sizable based on preliminary reactions of thearea departments.

4. The next section summarizes the IEO assess-ment of the PRSP/PRGF approach. Section III givesstaff views on the IEO evaluation and recommenda-tions, and Section IV outlines elements of a strategyfor moving forward. Section V presents some initialreflections on resource implications of the IEO recommendations.

II. A Summary of the IEO Evaluation of thePRSP/PRGF Approach

5. The IEO report comes out strongly in favor ofmaintaining the PRS approach, and observes impor-tant achievements . . .

• Ownership of the PRS has improved over previ-ous approaches.

• Participation in the formulation of PRSPs hasbeen more broadly based than in the past.

• Compared to previous development strategies,PRSPs provide more comprehensive and inte-grated poverty reduction strategies, with alonger-term perspective. They thus provide abetter framework for coordination with donorsand civil society.

. . . but also shortcomings.

STAFF RESPONSE TO THE INDEPENDENT EVALUATION OFFICE REPORT ON THE

EVALUATION OF POVERTY REDUCTION STRATEGY PAPERS AND THE

POVERTY REDUCTION AND GROWTH FACILITY


116

Staff Response

• The perception remains widespread that the newapproach is externally driven and dominated bydocument preparation, especially related to theHIPC Initiative.

• Institutions for policy formulation have not ingeneral been strengthened. Stakeholder andcountry participation in the formulation of themacroeconomic framework has been narrow.

• The quality of PRSPs remains weak in manyareas. For example, most PRSPs lack a strategicvision, especially in the area of macroeconomicand related structural policies. Many PRSPsavoid confronting strategic choices, such as crit-ical structural reforms and prioritization ofspending programs aimed at poverty reduction.

• The PRS approach has not in general deliveredon its promise of improving donor coordination.

• The Fund’s contribution to the PRSP processhas varied considerably across countries, fallingshort of the goals set in the original policy documents.

• The role of the Joint Staff Assessment (JSA) isnot sufficiently clear. Consequently, JSAs havenot always provided effective feedback to theauthorities on weaknesses in the PRS; they havehad little impact on the broader policy debate,and JSAs do not provide a satisfactory basis formaking judgments on lending decisions.

6. The IEO also provides a useful perspective onthe PRGF. It finds that program design has im-proved in many ways . . .

• fiscal targets are more flexible;

• projections do not suffer from a systematic “aidpessimism” bias;

• pro-poor expenditures have increased;

• there is no evidence of an excessive disinflation-ary bias.

. . . but concludes that too little has changed.

• program design has continued to lack an under-standing of micro-macro linkages;

• too few PSIAs have been conducted, and the in-tegration of the results into program design hasbeen too slow;

• alignment of the PRGF with PRSPs thus far hasbeen limited.

7. The IEO report is more suggestive than pre-scriptive in laying out the operational implications ofits recommendations. Nonetheless, a central thrustof the IEO’s recommendations is to clarify and, in

staff’s view, to strengthen the links between the Bret-ton Woods institutions’ (BWIs) lending decisions andthe PRS framework, while at the same time deepen-ing the reach of the PRS framework into domesticpolicymaking processes. In particular:

• Countries should define their own benchmarksfor improving their policymaking process. TheBWIs should then evaluate progress, and suit-ability for continued lending, based on thesebenchmarks.

• The Fund should be more involved in the PRSprocess, including in the public debate, but witha scope limited to its areas of comparative advantage.

• The IMF could restructure its conditionality tofit within a broader “partnership” approach tomonitoring and assessing progress in imple-menting the PRS.

III. Views on the IEO Evaluation andRecommendations

8. The IEO’s diagnosis of problems with the im-plementation of the PRSP approach overlaps sub-stantially with that of previous Board and staff as-sessments.1 The PRSP is an instrument charged withmultiple objectives, which result in tensions—forexample, long-term ambition versus immediate bud-get constraints; comprehensiveness in addressing thedifferent dimensions of poverty versus focus and pri-oritization; and meeting the expectations or require-ments of the international community versus countryownership.

9. The Fund shares the broad objectives underly-ing the IEO’s recommendations, including:

• Higher quality outputs of the PRS process, par-ticularly in terms of priorities, costing, and pol-icy analysis, and a more constructive Fund rolein its areas of competency;

117

1“Poverty Reduction Strategy Papers – Progress in Implemen-tation” (SM/02/250 and SM/03/279), “Review of the PovertyReduction Strategy Paper Approach—Early Experience with In-terim PRSPs and Full PRSPs” (SM/02/54, 3/27/02); “Aligningthe Poverty Reduction and Growth Facility (PRGF) and thePoverty Reduction Strategy Paper (PRSP) Approach: Issues andOptions” (SM/03/94); “Role of the Fund in Low-Income Mem-ber Countries Over the Medium Term—Issues Paper for Discus-sion” (SM/03/257); “Review of the Poverty Reduction andGrowth Facility—Issues and Options” (SM/03/51, 3/15/02);“The Fund’s Support of Low-Income Member Countries—Considerations on Instruments and Financing”(SM/04/53,2/24/2004); and “Fund Assistance for Countries Facing Exoge-nous Shocks” (SM/03/288, 8/11/2003).

STAFF RESPONSE

• Strengthened country ownership and greater inte-gration of the PRS process with domestic politi-cal and policy processes, particularly the budget;

• Reduced “mismatch of expectations” of variousstakeholders (governments, civil society, IFIs,donors) of the PRS approach;

• Better alignment of the PRGF with the PRSP,and more clarity on what happens if the PRSP istoo weak or unrealistically ambitious to serve asa framework for a PRGF;

• Better fit of the focus and mix of policies inPRGF-supported programs to country circum-stances, based on a deeper understanding ofmicro-macro linkages.

10. An important consideration for any evalua-tion of the PRS approach relates to how quickly re-sults from this new approach can be reasonably ex-pected to emerge. The PRS approach attempts toimprove the entire framework for, and direction of,policy formulation in low-income countries. This ap-proach is only five years old. Any assessment, andrecommendations, should recognize the variety ofcountry experiences, including important successes,the evolving nature of the initiative, the multiplicityof objectives, and the novelty of the approach in along-established multipartite international setting.The implication is that substantial scope exists forbetter implementation of the current approach, withappropriate mid-course corrections.

11. In light of the similarity of diagnosis and ob-jectives, Fund staff supports many of the IEO’s sug-gestions. However, in some cases, the objectives areshared, but the staff questions the feasibility of someIEO recommendations. After a discussion of thecommon ground, the implications of key differencesin perspectives are explored. The Appendix providesspecific staff reactions to each IEO recommenda-tions. Section IV presents our approach designed toachieve key objectives.

A. Areas of Broad Agreement

12. With respect to the design and implementa-tion of the PRS approach, staff agree that countriesshould themselves decide how policy formulation,implementation, and monitoring processes will beconducted and built up over time. Countries shouldalso determine the choice of output of the PRSprocess, in terms of documents and their periodicity,relying as much as possible on domestic institutionalarrangements and reporting vehicles (IEO Recom-mendation 1). The key issue is how donors shouldreact in cases where they believe that the countrycould aim higher, or commitment is insufficient.More candid and graduated assessments by Bank

and Fund staff and donors are key for appropriatelybalancing the need for ownership against the expec-tations of the international community.

13. With respect to the Fund’s role, the staff hasbeen attempting to clarify further the intensity andscope of the IMF’s role in the PRS approach, alongthe lines of the IEO suggestions and earlier Boardguidance. In particular, staff recognizes the need togive more emphasis to Fund staff involvement in thedomestic policy debate over macroeconomic policyand to open the rationale for IMF policy recommen-dations to broader scrutiny (Recommendation 4).The staff is also trying to explore in a more system-atic way the linkages between macroeconomic poli-cies and poverty reduction.

14. One area where progress has been made is inpoverty and social impact analysis (PSIA) of pro-posed policy changes. The Fund’s FY2004/05 bud-get includes dedicated additional budgetary re-sources to broaden and deepen the use of PSIA andto eventually mainstream PSIAs into the design ofPRGF-supported programs. Nonetheless, careshould be taken to avoid building unrealistic expec-tations as to what the Fund can deliver in PSIA in thenear future, in part because of resource limitationsdiscussed below.

15. Staff agree that the Fund should focus onareas in which it has a comparative advantage, thatthe Fund can play an important but only supportingrole in developing alternative aid and policy scenar-ios, and with the recommendation to avoid undue“on-off” signals from the Fund for donor financing(Recommendations 5 and 6). Strengthening theFund’s accountability for its own commitments inthe context of the PRS process is also an attractivegoal, particularly in ensuring the Fund’s contribu-tions are consistent with needs identified in countryPRSPs.

B. Differences with IEO Recommendations

16. The most important difference is over manag-ing the tensions between country ownership of thepolicies and programs in the PRSPs and the need ofdonors and IFIs for minimum standards. In particu-lar, the donor community wants to use the PRSP asthe basis for selectivity of resource allocation.

17. The IEO recommends that countries set ex-plicit criteria for judging progress towards key inter-mediate objectives related to the domestic policy for-mulation, implementation, and monitoring process,and that BWI staff provide candid assessments ofthose benchmarks (Recommendations 1 and 2).Donor decisions on the volume of resources pro-vided would then be linked to the progress countriesare making under the approach. The IEO contendsthat the PRSP approach already involves requiring

118

Staff Response

the countries to adopt various changes in their policyprocesses in return for financing and debt relief, inparticular from the Fund and the World Bank. It alsosuggests that the Fund’s criteria for making such de-cisions are unclear, and in particular are obscured bythe focus on documentation requirements. In theconcluding section on the role of the Fund, the IEOproposes that the IMF could restructure its condi-tionality to fit within a broader “partnership” ap-proach to monitoring and assessing progress in im-plementing the PRS. The precise operationalmodalities are not discussed.

18. Staff agrees that the quality of the PRS ap-proach should be a critical factor in donors’ re-source allocation decisions, in order to enhance aideffectiveness. Such selectivity will appropriately en-courage stronger PRSPs. In our view, however, theIEO recommendations imply excessive Fund in-volvement in assessing the country’s decision-mak-ing processes. The IEO suggests that the BWIscould assess achievement of country-definedprocess benchmarks. Combined with the suggestedtighter link to financing, this could be seen as a kindof process conditionality, which could underminethe legitimacy of these domestic institutions and de-cision-making processes. The IEO report itself un-derscores the costs associated with the perceptionthat the PRSP process is BWI-driven. Moreover,good policymaking processes are no guarantee ofgood policies.

19. In staff ’s view, the BWIs should continue toemphasize the country ownership of the PRSprocess. It will take time and effort for domesticallyanchored processes to result in high-quality, priori-tized PRSPs, and the BWIs can and should con-tribute to this institution building. The Fund willcontinue to work with the World Bank to developand promote best practices for improving domesticpolicy processes, but the Fund should stop short ofmeasuring countries against specific benchmarks inthis area. Instead, the Fund’s role should focus on as-sessing the quality of macroeconomic frameworks,and in providing advice and support for the develop-ment of sound frameworks.

IV. Options for Moving Forward

20. A promising strategy for creating a higher-quality and more fully country-owned PRS approachwould be to make the Fund’s contribution more substantive and less procedural. Currently, as the IEO notes, too much emphasis is placed on Fund assessment of PRS documents such as PRSPs and an-nual performance reports on PRSP implementation(APRs). In our view the direction forward is to lightenreporting requirements, while increasing the scope forthe Fund to make substantive contributions. The em-

phasis should be on providing more effective adviceon the formulation of the macroeconomic strategiesunderpinning PRSPs, particularly with respect to theirbudget implications, and on providing frank opinionson such strategies as the PRS process unfolds.

21. The links from the PRS to Fund financing, sig-naling to donors, and HIPC debt relief could bemade more flexible, in order to provide more policyspace and development of domestic policy processes.The incentives for the country to produce a goodPRSP must come from its benefits in terms of betterpolicy formulation and from donors, including theFund and the World Bank, respecting the views out-lined in a well-prioritized, domestically-drivenPRSP and (in the case of development assistance) al-locating more resources to countries that have high-quality PRSPs.2 This could increase ownership andencourage integration of the PRSP with the budgetand other domestic processes while still providing fi-nancial incentives.

A. Improving the PRS Approach

22. As noted above, staff agrees with the IEO thatthe staffs’ assessments of the countries’ PRSPs needto be more candid and more graduated. The currentoperational requirement is that the JSA conclude byfinding either that the PRSP does or does not providea credible basis for BWI lending. This can hamstringthe potential candor of the assessment and puts allPRSPs into one of two categories (and in practiceonly one). PRSPs are seen by staff as becominglonger, more detailed, and more resource intensive,while providing little of apparent value to staff, theauthorities, or donors.

23. A reformulation of the instrument used byBWIs to assess PRSPs is needed. As noted before,staffs intend to come back to the Board with morespecific recommendations in the context of thePRSP annual implementation report, but early Boardguidance would be appreciated. Bank and Fundstaffs are currently considering three elements ofsuch a reformulation.

• What is the purpose of the assessment: to shapeFund lending decisions; to provide candid feed-back; and/or to help coordinate with the Bank(and donors)?

• Who is making the assessment? If the assess-ment is meant to provide a basis for the BWIlending decisions, then presumably Board en-dorsement would be necessary. If it is meant to

119

2This idea was at the origin of the PRSC, which was originallythought of in part as a vehicle to reward countries with suffi-ciently good PRSPs with an appropriate reduction in traditionalred tape and conditionality.

STAFF RESPONSE

give candid feedback to the country or todonors, then would it be more appropriatelytreated in a similar manner to assessment lettersand sent to the Board for information? If coordi-nation of donors is an important objective, arethere continued benefits to a joint assessment?Could other donors be involved in the assess-ment? What is the role of the low-income coun-try itself?

• What is being assessed? Should the assessmentevaluate the process by which the document wasformulated, the policies in the document, the au-thorities commitment to it, and/or the donors’commitment to financing the strategy?

24. Besides modifying the JSA, the Fund couldparticipate more fully in the PRSP process throughproviding policy advice and, where appropriate, par-ticipation in the public discussion leading to the for-mulation of PRSPs, particularly encouraging and par-ticipating in discussions of the macroeconomicoptions and the options for structural reforms that aremacro relevant. There are important roles to play bothas participants in the public debate, and in helping thegovernment make its case for its macroeconomic pol-icy choices. This would include bringing the results ofany relevant poverty and social impact analysis(PSIA) into the discussions of overall macroeconomicframeworks and related policies. In its ultimate as-sessment, however, the Fund should make clear itsviews on macroeconomic frameworks ultimately in-cluded in PRSPs and be prepared to provide convinc-ing arguments for any proposed revisions.

25. In this context the Fund’s efforts to build ca-pacity and strengthen institutions through technicalassistance will be particularly important. Further ef-forts will be needed to provide appropriate advice andtechnical assistance where required to strengthen bud-get policies and processes aimed at promoting sus-tainable growth and poverty reduction.

26. One underemphasized issue has been the roleof building capacity through technical assistance(TA), which is central to the vision of the Fund as aservice provider in the context of the PRS approach.In its recent review of TA, the Board pointed toprogress in pursuing strategic focus and enhancingeffectiveness.3 Nonetheless, further work needs to bedone to better coordinate with other TA providersand integrate Fund TA into the broader agenda of theFund’s relationship with low-income countries. Inthis regard, AFRITACs (Africa Regional TechnicalAssistance Centers) and the forthcoming MiddleEast Technical Assistance Center can play an impor-tant role. Area departments will also need to take a

more strategic view of the technical assistance needsof member countries in close consultation withcountry authorities.

B. Implications for Fund-Supported Programs

27. The Fund Board previously considered issuesin working towards a better alignment of PRGF-sup-ported programs with country PRSPs.4 The Boardstressed that the process of alignment will need to beimplemented pragmatically and flexibly, taking dueaccount of country-specific circumstances and ca-pacity constraints, as well as the need to ensure thatPRGF-supported programs continue to be strong anddesigned to help countries accelerate growth and thepace of progress towards the Millennium Develop-ment Goals.5

28. In the staff’s view, a key element in improvingthe Fund’s contribution to the PRS process is to in-troduce flexibility in the link between PRGF-sup-ported programs and the PRSPs, based on country-specific circumstances. PRGF-supported programswould continue to be drawn from the macroeco-nomic frameworks in PRSPs to the extent possible.Where a PRSP provides a good operational roadmap, the PRGF should be tightly linked to its frame-work. Program and mission cycles should be alignedwith the PRSP/budget cycles and other domesticprocesses as much as possible.

29. However, some countries may not be able toproduce operationally viable PRSPs for some time.A PRSP would still be required for Fund (and Bank)financial support, but the emphasis in these cases,for the Fund, would shift from artificially insistingon immediate tight alignment between the PRGF-supported program’s macroeconomic frameworkand in the PRSP. Rather, Fund staff need to workwith countries to strengthen the macroeconomicframeworks in their PRSPs so as to move towardalignment over time. Areas that required work couldbe clearly highlighted in the assessment of the PRSP.The Fund would still seek to apply the PRSP princi-ples to the Fund’s role. This would include (i) seek-ing to open up the policy debate on a few key prior-ity issues (for example, through analytical inputs andprioritization of PSIA that would be designed to explore various policy options); and (ii) contribu-tions to capacity development that follow country-driven priorities and understandings on what a rea-

120

3“Review of Technical Assistance” (SM/04/41, 2/17/04).

4“Aligning the Poverty Reduction and Growth Facility (PRGF)and the Poverty Reduction Strategy Paper (PRSP) Approach: Is-sues and Options” (SM/03/94).

5“Concluding Remarks by the Acting Chair, Aligning thePoverty Reduction and Growth Facility (PRGF) and the PovertyReduction Strategy Paper (PRSP) Approach—Issues and Op-tions” (BUFF/03/55).

Staff Response

sonable road map toward a operationally orientedPRSP would be.

30. In such cases, the PRGF-supported programswould provide additional coherence, precision, andclarity to the authorities’ policy intentions, and staffwould need to affirm that the authorities’ commitmentto the resulting framework is nonetheless adequate.Rather than derailing the incipient participatoryprocesses by integrating them into PRGF-supportedprogram discussions, in this manner the Fund couldcontinue to seek to approve arrangements with suchcountries on the basis of adequate PRGF-supportedprograms that would still be linked to PRSPs, albeitwith clearly explained modifications.

31. The introduction of such flexibility in the linkbetween the PRSP and PRGF is seen as a way tostrengthen the Fund’s involvement in the formulationand implementation of macroeconomic policies in thePRS approach. Making the Fund’s concerns with thePRSP macroeconomic framework transparent in theassessment will give the authorities and their develop-ment partners a more precise and nuanced assessmentof the country’s macroeconomic challenges. It willalso underscore the evolving nature of the macroeco-nomic framework in the PRGF-supported programs,which are inherently dependent on shorter-term con-siderations such as availability of donor financing. Ofcourse risks will also need to be addressed. In particu-lar, the existence of different macroeconomic frame-works in the PRSP and the PRGF-supported programcould be seen as evidence of Fund intrusion in domes-tic policy setting or of the Fund undermining of thePRSP process. Thus, the government should clearlydemonstrate its ownership of the program, and theFund its commitment to the PRS process, by explain-ing clearly their position to other development part-ners and domestic stakeholders.

32. As part of the effort to strengthen its involve-ment in the PRS process, the Fund will also continueto seek ways in which to improve PRGF-supportedprogram design. More can be done, and the Fundwill take due consideration of the IEO’s report in itsforthcoming work on the design of PRGF-supportedprograms. This review of program design will seekto provide recommendations on how to increasethese programs’ effectiveness in enhancing growthand reducing poverty. In this regard, it is importantthat the Fund be sufficiently selective in its alloca-tion of PRGF resources. Areas where the design ofPRGF-supported programs might be refined include:

• ways in which programs, along with TA, canhelp build the institutions necessary to underpin

sustainable growth, particularly in the monetaryand fiscal areas most directly related to theFund’s comparative advantage;

• appropriate strategies to facilitate developmentof the private sector, including monetary frame-works for facilitating higher private sector creditgrowth;

• means by which macroeconomic frameworkscan best accommodate higher aid inflows.

V. Resource Implications

33. In considering the IEO recommendations, re-source constraints must be taken into account. TheFund’s participation in the PRS approach can, de-pending on its nature, require substantial Fund re-sources, including staff time and travel, in additionto that necessary to carry out the Fund’s other re-sponsibilities. These demands have a tendency tocrowd out critical analytic and substantive work.

34. Many of the IEO recommendations would in-crease Fund responsibilities considerably.6 The IEOsuggests that staff become more systematically anddirectly involved in the PRS process, including inthe public debate surrounding the PRSP and APR. Incountries where policy formulation procedures areweak, considerable additional Fund resources wouldbe needed to deepen the policy dialogue. In suchcases, the IEO recommendations would represent amajor additional demand on staff time. Thus, inter-nal Fund discussion is needed on how to balancethese needs against other priorities. The suggestionsin Section IV are formulated in part with a view toreducing the procedural burden on staff of the PRSprocess.

35. The staff resource implications of deepersubstantive involvement in the PRS process wouldneed to be considered carefully. In the context of adesire to maintain closer and more informal con-tacts with a large number of stakeholders, onequestion is whether, as the IEO has suggested, con-sideration should be given to adjusting somewhatthe mix of headquarters and resident staff. In par-ticular, the role of the resident representatives re-quires careful consideration.

121

6The IEO considers that none of its proposals are outside the original framework of the PRSP approach as reflected in theBoard documents.

STAFF RESPONSE

122

STAFF VIEWS ON THE SPECIFIC IEO RECOMMENDATIONS

A. Recommendations Relating to the Design and Implementation of the PRS Approach

IEO Recommendation Staff Response

1. Introduce greater flexibility in the implementation of the PRS Staff agrees that countries should be encouraged to customize approach to fit better the needs of countries at different stages of the PRS approach to country circumstances. This notion was the process and with different capacities and political and admin- embedded in the PRSP approach at its inception, and was istrative systems. Countries need to be put even more firmly in reiterated in the 2002 review of the PRSP, which stressed that the driver’s seat by determining themselves: the approach should reinforce—not undermine—existing

national institutions, processes, and governance systems. That (a) How the policy formulation, implementation, and monitoring this has not happened to the extent that might be wished may

processes will be conducted and built up over time, and with have more to do with broader questions of ownership than with what rules of the game (e.g., for opening up the policy pro- a failure to articulate this objective. As staff argues in the text cess to previously excluded groups or strengthening budget- and with respect to Recommendation 2 below, other aspects of ary processes). Progress would be monitored against an the IEO’s recommendations risk further reducing the extent to explicit set of country-determined intermediate benchmarks; which the process is country-driven.

(b) What the output of these processes will be in terms of Moreover, leaving the criteria for the documents entirely to documents (e.g., PRSP, Progress Reports, etc.) and on what countries raises a number of issues. The report does not clarify periodicity they will be prepared, relying as much as possible how donors (including the BWIs) should react in cases where on domestic institutional arrangements and reporting vehicles. they believe that the country could aim higher, or where they IMF process requirements (e.g., linking reviews under the view the pace of progress chosen as indicative of a lack of PRGF to completion of specific PRSP documents) should be commitment on the part of the authorities to objectives donors minimized and oriented around domestic processes. can support.

2. Shift the emphasis of the initiative from the production of docu- Staff agrees that the PRS process is too document-driven and ments to the development of sound domestic policy formulation that PRSPs should be more tightly integrated with domestic and implementation processes. This means: policy process and institutions. It is also desirable that the

quality of the PRS approach be a critical factor in donors’ (a) Build in greater results-orientation. Countries should be encouraged resource allocation decisions.

to establish (with help from BWIs where needed) substantive criteria for judging progress toward key intermediate objectives However, the thrust of this recommendation is to attempt to such as developing (i) an operational road map that provides use the financial leverage of the BWIs to improve domestic strategic guidance for setting priorities and resolving trade-offs; policy processes. This effort is unlikely to succeed. As the IEO and (ii) effective institutional arrangements for formulating, report itself underscores, one reason the PRS process remains implementing, monitoring, and updating this road map, with a separate from domestic modalities is the perception that it is firm link to budget processes. The choice of the criteria/bench- BWI-driven. It will take time and effort for domestically-marks judged to be most important would likely vary by country, anchored processes to result in high-quality, prioritized PRSPs.but in some areas broad guidance could be developed by BWI The IEO recommendations risk weakening domestic ownership staff or others, upon which countries could draw as a starting (if the BWIs ask for minimum standards in the PRSPs) or point. However, the types of benchmarks chosen by each country weakening the programs supported by the BWIs (if the BWIs do will depend on the improvements that their PRS process not set minimum standards for lending).identifies as a high priority.

The recommendation also raises a variety of issues that reflect (b) Shift in emphasis of the incentives structure faced by countries these tensions. It is not clear who would be doing the proposed

from procedural aspects and production of documents to monitoring of country-chosen progress benchmarks. If countries achieving substantive changes in domestic processes and do the monitoring themselves, donors may seek to establish policies objectively measured as described in (a) above. The their own benchmarks to ensure the country is in fact doing new set of incentives would include: (i) Countries should present what it says.their intentions and objectives, along with the benchmarks selected to monitor progress in a manner open to public scrutiny; (ii) IMF (and World Bank) staff would be responsible for providing clear and candid assessments of the progress made by each country in implementing the PRS approach, both in relation to the goals set by the country itself and against initiative-wide benchmarks; (iii) IMF (and World Bank) staff would help countries identify key constraints in making progress towards PRS object-ives and support efforts to ameliorate them; (iv) Ideally, donor decisions on the volume of resources provided should be linked to the progress countries are making under the approach. To facilitate this, IMF assessments in its area of expertise need to provide as clear and candid a signal as possible. The criteria guiding the IMF’s own lending decisions under the PRS approach could also be improved in this regard.

Staff Response

123


3. Clarify the purpose of the JSA and redefine the vehicle accordingly This recommendation includes many shared objectives for the The JSA is a useful concept whose potential has not been realized. JSA and indeed for Fund analysis more broadly, such as candor,The IEO recommends making the following changes: reporting the views of other stakeholders, and pointing out

areas of weakness in the PRSP and how to address them.(a) JSAs should focus on the adequacy of domestic policy choices

and the quality of domestic processes as well as actual progress By expanding the reach of the JSA to include “the quality of towards intermediate objectives, and less on the quality of the domestic processes,” the IEO would significantly increase the PRSP as a document. resource costs of the JSA and its overlap with other documents,

such as Article IV staff reports. At the same time, the attempt to (b) To foster clear and candid assessments, we propose: (i) making tighten the link between BWI assessment and these domestic

explicit the criteria and benchmarks used by staff to form their processes is likely to be counterproductive, as described in judgments; (ii) reporting on the views of third parties (especially staff ’s response to Recommendation 2.local stakeholders and donors) when available, and discussingdifferences of view; and (iii) eliminating the need for JSAs to Staff agrees with the suggestion to eliminate the binary conclusion reach a binary (yes or no) conclusion as to the adequacy of the in the JSA. Section IV of the staff response goes further andPRS as a basis for BWI concessional lending. JSAs should aim to discusses elements of a possible reformulation of the instrument provide a graduated assessment of the strength of the PRS and used by BWIs to assess PRSPs, in order to increase the scoperelated processes as well as of the quality of policies. for candor and textured assessments, allow for more effective

feedback, and counter the perception that the PRSP and the(c) The effectiveness of the above recommendations could be APR are mainly instruments of BWI conditionality.

enhanced if JSAs were produced on an independent schedule (e.g., once a year), rather than being linked to a specific PRSP document.

(d) The JSAs would be more effective if, in addition to flagging weak-nesses in the PRS, they indicated clearly what are the main obs-tacles to overcome; what the IMF proposes to do to help address them in its areas of responsibility; and what needs remain un-addressed, especially in the area of capacity building.

B. Recommendations Relating to the Fund’s Role in the PRS Approach


4. Clarify what the PRS approach implies for the IMF’s own operations (a) PRS process. Staff shares the objectives of clarifying the and strengthen the implementation of the agreed role. expectation about the IMF role under the PRS approach and

finding ways to improve its effectiveness in meeting these (a) IMF engagement in the PRS process: (i) More emphasis should be expectations.The Fund should participate in the PRSP process

given to IMF activities that help to better inform broad-based through providing policy advice and, where appropriate,policy discussions in its areas of competence. Guidelines to staff participation in the public discussion leading to the formulation need to be clarified so as to encourage more active inputs to such of PRSPs, particularly encouraging and participating in discussions, including analyzing alternative policy options and discussions of the macroeconomic options and the options for trade-offs; (ii) Rather than establish uniform “standards” for the structural reforms that are macro relevant. The Fund should IMF’s role, expectations should be tailored to country-specific provide advice and technical assistance where required to circumstances, including the government’s wishes. These country- strengthen budget policies and processes aimed at promoting specific “rules of the game” should be made public and could sustainable growth and poverty reduction. Resource describe how IMF staff expected to participate in the broader implications of further initiatives to strengthen Fund policy debate, including what supporting analytical work they will involvement in the PRS process would, however, have to be undertake. This is also likely to require a combination of more carefully considered.“stand alone” missions, set apart from program negotiations, and enhancing the role of resident representatives; (iii) There should be more systematic explorations of country-specific macro-micro linkages; (iv) Article IV surveillance reports should be used system-atically as a vehicle to convey the IMF’s own thinking and analysis on key issues; (v) Assess systematically, as part of broader IMF sur-veillance activities, obstacles to the achievement of PRSP objectives originating in trade and subsidy policies of main trading partners.

(b) PRGF-related activities: (i) The rationale for IMF policy recom- (b) PRGF-related activities. Staff agrees with these objectives. Staffmendations and program design should be subjected to broader considers that the Fund should make clear its views on macro-scrutiny and debate; (ii) Clarify the approach to be taken by the economic frameworks ultimately included in PRSPs and beIMF in those cases where the PRS approach has added some value prepared to provide convincing arguments for any proposed but has not yet produced an operational road map or the neces- revisions in a PRGF-supported program.sary institutional framework for implementation; (iii) Clarify whatthe BWIs are trying to achieve jointly through the streamlining of conditionality and how this fits with stronger domestic ownership;establish a system for the monitoring of aggregate Bank-Fund conditionality at the country level.

STAFF RESPONSE

124


(c) Streamline IMF documentation and Board scrutiny of PRS documents. (c) Staff agrees with the need to streamline documentation and theprocedural burden of the PRS more broadly. Indeed, thisrepresents an important disadvantage of those IEO recom-mendations that would seem to move in the other direction(increasing the importance and areas of coverage of the JSA,for example).

5. Strengthen prioritization and accountability on what the IMF itself is Again, staff agrees with the objectives of prioritization and supposed to deliver within the broader partnership framework, built accountability, and takes note of the IEO’s suggestions.around the priorities emerging from the PRS process, and ensureresources match commitments. Prioritization and coordination could However, while the IEO acknowledges the need to prioritize,be improved as follows: many components of the IEO’s recommendations (e.g., greater

staff role in “framing the policy debate” and more exploration of(a) Generate, as part of the PRS process, specific priority actions for macro-micro linkages) would require substantially more staff

the IMF to assist the country concerned to reach its objectives, resources, while suggestions for reducing activities are few.including for analytical inputs and long-term capacity building. Ifthe IMF is not well-suited to produce a particular input (e.g., some Section IV of the response provides a broad outline of how staff PSIA), but this input is judged critical for program design, a clear sees the Fund’s role in the PRS approach, emphasizing identification of who has agreed to do what and by when would substantive contributions to the formulation and debate of strengthen incentives. Such key deliverables would be agreed with macroeconomic frameworks in PRSPs. The Fund has also been the government and made public. These priority actions would moving towards more systematic explorations of micro-macro then be the basis for evaluations of IMF performance, both at the linkages through the use of poverty and social impact analysis of country level and in aggregate. proposed policy changes. However, care should be taken to

avoid building excessive expectations as to what the Fund can (b) The IMF’s own budget decisions on allocation of administrative realistically deliver in this area in the near future.

resources should be geared to these priorities. Linking realistic resource allocation decisions to a more explicit set of priorities The Fund is accountable to its members and reports to them,should help improve decision making and make clearer to all what through the Executive Board, on a regular basis. Strengthening the IMF has committed to, and what it has not. the Fund’s accountability for its own commitments in the context

of the PRS process is desirable; putting it into practice is a chal-(c) Experiment with broader “external reviews” of the PRS/PRGF lenge. For example, a large part of the Fund’s deliverables would

process, monitoring in particular the performance of donors and be the provision of technical assistance and policy advice, but IFIs in providing support, and not just performance of the national measuring the outcome of these activities is difficult at best. At aauthorities. minimum, it is important to work to ensure that the Fund’s con-

tributions are consistent with needs identified in country PRSPs.

Staff continues to welcome external reviews of all aspects of thePRS/PRGF process.

6. The IMF should encourage a strengthening of the framework for This recommendation is consistent with staff ’s views on the establishing the external resources envelope as part of the PRS appropriate substantive role of the Fund. The Fund should focus approach. The country itself, not the IMF or World Bank, should on areas in which it has a comparative advantage. The Fund can eventually play the central role in elaborating macro frameworks play an important but only supporting role in developing and catalyzing donor support. The IMF role would be to provide alternative aid and policy scenarios. Staff should aid country debt and macroeconomic sustainability assessments and judg- authorities wherever possible in elaborating alternative ments on the policy framework, but it would not be responsible macroeconomic scenarios.for the “normative” judgment on appropriate aid levels over the medium term. The tension between “ambition” and “realism” in Staff also agrees that the Fund should avoid giving excessive “on-determining the external resource envelope can perhaps be off” signals for donor financing. The Fund is working with low-handled by presentation of alternative projections. The IMF should income countries and the entire development community to provide increased analytical support for such approaches when improve donor coordination and provide sufficiently calibrated requested. But the choice to prepare alternative projections signals. The Board will have an opportunity after the Annual should remain with the country and not be a uniform requirement. Meetings, for example, to consider options for signaling Fund

assessments of low-income country policies.Improving aid predictability is a wider problem that the IMF cannot resolve on its own. The challenge is how to reconcile PRSP countries’ concern for aid predictability with donors’ concern for aid effective-ness, which implies some performance-based selectivity. For the IMF,this requires finding a way, perhaps through a strengthened JSA, to provide signals to the donor community on macroeconomic perform-ance that are sufficiently calibrated, and take account of the longer-term framework of donor involvement, to be a useful input into selectivity decisions without providing excessive “on-off” signals for financing.

The comprehensive nature of the management/staff responses is welcome. They provide a basis fora substantive debate on the major issues raised in theevaluation. The staff has noted a number of areaswhere it broadly agrees with the messages and rec-ommendations contained in the evaluation report,but also areas of disagreement. Since the evaluationand the management/staff response inevitably covera lot of ground it might be useful to highlight a fewkey questions of particular interest:

• Whether and how to modify the design of thePRS approach to give more emphasis to the in-termediate objective of improving domesticprocesses for policy formulation, implementa-tion and monitoring, in a way and at a pace thatbest fits each country’s circumstances.

• Whether and how to ensure that incentives facedby participating countries—especially those re-sulting from BWI procedural requirements, as-sessment procedures, and provision of assis-tance—are aligned with this objective and aretransparent.

• How the BWIs should assess PRSPs and the na-ture of the instrument they should use. (In thiscontext, we think the questions posed by thestaff in para. 23 are the right ones.)

• How to clarify or adjust expectations about theextent and nature of the IMF’s involvement inthe PRS process, and ensure that commitmentsare consistent with available resources.

• How the IMF’s “way of doing business,” includ-ing in the context of the PRGF, needs to be mod-ified in support of the PRS approach, both in rel-atively mature cases and in those “difficult”cases where the PRSP is unlikely to provide aneffective operational road map for the foresee-able future.

To facilitate this discussion, our comments willfocus on a few issues where we see the messages

emerging from the evaluation as somewhat differ-ent than those emphasized by the staff and will re-spond to several concerns raised about our recom-mendations. Paragraph references are to the staffcomments.

The staff observes (para. 8) that the IEO diagnosisof problems with implementation of the PRSP ap-proach overlaps substantially with those of earlierinternal reviews. In our view, the evaluation raisestwo issues that are more fundamental than thoseraised in previous assessments and that will not beresolved merely by allowing more time for the ap-proach to have an impact. First, some significant de-sign problems have emerged as the approach hasbeen implemented. While some of the tensions thatgave rise to these problems (e.g., between countryownership and BWI/donor selectivity) have been ac-knowledged previously, how best to deal with themhas not been adequately addressed. Second, the IMF“way of doing business” in low-income countrieshas not adapted sufficiently to the implications of thePRS approach—an issue that goes beyond resourceavailability.

We agree, however, that the issue of staff re-sources is a critical one. What the IMF is expectedto contribute to the PRS approach needs to be tai-lored to fit the resources that can reasonably be ex-pected to be available. Contrary to the staff’s obser-vation (para. 34), none of our recommendationsinvolve an expansion of the IMF’s “responsibilities”beyond what was indicated in the original policy pa-pers establishing the PRSP and the PRGF. The dis-cussion in [the section “What Was Expected of theIMF Under the New Initiative?”] of the evaluationreport illustrates how bold were the original expec-tations on the IMF’s role. If these expectations arenow judged to have been too ambitious, it would bebetter to clarify that role explicitly. But whether a“larger” or a “smaller” role for the IMF is expectedin low-income countries in the future, it should notbe a “business as usual” role; if the PRS approach isto be the key framework for IMF involvement in

INDEPENDENT EVALUATION OFFICE COMMENTS ON

MANAGEMENT/STAFF RESPONSES TO THE EVALUATION OF POVERTY REDUCTION

STRATEGY PAPERS AND THE POVERTY REDUCTION AND GROWTH FACILITY


125

INDEPENDENT EVALUATION OFFICE COMMENTS

low-income countries, IMF activities will need toadapt accordingly.

In this context, the most effective approach tomatching expectations with resources is to introducegreater scope for customization of the IMF’s role tocountry needs, building on the twin principles of pri-oritization and partnership. Prioritization impliesthat the key deliverables from the IMF (e.g., for ana-lytical work to strengthen the broader policy debateor assistance with capacity building) should them-selves be derived—to the extent possible—from thecountry-driven PRS strategies. This would then feedinto more transparent budgetary choices within theIMF and realistic indications of what can and cannotbe delivered. The partnership element means thatstrong coordination with other donors is needed,built around the country-driven strategy and priori-ties, and that “stand alone” approaches should be re-sisted, even if they appear administratively easier inthe short run.

The staff agrees that managing the tension be-tween ownership and selectivity is a central one forthe design of the PRS approach but disagrees with theapproach proposed in the evaluation report (paras.17–19). Clearly, different solutions to this problemare possible, but we would like to emphasize twopoints. First, the solution proposed in the report isbased on (i) greater scope for, and openness to, coun-try-driven choices on the road map for implementingthe core principles of the PRS approach; (ii) trans-parency about the choices (so other stakeholders, in-cluding civil society, can contribute and indicatewhere they disagree) along with clear country-drivenbenchmarks for monitoring progress; (iii) candid as-sessments by the BWIs of the country-driven choicesand progress made; and (iv) transparent BWI anddonor decisions on financing selectivity, drawinginter alia on these assessments. We do not propose anapproach based on universal minimum standards, forreasons given in the report. Second, we are not sug-gesting that the IMF (or World Bank) artificially forcethe pace of reforms of domestic processes throughconditionality; on the contrary, the thrust of our rec-ommendations is to allow greater scope for country

diversity, recognizing the wide divergence in startingconditions and political structures. While we agreethat our proposal could be seen as a kind of “processconditionality” (para. 18), this is inherent to thewhole approach of requiring countries to produce aPRSP. Our proposal has the merit of allowing greatercustomization of the process to country needs and cir-cumstances. While everyone recognizes that countryownership is critical, a candid discussion is needed ofhow best the design of the approach can manage op-erationally the tensions between ownership, BWI as-sessments, and selectivity in financing decisions. Asillustrated in the report, such judgments are obviouslymade in practice, but in a manner that is less transpar-ent than under our proposal, and that gives too muchweight to meeting BWI procedural requirementsrather than more fundamental improvements in do-mestic policy processes.

We agree that a key challenge for the IMF is howto apply PRS principles in “difficult” cases, includ-ing those where there is not yet an operationally vi-able PRSP. The particular steps mentioned by thestaff in para. 29—opening up the policy debate on afew key priority issues, including through PSIA de-signed to explore various policy options, and con-tributing to capacity development following country-driven priorities—follow the thrust of ourrecommendations and would be very helpful. How-ever, it is important that PRSP and PRGF-related ac-tivities not be seen as proceeding on separate tracks,as the staff’s proposal in para. 30 seems to imply. Inparticular, we do not see how efforts to better inte-grate the two sets of activities would risk “derailingthe incipient participatory process.” Quite the con-trary, the aim should be to seek opportunities to usethe PRS principles, including through informing abroader policy debate, to ease political economyconstraints that have made progress so difficult insuch cases. Otherwise, there is a serious risk thatdelinking PRSP and PRGF activities would elimi-nate a key incentive—both for countries and for theFund as an institution—to make progress on thePRSP front, while reverting to a “business as usual”mode on the PRGF side.

126

General Remarks

Executive Directors welcomed the timely reportof the Independent Evaluation Office (IEO). The re-port has a number of constructive recommendations,which will require careful further follow-up and im-plementation with a view toward strengthening thePRSP approach, clarifying the Fund’s role in this ap-proach, and enhancing the Fund’s advice and assis-tance to low-income countries. More broadly, theIEO report is a valuable contribution to the ongoingreview of how to improve the effectiveness of theFund’s engagement with low-income countries.

Most Directors considered that the PRS approachhas had a positive impact on economic policy designand implementation in low-income countries. It hashelped to improve country ownership, enhance par-ticipation, and provide better-quality strategies, andit is generally evolving in the right direction. At thesame time, Directors stressed that substantial scopeexists for better implementation of the current ap-proach, based on the evolving experience and the di-rections for change identified in the IEO report. Inthis context, Directors noted that the report foundthat PRSPs suffer from a multiplicity of objectives,which contribute to significant weaknesses in theirimplementation, observing that: the approach is per-ceived to be externally driven; participation some-times has been narrow, particularly in formulation ofthe macroeconomic framework underlying thePRSP; and that PRSPs often have lacked opera-tionally viable strategies. At the same time, Direc-tors cautioned against drawing premature conclu-sions about the ultimate success of the PRSPapproach based on only five years of experiencewith its implementation. It was noted that resolvingthe mismatch between the ambitions of this ap-proach and what can realistically be achieved willcontinue to remain a challenge. In light of this, the

staff should draw on the IEO’s recommendations invarious aspects of its work on the role of the Fund inlow-income countries to deepen and refine its analy-sis on ways to enhance the effectiveness of the PRSPapproach. In particular, the upcoming Annual Bank-Fund PRSP Progress in Implementation Report, andforthcoming discussions on the role of the Fund inlow-income countries, will offer such opportunities.

Directors agreed that the report provides a usefulperspective on the PRGF. They were encouragedthat PRGF-supported programs increasingly arebeing aligned with the country-owned PRSP, eventhough such alignment is still somewhat limited.They welcomed the finding that the design of theseprograms has improved in a number of ways. For ex-ample, fiscal targets have become more flexible toaccommodate increased expenditures on pro-poorprograms, and there is no evidence of an excessivedisinflationary bias. At the same time, Directorsnoted the challenge of basing Fund-supported pro-grams on a full understanding of micro-macro link-ages, which are by their nature quite complex. Di-rectors also considered that more should be done tointegrate the results of poverty and social impactanalysis. They noted that, as is the case with thePRSP process, the Fund’s adaptation of its programsand policy advice in low-income countries towardgrowth and poverty reduction is an evolutionaryprocess. In this context, they looked forward to therecommendations of the management-led Commit-tee on Low-Income Country Work to provide newimpetus and focus to Fund work on low-incomecountries. Directors also considered that further staffanalysis will be needed on improving the design ofPRGF-supported programs, and gearing them to ad-dress macroeconomic policy challenges of achievingthe Millennium Development Goals (MDGs).

Directors welcomed the IEO report’s discussionof the broader implications of the PRS approach for

THE ACTING CHAIR’S SUMMING UP

EVALUATION REPORT OF POVERTY REDUCTION STRATEGY PAPERS AND THEPOVERTY REDUCTION AND GROWTH FACILITY BY THE

INDEPENDENT EVALUATION OFFICE


127

SUMMING UP BY THE ACTING CHAIR

the Fund’s role in low-income countries. Theylooked forward to work on refining the role of theFund in low-income countries, including the paperon instruments and financing in August, and betterdefining the Fund’s strategy to help low-incomecountries make progress toward meeting theMDGs. The joint paper with the Bank on aid, aideffectiveness, and financing of the MDGs will dis-cuss how aid might be better mobilized. The up-coming discussion on PRGF financing and instru-ments will consider how to finance PRGFoperations beyond 2005 and will discuss proposalsfor instruments to allow the Fund to respond moreeffectively to low-income countries faced withshocks. The Board will also have an opportunityafter the Annual Meetings to discuss the Fund’srole in providing appropriate signals about low-in-come members’ macroeconomic policies in the ab-sence of a need for Fund financing.

Individual Recommendations

Recommendation 1. Introduce greater flexibilityin the implementation of the PRS approach to betterfit the needs of countries at different stages of theprocess and with different capacities and politicaland administrative systems.

Directors agreed that the PRS approach will needto be implemented pragmatically and flexibly, takingdue account of country-specific circumstances andcapacity constraints as well as the need to ensure thatFund-supported programs continue to be designed toassure macroeconomic stability and help membersaccelerate the pace of progress towards the MDGs. Inthis regard, it will be important to have a clear idea ofthe core objectives of the PRS approach.

Directors noted that the PRGF should be linkedtightly to PRSPs that provide a sound operationalroad map. However, where PRSPs are not yet opera-tionally viable, the Fund should not insist on immedi-ate tight alignment between the PRSP and the PRGF-supported program. Instead, Fund staff should workwith these members to strengthen the macroeco-nomic frameworks in their PRSPs so as to move to-ward eventual alignment. In this context, Directorsnoted that the Fund’s efforts to build capacity andstrengthen institutions through country-driven andproperly targeted technical assistance to these mem-bers will be particularly important. At the same time,Directors cautioned that increased flexibility shouldnot imply delinking the PRGF from the PRSPprocess, and noted that the Fund would still seek toapply the PRSP principles in its program work.Moreover, maintaining a link between the PRGF andthe PRSP provides a key incentive for members andthe Fund to improve the PRSP process.

Recommendation 2. Shift the emphasis of the ini-tiative from the production of documents to the de-velopment of sound domestic policy formulation andimplementation processes.

Directors agreed that the PRS approach has thepotential to encourage the development of country-owned and credible long-term strategies for growthand poverty reduction. To accomplish this objec-tive, there should be less emphasis on documentpreparation, and more emphasis on improving thecapability of countries to develop and implementpolicies supportive of growth and poverty reduc-tion. Directors underlined that governments are ex-pected to be in control of the development processand policy options.

Directors noted that managing the tension be-tween country ownership of the policies and pro-grams and donor selectivity is central to the designof the PRS approach. Some Directors agreed withthe report’s recommendation that countries shouldset explicit criteria for judging progress towards keyintermediate objectives related to the domestic pol-icy formulation, implementation, and monitoringprocess for their PRSPs, and that Fund and Bankstaff should provide candid assessments of thosebenchmarks. Many Directors cautioned, however,that this could imply excessive Fund involvement inassessing the country’s decision-making processes,and establish an unwarranted direct linkage betweensuch assessments and the Fund’s lending decisions.They believed that the use of such conditionality bythe Fund to improve domestic policy processes is in-appropriate because it could undermine the legiti-macy of domestic institutions and processes. Direc-tors noted that further reflection and discussion willbe needed on how the Fund should react in caseswhere it believes that the member could aim higher,or where the pace of progress chosen is not ambi-tious enough.

Recommendation 3. Clarify the purpose of theJSA and redefine the vehicle accordingly.

Directors emphasized the need for a reformula-tion of the Joint Staff Assessment—the instrumentused by the Fund and the Bank to assess PRSPs,with an emphasis on graduated rather than binary as-sessments. Many Directors noted that the purpose ofthe staff assessment of PRSPs should be to providecandid feedback to countries and to help coordinatewith the Bank. They believed that not using the as-sessment to shape Fund lending decisions and re-moving the requirement of Board endorsementwould help enhance country ownership of PRSPs. Inthe absence of any other coordination vehicle, Direc-tors favored keeping a joint instrument for the as-sessment. They noted that any reformulation of the

128

Summing Up by the Acting Chair

assessment instrument should reflect and maintainthe two institutions’ commitment to the PRSP initia-tive, while reducing the documentation requirementsfor the PRSP process and utilizing the limited staffresources more effectively. Directors looked forwardto discussing specific recommendations by the staffon options for the assessment instrument in the con-text of the Annual Bank-Fund PRSP Progress in Im-plementation Report.

Recommendation 4. Clarify what the PRS ap-proach implies for the IMF’s own operations andstrengthen the implementation of the agreed role.

Directors agreed that the Fund needs to set outmore clearly its own role in the PRS approach in eachcountry, based on the Fund’s core mandate in macro-economic and related structural policy issues, and usethe tools—including surveillance and other monitor-ing mechanisms—that are best suited to the individ-ual case. In doing so, the Fund would need to con-tinue to strengthen its collaboration with the WorldBank and other donors. While Directors welcomedthe streamlining of Fund conditionality, somestressed that aggregate Fund-Bank conditionalityneeds to be monitored and reduced as well.

Many Directors also supported a more active rolefor the Fund in the public debate on macroeconomicpolicy design and implementation—especially re-garding assumptions, alternative policy scenarios, andtrade-offs. This role could be facilitated through Fundanalysis and research, the Article IV consultationprocess, the provision of technical assistance, and abetter-defined and enhanced role of resident represen-tatives. These Directors believed that such a roleshould be fully compatible with country ownership ofprograms and policies. However, other Directors wereof the view that a more proactive public role could beseen as influencing the political decision-makingprocess, which is not part of the Fund’s mandate.Some of these Directors also commented on the limi-tations of the participatory approach, including a ten-dency to produce broad and unfocused documents.

Recommendation 5. Strengthen the prioritizationand accountability on what the IMF is supposed todeliver within the broader partnership framework,built around the priorities emerging from the PRSprocess, and ensure resources match commitments.

Directors welcomed the IEO report’s emphasis onthe need to define priorities for the work of the Fundin low-income countries in the context of its re-source constraints. In this regard, the prioritizationof budget resources must be guided by the Fund’soverall mandate. Directors called for a careful as-sessment of the resource implications of adaptingthe Fund’s role along the lines of the report’s recom-mendations. Some Directors felt that many of the re-port’s specific recommendations—in particular,more direct and deeper involvement in the PRSPprocesses—would increase the call on the Fund’sstaff resources considerably, and would require bal-ancing these additional demands against other prior-ities. Some others thought that the report’s recom-mendations would not involve an expansion of theFund’s responsibilities beyond what was indicated inthe original policy papers establishing the PRSP andthe PRGF. They noted the need to discuss the extentto which the original expectations for the Fund’s rolewere too ambitious. Several Directors felt that thereis scope for changing the Fund’s way of doing busi-ness in low-income countries within existing re-source constraints. Directors looked forward to staffviews on ways to improve the Fund staff’s involve-ment in the PRSP in the forthcoming Annual Bank-Fund PRSP Progress in Implementation Report andin the review of the resident representative program.

Recommendation 6. The IMF should encouragea strengthening of the framework for establishingthe external resources envelope as part of the PRSapproach.

Directors stressed that adequate, timely, and pre-dictable donor support is crucial to the success of thePRS approach and the Fund should play a supportiverole with donors and low-income members to helpensure adequate provision of aid to achieve theMDGs. In this regard, the Fund needs to considerhow its signals can be clear and useful to its mem-bers. In particular, Fund signals should not lead tothe inappropriate interruption of long-term develop-ment and poverty reduction finance. The Fundshould also work with the donor community to en-hance aid coordination and encourage medium-termsupport linked more effectively to country-ownedpoverty reduction strategies.

129

Date post:	25-Mar-2022
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times