- International Monetary Fund. Independent Evaluation Office
- Published Date:
- April 2003
© 2002 International Monetary Fund
Production: IMF Multimedia Services Division
Typesetting: Alicia Etchebarne-Bourdin
Cover design: Martina Vortmeyer
Figures: Sanaa Elaroussi
Evaluation of prolonged use of IMF resources—
[Washington, D.C.]: International Monetary Fund, 2002.
Includes bibliographical references.
1. International Monetary Fund-Evaluation. 2. Economic assistance. 3. Economic assistance-Case studies. I. International Monetary Fund. Independent Evaluation Office.
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
Summary of Contents
The following symbols have been used throughout this report:
– between years or months (e.g. 2000–01 or January-June) to indicate the years or months covered, including the beginning and ending years or months;
/ between years (e.g. 2000/01) to indicate a fiscal (financial year).
“Billion” means a thousand million.
Minor discrepancies between constituent figures and totals are due to rounding.
On September 30, 2002, SDR 1 = US$1.3227.
Some of the documents cited and referenced in this report were not available to the public at the time of publication of this report. Thus, they have not been listed in the bibliography. However, under the current policy on public access to the IMF’s archives, some of these documents will become available five years after their issuance. They may be refer-enced as EBS/YY/NN and SM/YY/NN, where EBS and SM indicate the series and YY indicates the year of issue. Certain other documents are to become available ten or twenty years after their issuance depending on the series.
The Independent Evaluation Office (IEO) was established by the International Monetary Fund’s Executive Board in July 2001. The office operates independently of IMF management and at arm’s length from the IMF’s Executive Board. Its mission is to provide objective and independent evaluation of issues related to the IMF’s mandate and thereby support the Executive Board in its governance and oversight responsibilities, to contribute to enhancing the learning culture of the IMF, and to promote understanding of the IMF’s work.
This Evaluation of Prolonged Use of IMF Resources is the first report produced and published by the IEO. The terms of reference for the evaluation were prepared following extensive consultation with various stakeholders. An initial draft of an issues paper for the evaluation was posted for comments on the IEO’s website (www.imf.org/ieo). After input from a range of groups and individuals, the final terms of reference were determined by the IEO, and also posted on the website. Interested parties were invited to submit material relevant to items included in the final terms of reference. This consultative approach will be followed for all IEO evaluations.
The process for review and publication of the final report followed the standard rules that have been established by the IEO; a detailed description of these procedures is also available on our website. The final report was circulated to IMF management and the Evaluation Group of Executive Directors at the same time. Subsequently, management’s comments, and the IEO’s response, were circulated to the Executive Board, but the report was not changed in light of these comments. The Executive Board discussed the evaluation on September 20, 2002, and the Chairman’s summing up of that discussion is also included.
This report was prepared by a team headed by David Goldsbrough and including Kevin Barnes, Isabelle Mateos y Lago, and Tsidi Tsikata. Sections of the report have also benefited from comments and other input from Professor Graham Bird, Professor Jong-Wha Lee, David Peretz, and Professor Andreas Wimmer, but the final judgments are the responsibility of the IEO alone. Research assistance from Mwaffak Taib, Ned Rumpeltin, Yothin Jinjarak, and Sergei Peredriy and administrative support by Annette Canizares and Arun Bhatnagar are gratefully acknowledged.
Montek S. Ahluwalia
Independent Evaluation Office
List of Abbreviations
Agricultural Sector Adjustment Credit
Asian Development Bank
Association of South East Asian Nations
Banque Centrale des Etats de l’Afrique de Ouest
Bureau of Internal Revenue
Budget reporting system
Compensatory and Contingency Financing Facility
Caribbean Development Bank
Consolidated public sector deficit
Caisse de Péréquation et de Stabilisation des Prix
Country strategy paper
Comprehensive Tax Reform Package
Debt sustainability analysis
Enlarged access policy
Extended Fund Facility
Emerging Market Global Bond Index
Enhanced Structural Adjustment Facility
Fiscal Affairs Department
Foreign currency deposit
Financial Sector Assessment Program
Gross domestic product
Government Finance Statistics
Gross national product
General Resources Account
General sales tax
Heavily Indebted Poor Country
Inter-American Development Bank
Independent Evaluation Office
International financial institution
International Financial Statistics
Information Notice System
Letter of intent
Memorandum of Economic Policies
Net present value
Overseas development assistance
Operations Evaluation Department
Office of Internal Audit and Inspection
Policy Development and Review Department
Public Information Notice
Produce Index Unit
Policy Framework Paper
Philippine National Bank
Poverty Reduction and Growth Facility
Poverty Reduction Strategy Paper
Structural Adjustment Facility
Structural Adjustment Loan
State Bank of Pakistan
Special Contingent Account
Special drawing right
Supplemental Reserve Facility
Use of Fund resources
U.S. Agency for International Development
West African Economic and Monetary Union
Water Resources Administration and Power Development Authority
World Economic Outlook
Report of the IEO on Prolonged Use of IMF Resources
1. In this evaluation, which is the first by the Independent Evaluation Office (IEO), we examine the issue of the prolonged use of IMF resources, which has been the subject of external criticism and internal concern. Some critics have argued that prolonged use constitutes a departure from the IMF’s traditional mandate of providing temporary balance of payment support and suggests a lack of effectiveness of IMF-supported programs. Others have argued that frequent recourse to IMF arrangements is less of a problem than critics contend, that it can take place for good reasons in countries with deep-seated adjustment problems, and that it can be fully compatible with the IMF’s mandate. We analyze the various reasons that have caused prolonged use to expand over the last three decades, be they related to IMF lending policies, to specific characteristics of the countries concerned, to shortcomings in program design and implementation, or to broader institutional factors. In so doing, we inevitably touch upon some issues that have been at the heart of recent controversies about the effectiveness of IMF-supported programs. Given the IEO’s mandate, the focus is on the role of the IMF, but we are not suggesting that the IMF is responsible for all problems that contributed to prolonged adjustment difficulties. Clearly, the governments of the countries themselves bear primary responsibility for their policies.
2. We used a combination of methodological approaches in the evaluation, including empirical and econometric analyses of the broader group of prolonged users; three main country case studies (of Pakistan, the Philippines, and Senegal); two more limited case studies of “graduators” from prolonged use (Jamaica and Morocco); interviews with a wide range of stakeholders in the countries studied and with IMF staff; and a series of questionnaires to prolonged users’ authorities, official and private creditors, and IMF mission chiefs.
3. We find that the increase in prolonged use is partly a reflection of systemic factors arising from the changed role that the international community expects the IMF to perform-where the implications for the extent of prolonged use have not been fully recognized- and is also linked to program design and implementation problems, many of which are not specific to prolonged users.
Definition and Scope of Prolonged Use
4. There is no single definition of prolonged use and different definitions have been used in earlier studies on the subject. For this evaluation, we have used a definition that treats a country as a prolonged user if it has been under IMF-supported programs for 7 or more years in a 10-year period. (For example, a country would have to be in more than two consecutive three-year arrangements to be classified as a prolonged user.) The “fixed” version of this definition looks at all countries that met this definition at some point in the period covered by the evaluation. On this basis, 44 countries (29 of them PRGF-eligible) were prolonged users at some point during 1971–2000; a further 7 countries were prolonged users if precautionary arrangements are included. For some purposes, we have also used a “dynamic” definition according to which a country is treated as a prolonged user in a particular year if it has had an arrangement for 7 of the previous 10 years.
5. Prolonged use has expanded consistently since the 1970s in terms of number of countries, share of the IMF’s membership, and total financial exposure. This conclusion holds regardless of how prolonged use is defined and also if the analysis is done separately for low-income countries and middle-income countries. In terms of number of countries, most of the expansion is accounted for by those eligible for the concessional facilities; but in terms of financial exposure, it is predominantly associated with users of the IMF’s general resources. Furthermore, prolonged use is persistent, in the sense that countries are generally slow to “graduate” from such use. In 2001, arrangements with prolonged users represented about half of the total number of ongoing IMF-supported programs, with a total exposure of SDR 24 billion (i.e., about half of total outstanding obligations to the IMF).
Evolution of the IMF’s Policies with Respect to Prolonged Use
6. The emergence of the phenomenon of prolonged use is to some extent the consequence of the evolution of IMF policies on the length of use of its resources. A review of the history of this evolution (in Chapter 3) suggests the following broad conclusions:
Initially, the IMF was expected to provide financing only for a relatively short period, but over time the Executive Board accepted that many balance of payments problems, especially in the case of low-income countries, arose from deep-seated structural problems that required a longer time frame for adjustment. This led to acceptance of IMF financing being provided over a longer period. Even so, the departure was initially viewed as exceptional and access was not expected to be overly prolonged.
With the establishment of concessional financing facilities in the late 1980s to address deep-seated external problems in low-income countries, policies applied to the use of concessional and general resources began to diverge, especially in the 1990s. There was a gradual relaxation of any implied upper limits on the time spent under IMF-supported programs in the concessional facilities. Simultaneously, there was a tightening of the rules governing the length of use of the IMF’s general resources, although not to the point of setting a ceiling on that length.
The evolution of these facilities did not fully and explicitly recognize some of the consequences in terms of the extent of prolonged use and its potential adverse impact. This was in part a consequence of differences of view within the Executive Board over what the longer-term role of the IMF should be in such cases. As a result, the IMF has been left with a mismatch between its core operational approach (which is still focused on promising and achieving a restoration of sustainability within a relatively short time frame) and some of the new tasks it is being asked to perform, especially the provision of support for longer-term structural adjustment.
The approach to prolonged use as it evolved over time reflected the nature of the financing constraint. Initially, the limited availability of funding for the IMF’s concessional facilities precluded any consideration of a long-term involvement. Subsequently, the reluctance of donors to maintain their aid flows at a level consistent with the intended diminished reliance on IMF lending, implied that the IMF would either have to remain involved through repeated programs until a sustainable external situation could be reached, or withdraw from the countries concerned midway in the process before external viability could be achieved.
Over time, a set of guidelines was approved by the Executive Board to reduce the likelihood of prolonged use of IMF resources. These guidelines called for front-loading the adjustment effort and monitoring implementation closely; calibrating access to IMF resources to ensure a diminishing reliance over time; comprehensive ex post assessments of previous programs; and the preparation of explicit exit strategies for prolonged users. The evaluation indicates that these guidelines were not implemented consistently. One reason may have been the Board’s reluctance to endorse a specific definition of prolonged use, which led to ambiguity about the application of the guidelines in individual cases.
Characteristics of Prolonged Users
7. Prolonged users appear to face external circumstances and have fiscal characteristics that are less conducive to swift adjustment. These include lower trend export growth and more volatile terms of trade, as well as a more rigid structure of government expenditure, lower tax revenues, and a higher public debt burden. Although the statistical significance of these differences between prolonged and “temporary” users is limited and directions of causation are hard to disentangle, some of these characteristics have potential implications for program design. For example, they suggest the need for caution against the danger of overambitious projections for exports or tax revenues being built into program targets, since these could cause problems for program implementation. Evidence from the case studies and cross-country analyses—discussed in Chapter 5—suggests that such overoptimism often was a problem in program design. However, there are also indications that prolonged users suffered from more serious imbalances than “temporary” users at the start of their prolonged use period, which might account in part for the greater length of their adjustment process.
Effectiveness and Design of Prolonged Users’ IMF-Supported Programs
8. While it is hard to disentangle the effects of poor implementation from design problems, IMF- supported programs in prolonged use cases achieved much less than projected:
Although the methodological problems are considerable, econometric evidence suggests that in cases of prolonged use of general resources, IMF-supported programs tend to be associated with a negative impact on growth—after controlling for the endogeneity of a country’s decision to seek IMF assistance and for other determinants of growth. There is no such negative association for “temporary” users. Prolonged users of concessional resources do not show any negative impact on growth.
A cross-country analysis of the extent of fiscal adjustment suggests that, in multiyear arrangements, prolonged users adjusted less than “temporary” users over the program period. In Stand-By Arrangements (SBAs), prolonged users also appear to have achieved a slightly smaller reduction in the overall public sector deficit than “temporary” users and the shortfalls from the targeted fiscal adjustment seem to have been larger, although lack of a fully consistent database makes such comparisons difficult.
The case studies show that, during the long period of IMF program involvement, significant progress toward solving these countries’ economic difficulties was eventually achieved in the Philippines and Senegal and even more so in Morocco, although with a mixed record across areas of economic policy and at a much slower pace than originally envisaged. The record in Pakistan and Jamaica was more disappointing. In all cases, substantial challenges remained at the end of the prolonged use period reviewed, especially as regards institutional reforms in tax administration and the broader public sector.
9. A cross-section analysis of program design in prolonged use cases reveals interesting differences with “temporary” user cases. Some of these might have contributed to lower program effectiveness:
Prolonged users’ programs had an optimistic bias as regards projections of real GDP growth and (for users of concessional facilities) export growth.
In both SBAs and multiyear arrangements, the magnitude of the targeted fiscal adjustment (as measured by the primary fiscal balance) was markedly lower in prolonged use cases. In SBAs, the targeted adjustment of the external current account was also much lower. Contrary to established guidelines the targeted adjustment effort in multiyear arrangements was less front-loaded in prolonged use cases.
Conditionality in prolonged users’ programs was, on average, less extensive than in other programs and its modalities involved fewer performance criteria and prior actions (as opposed to structural benchmarks and reviews). Moreover, a higher proportion of prolonged users’ performance criteria were waived. However, the evidence from the case studies, discussed below, suggests that it was not so much the magnitude of conditionality as its prioritization and integration into program design that was critical.
Programs with prolonged users, particularly those supported by general resources, were frequently subject to interruptions (both temporary and irreversible), to a greater extent than were “temporary” users’ programs.
Access to IMF resources in successive prolonged users’ arrangements declined only in a minority of cases, and disbursements were on average somewhat more front-loaded than in arrangements with “temporary” users.
Staff inputs in programs with prolonged users were smaller than in other programs, although the evidence suggests that this had no significant impact on program outcomes. Staff turnover on country assignments was very high in prolonged user countries, although not worse than in “temporary” users.
10. The case studies broadly confirm the pattern emerging from the cross-section analysis and highlight a number of additional issues that hampered program effectiveness. We note that not all programs suffered from these problems and there is some evidence of improvement over time. While these problems likely contributed to prolonged use, they were not special to such cases.
Discrepancies between the time frame of programs and the magnitude of their objectives. There was a tendency to overoptimism about the effectiveness of the structural reform agenda in all three of the country cases. On the macro side, the tendency was most marked in Pakistan. Too little attention was often paid to analyzing how the real economy would respond to key policy measures and to assessing the expected sources of growth leading to overly optimistic forecasts of key magnitudes.
The risk to programs of weak ownership and political commitment was often understated, and not enough attention was paid to assessing and developing implementation capacity. Assessments of such issues in internal documents were variable and, in those relatively few cases where a clear assessment was made, subsequent Board papers were much less candid. There is evidence, however, of a somewhat greater attention to such issues in some more recent programs.
Many programs had difficulty in dealing with uncertainty, in part because program documents often did not analyze the key risks to a program and specify how policies would broadly respond to those risks.
Relatively few systematic ex post assessments of programs were undertaken and (with the exception of the Philippines) there was generally limited discussion of exit strategies from prolonged program involvement.
11. The three main country cases also provide lessons for approaches to structural conditionality. Many of these lessons are not specific to prolonged users, but they are especially important in these cases.
The extent and structure of conditionality was much less important than an underlying political commitment to core policy adjustments. The three country cases illustrate a wide variety of approaches to conditionality, ranging from extensive and detailed in the case of Pakistan to a heavy reliance on reviews in the case of the Philippines; none proved especially effective in periods when ownership was weak. Likewise, very detailed conditionality (e.g., a detailed matrix approach) was not effective in enhancing implementation when political commitment was lacking.
Structural conditionality was often poorly prioritized, so that compliance with a subset of these conditions did not ensure that the most critical problems were being addressed even though it was often sufficient for continued access to IMF resources.
Forms of conditionality that focused on policy rules or procedures, rather than one-time discretionary actions, appear to have been ultimately more effective, especially when dealing with deep-seated structural issues.
Prior actions were not always well integrated into program design and did not always focus on the most critical issues. This might explain in part the general finding that the number of prior actions has little influence on successful program implementation. The history of prior actions in the area of agricultural income taxation in Pakistan is one illustrative example.
Influence of IMF Governance and Other Institutional Factors on Prolonged Use
12. Internal institutional factors have contributed importantly to the program design problems outlined above:
The approach to structural reforms was, until recently, often characterized by insufficient emphasis on fostering the deep institutional changes needed in critical areas. These problems were compounded by the fact that collaboration between the IMF and the World Bank often did not yield an operationally effective integration of priorities and work programs. The case studies illustrate some of the difficulties that programs encountered in addressing some especially intractable structural problems that were central to the sustainability of adjustment (e.g., effective tax collection procedures in all three countries). Recent initiatives to streamline conditionality should help to address these problems and are welcome.
The IMF’s approach to program design has until now often given insufficient priority to a proper assessment of the implementation capacity constraints that a program might face, be they related to political feasibility or to administrative capacity. In best practice cases, efforts are made to take account of these constraints, but there are insufficient systemic incentives to ensure that such an approach would be followed more generally.
Some of the case studies (most notably Pakistan) and a survey of IMF mission chiefs suggest that political considerations have been an important factor in program-related decisions on some occasions. While political considerations are bound to enter into decision making for an institution where the ultimate power of approval rests with shareholder governments, it is necessary to ensure that technical judgments and political considerations do not get blurred in these cases. This only dilutes accountability and tends to reduce the credibility of those programs.
There is evidence that internal incentives in the IMF encourage overpromising in programs. This results from both the relatively short time frame of programs, forcing optimistic assumptions about the pace of adjustment and also from a desire to maximize the program’s catalytic role. This led to a tendency to downplay risks. Even when, as was often the case, they were well identified during the internal review process, the assessment of risks was not candidly presented to the Executive Board.
The case studies suggest that, by and large, surveillance failed to play a major independent role in prolonged user cases. In some cases (most notably the Philippines), doubts that surveillance would be a sufficiently strong instrument for some purposes (e.g., to provide a continuing framework for “good” policies and to send a signal to other lenders) appear to have contributed to continued program involvement even when there were questions about the balance of payments need.
The IMF’s ability to learn from experience is hampered by (i) the relative scarcity of systematic ex post assessments of programs and (ii) the slow pace at which lessons learned in the context of cross-country policy reviews—which are often insightful—permeate operational practices. Moreover, many of the most candid internal assessments and debates on alternative policy strategies in individual countries were not reflected in subsequent Board papers.
13. The international community’s evolving expectations of the IMF’s role contribute in important ways to the expansion of prolonged use. In particular, official donors and creditors have tended to link increasing parts of their financing flows to the existence of an IMF lending arrangement acting as a “seal of approval” on recipient country policies. While this tendency can be justified by the legitimate desire to ensure aid effectiveness, it is not clear that the “seal of approval” provided by IMF-supported programs, especially as currently designed, suitably addresses all the concerns of donors and creditors—in particular those related to long-term sustainable growth and the deep institutional changes that it requires. Furthermore, there is some evidence that “seal of approval” demands are the source of pressures to agree to programs with a low probability of success, which only undermines the seal of approval.
Implications of Prolonged Use of IMF Resources for the Member Country and the IMF
14. Prolonged use appears to have both positive and negative implications for institutional development in prolonged users. Most prolonged users’ authorities acknowledged that successive programs had been accompanied by a positive transfer of economic management skills, although views differed regarding the scope of that transfer. However, there was a general sense—both from the case studies and questionnaire responses from a broader group of prolonged users—that throughout its involvement, the IMF had paid insufficient attention to institutional reform and to the development of implementation capacity, even in cases where programs had facilitated access to IMF technical assistance. The views expressed regarding the impact of prolonged use on the policy formulation process were generally negative, in particular because program negotiations were often characterized as proceeding in a way that left too little space for policy debate and the formulation of homegrown policies. While this is a general problem, it is particularly damaging in prolonged use cases because of the extended period involved. However, a number of officials acknowledged that the primary responsibility for any failures in policy implementation had to lie in the countries themselves and that a “blame-the-IMF” approach could, in itself, be detrimental to ownership.
15. Prolonged use may also affect the IMF itself, in two ways: it has a potential impact on its finances, and there are indications that it also affects its credibility. In terms of financial impact, our analysis indicates that, in spite of its expansion, prolonged use does not appear to have been a binding constraint on the IMF’s lending ability, either in the General Resources Account (GRA) or in concessional trusts. However, since decisions on the size of access to IMF resources and on quota increases are endogenous, it is difficult to determine ex post whether prolonged use led to implicit rationing of resources to other users. Prolonged users’ programs have also contributed significantly to the growing mismatch between the IMF’s staff resources and its lending activities. The impact of prolonged use on the IMF’s credibility cannot be quantified rigorously but there is anecdotal evidence that the willingness of the IMF to maintain or renew its support in the face of uneven implementation weakened the leverage of conditionality. There is also some econometric evidence that, where IMF resources have been found to have a catalytic effect on other financing flows, that effect weakens with prolonged use.
Conclusions and Recommendations
16. Prolonged use, even when it takes place for “good” reasons, can have significant adverse implications, both for prolonged users and for the IMF. The expansion of prolonged use reflects both systemic factors and weaknesses in program design. The latter largely arise from a mismatch between the functions taken on by the IMF for systemic reasons and its institutional culture. Implementation weaknesses also played a part. While they largely reflect policy choices in the countries themselves, which are beyond the scope of this evaluation, their prevalence is also related to the systemic and institutional factors mentioned above. The IMF is in a process of change, and many initiatives have been adopted in recent years that, if implemented faithfully to their spirit, should go a long way towards addressing some of the problems highlighted by this report, particularly as regards ownership, the design of conditionality, surveillance in program countries, and IMF-World Bank collaboration.
17. The recommendations summarized below have two objectives: to diminish the incentives for prolonged use, including by enhancing program effectiveness, and to reduce its adverse consequences. Some of these recommendations would be applicable only to actual prolonged users. Others are of more general applicability to the IMF’s approach to programs, not just prolonged use cases, but they constitute a preventive strategy that would reduce the likelihood of prolonged use in future. We recognize that some of the recommendations are likely to require some increase in resources, especially in the short term when offsetting resource savings will not yet have materialized. Several of the recommendations are not new, either because they merely give renewed emphasis to existing policies and guidelines that do not appear to be uniformly implemented, or because they constitute an endorsement of recent initiatives, to which they only add specific proposals regarding their implementation.1
Recommendations on the rationale for IMF involvement
1. The Executive Board should adopt an explicit definition of prolonged use, as a trigger for the adoption of automatic due diligence procedures. The definition could use different criteria for the general and concessional resources.
2. In view of the experience with ineffectively implemented programs, greater efforts should be made at judging when countries are ready to implement programs, especially in situations of prolonged use. On this basis, the IMF should be willing to be more selective in extending financial support.
3. The IMF should aim to provide the international community with credible alternatives to the current situation where IMF-supported programs have become a precondition for the provision of many other sources of financing by donors and creditors. This could be done by developing a mix of tools (e.g., strengthened surveillance, PRSP assessments, precautionary arrangements, or shadow programs) to deliver seals of approval suited to the varied needs of donors and creditors.
4. An explicit “exit strategy” should be developed for identified prolonged users, although without setting rigid limits on the duration of IMF program involvement.
5. A differentiated rate of charge could be considered for prolonged users. While there is no evidence that cost of IMF resources is a factor in prolonged use, it would serve as a signaling device that could possibly provide at least a political incentive against prolonged use.
Recommendations for program design and implementation
6. Since implementation is critically dependent on ownership, it would be desirable to evolve specific operational procedures to ensure that program design places greater emphasis on ownership and the nature of the domestic policy formulation process. This approach is already embedded in initiatives currently under way. The following specific suggestions are offered to operationalize the process.
The IMF should move toward a situation where the normal procedure would be for the authorities to have the initial responsibility for proposing a reform program, which should be the starting point for negotiations. The speed at which this can be done will vary from country to country, depending on administrative capacity. Such an approach should not be an additional prerequisite for financial support, but countries should be encouraged to adopt it.
The aim should be to move as soon as possible to a situation in which the core elements of a program are subject first to a domestic policy debate within the member country’s own policy-making institutions.
As is already “best practice,” Article IV surveillance reports should actively seek to present alternative policy options and to analyze the tradeoffs between them so as to encourage an open debate on alternative policy options.
7. Programs should give much greater emphasis to fostering key institutional changes and to strengthening implementation capacity. Staff reports should include an explicit assessment of key implementation challenges foreseen and ways to address them.
8. There should be greater selectivity in program content, based on a clearer prioritization of conditionality and a better integration of the latter in program design. This is the essence of ongoing efforts to streamline conditionality, which we strongly support. To reach this objective, the following steps would be key:
A deepening of operational collaboration with the World Bank, to ensure an effective meshing of priorities and time frames of the two institutions on key issues. Recent initiatives in this area are welcome, but substantial changes in operating approach and sustained emphasis by management will be needed to make collaboration effective.
Systematically incorporating into program documents more in-depth analyses of real economy responses to the key policy elements of programs and of the sources of growth, while spending proportionately less time fine-tuning traditional financial programming exercises. Such analyses would draw, where appropriate, on the expertise of the World Bank.
Using conditionality as a means to direct attention to critical reforms and emphasizing substantive rather then formal progress toward the program’s objectives both in setting conditionality and in assessing compliance with it. Greater selectivity in the use of conditionality should be accompanied by less recourse to waivers.
Making greater efforts to tailor the effective time frame of program design to the foreseeable length of the reform and adjustment process, not necessarily by lengthening the time frame of arrangements themselves, but by casting individual arrangements within a longer-term strategic framework and recognizing upfront the need for repeat arrangements where adjustment is expected to be protracted. This approach would build on the existing internal country strategy papers, but with the core elements included in reports to the Executive Board.
9. UFR reports should include more explicit discussions of the major uncertainties faced by the program and of how policies will be adapted if underlying assumptions are not borne out by subsequent developments. To counteract any bias to overoptimism, the reports should discuss explicitly how programs would be adapted if other available forecasts were to prove more accurate.
Recommendations for IMF governance
10. Systematic ex post assessments of programs should be undertaken, with the key messages reported to the Executive Board, as part of a broader effort to disseminate more effectively “best practices” and lessons learned, and to maximize the effectiveness of the review process. As part of this process, the following points deserve emphasis:
Staff reports, especially those involving UFR requests by prolonged users, should provide more perspective on the history of the IMF’s program involvement in the country, highlighting what has been achieved and where previous programs have fallen short of their objectives.
More follow-up monitoring should be under-taken when programs go off-track, especially for prolonged users. This will require improvements to the existing (MONA) database.
11. Steps should be taken to strengthen further surveillance in prolonged use cases, going beyond the recent revisions to surveillance guidelines aimed at improving surveillance in program countries. Further steps should include:
A clarification of the expectations on the role of surveillance in program cases, going further than the simple reassertion of existing, but unevenly implemented, guidelines and stressing the need for ex post assessments of programs’ achievements.
There is a case for creating greater operational separation between UFR and surveillance missions in prolonged use cases, although implementing this separation raises delicate tradeoffs. The appointment of a mission chief chosen from outside the area department for surveillance missions to prolonged users could be considered provided there are suitable continuity and coordination safeguards.
There is merit in seeking a second opinion—including from outside the IMF—on key policy issues that appear to be contributing to prolonged use.
The precise frequency of Article IV consultations with program countries is less important than that they take place at times when a “fresh look” would be most valuable.
12. The IMF should strengthen the ability of its staff to analyze political economy issues in order to achieve a better understanding of the forces that are likely to block or enhance reforms and to take these into account in program design.
13. A broad review of explicit and implicit incentives facing IMF staff should be undertaken, in particular with a view to reducing the excessive turnover of staff working on countries and to fostering increased candor in staff reports and greater accountability.
14. The appearance of undue political intervention in the IMF’s decisions to grant a country access to its resources undermines the credibility of programs. Procedures should be evolved so that political considerations, which are inevitably present in these decisions, are seen to be taken into account in a transparent manner, with decisions and accountability clearly at the level of the Executive Board on the basis of a candid technical assessment by staff of the risks and potential trade-offs.
Country Case Studies
1. Since 1988, Pakistan has been almost continuously under IMF-supported programs. However, all but the last arrangement (the 2000 SBA) suffered from substantial policy slippages and soon went off-track. This intensive UFR period coincided with a marked fall in GDP and export growth, along with a steady increase in poverty, only minor corrections of fiscal and external imbalances, and a buildup in vulnerabilities associated with foreign currency deposits. Substantial progress was achieved in a few structural reforms areas: liberalization of interest rates, trade regime and external payments, modernization of indirect taxes, and utilities pricing. However, the reform process was protracted, and considerable problems remain for lack of in-depth institutional reform, particularly in the areas of tax administration and public enterprise management. Economic governance problems also became more acute and widespread.
2. The limited achievements of successive programs were associated with considerable political and regional instability, but also reflected design and implementation problems. Most of the programs negotiated from 1988 onward suffered from overoptimistic assumptions regarding the path of exports, GDP, and domestic savings/investment growth. They also proved to be unrealistic about the pace and breadth of the structural reform agenda that was likely to be implemented by the government, including in the area of tax reform. While it is difficult to distinguish between failure to implement agreed policies from ex ante overoptimism about key economic variables, there is little doubt that limitations on implementation ability were not fully taken into account in the program design, and the reform agenda was not sufficiently prioritized. Ex ante overoptimism was also important as some of the programs were based on projections (e.g., for tax revenues and exports) that would have been difficult to achieve even with full implementation of agreed policies.
3. A basic problem was that successive governments had limited ownership of the programs and were often unable to sustain the reform and adjustment effort for more than a short period of time. As a result, some of the policy prescriptions of the programs, implemented in isolation from other equally important parts of the agenda, turned out to have adverse side effects (e.g., the net revenue shortfall—leading to excessive expenditure compression—caused by the substantial lag between the fall in revenues associated with import tariffs cuts and the full effectiveness of domestic tax reforms). Although there were limits to what an external agency could achieve in such circumstances, not all of the mitigating devices at the disposal of the IMF were utilized consistently, such as track record requirements, front-loading the adjustment effort while back-loading the disbursement of funds, or focusing conditionality on a few truly key objectives and pursuing them unwaveringly.
4. Programs with Pakistan had more than the average number of structural conditions, but this structure was not effective in overcoming weak political commitment and some of the conditions were not well integrated into program design. The history of the largely unsuccessful attempts to impose conditionality in the area of agricultural income taxation is an important example; a number of prior actions were met, but without a significant effect on the program’s prospects and without yielding any significant advance in taxation of agricultural incomes. Moreover, the perception that the IMF would resume financing within a relatively short time whenever programs went off-track appears to have undermined incentives to undertake difficult adjustment decisions, especially in the fiscal area. However, the most recent programs (from late 1997 onward) do indicate some learning over time.
5. Most of these design and implementation problems had their roots in institutional factors. In particular, decisions regarding the IMF’s involvement in Pakistan appear to have been heavily influenced by geopolitical considerations; the presumption of a short- to medium-term involvement only; the obligation to give IMF member countries the benefit of the doubt which, in combination with relatively frequent changes of government, contributed to many programs in succession; and collaboration with the World Bank that did not always produce an effective integration of priorities and time frames in dealing with key structural issues. As in other countries under program, IMF surveillance tended to be crowded out and, in particular, failed to give sufficient attention to the long-term implications of short-term stabilization measures for the sustainability of adjustment, to flag downside risks, and to highlight sufficiently the buildup of major vulnerabilities in the capital account (associated with the foreign currency deposits). Finally, as in many other prolonged use cases, there were too few occasions when the IMF was able to “step back” and reconsider the overall strategy on the basis of candid assessments of previous programs.
6. The main lesson to be derived from this experience is the need for greater selectivity, based on candid ownership assessments before committing the IMF’s resources, along with greater efforts to tailor program design to the specific circumstances and long-term needs of the country, focusing especially on key institutional changes.
7. The Philippines is the most extreme case of the prolonged use of IMF resources, with 23 programs between 1962 and 2000. The near-continuous involvement of the IMF in the 1960s and 1970s did not prevent an unsustainable debt buildup, which culminated in the early 1980s’ debt crisis. In this period, the IMF was dealing with a very difficult situation, epitomized by deep-seated governance problems. Although program design did not address many of the key structural problems, a lack of political commitment was the core issue, and in the circumstances, it would have been better for the IMF to have refrained from lending.
8. In the ten years following the 1983 debt crisis, the Philippines underwent a protracted adjustment process along with the restructuring of debt by private and official creditors tied to a succession of IMF-supported programs. As this process was completed, an EFF that was explicitly intended as an “exit program” was agreed in 1994, but before it could be completed it was overtaken by the Asian crisis. Subsequently, the IMF’s program engagement lasted until 2000, even though there were growing problems with governance and doubts about political commitment after the Estrada administration took office in 1998. Since 2000, the Philippines has been engaged in post-program monitoring with the IMF.
9. IMF-supported programs did encourage macro-economic discipline and did assist in the substantial transformation of the Philippine economy that occurred between the mid-1980s and the second half of the 1990s under the Aquino and Ramos administrations. Considerable progress was made in liberalizing the trade and exchange system, abolishing agricultural marketing monopolies, opening the economy to competition, and eventually deregulating oil pricing. However, many of these reforms took much longer to implement than originally planned, in part because some of the original time frames were overambitious and in part because, with the exception of some important periods when political leadership was strong, domestic ownership of the reforms did not extend to the legislative branch of government, which was critical for implementation. The 1989–91 EFF, in particular, had a reform agenda that proved overambitious in retrospect. A few long-standing and critical weaknesses, including a low saving rate and weak tax collection (despite a temporary improvement in the first half of the 1990s) remained unresolved at the end of the period and contributed to prolonged use. Despite some progress, the incidence of poverty remained high.
10. The IMF’s long involvement in the Philippines reflects a number of factors, including a difficult starting position; “seal of approval” influences; program design and implementation problems (especially difficulties of matching the time frame of longer-term institutional reforms to shorter-term programs); and fluctuations in the degree of political commitment to and ownership of economic reforms. Toward the end of the period, the rationale for program involvement seems at times to have been too broad, reflecting the perception that surveillance was not an effective instrument for some tasks. Thus, efforts to encourage reforms and reformers or to bind an incoming administration to “sound” macroeconomic policies appear to have been important factors behind continued programs, even when a balance of payments need had become questionable.
11. Institutional constraints on implementation, particularly the Philippines’ congressional-style political system, posed special challenges for program design and required structural conditionality to be exercised flexibly, primarily through reviews. This was probably the correct approach, but combined with the tacit understanding that the program relationship was likely to continue for an extended period, there was a tendency to “overpromise” to meet the requirements of short-term programs in the belief that recontracting would be possible. This led to some erosion of the credibility of conditionality. For some elements of reform—such as tax administration, where repeated efforts yielded some results in raising the tax effort, but which were ultimately not sustained—it would have been preferable if a longer time frame had been adopted from the start, combined with a greater emphasis on strengthening implementation capacity, and more direct attention to governance concerns. While there were more efforts at ex post assessment of programs than in the other two country cases, not all of the results of these assessments were conveyed to the Executive Board.
12. Except for a nearly two-year period during 1992–93, Senegal has had a continuous succession of IMF arrangements since 1979. From 1986, the bulk of IMF lending to Senegal has been under multi-year arrangements supported by resources from the IMF’s concessional facilities—one SAF and three ESAF/PRGF.
13. The main reasons for Senegal’s prolonged use of IMF resources include (i) large initial imbalances, deeply rooted in structural weaknesses of the economy, which were likely to require a long time to address in a sustainable manner; (ii) a broadening of objectives of programs, in step with the evolution of the SAF to the ESAF and the transformation of the ESAF into the PRGF; (iii) the use of IMF arrangements as a seal of approval for the provision of external finance by several multilateral and bilateral creditors and donors, combined with an approach to dealing with the debt problem that only brought about a resolution gradually; (iv) weaknesses in program design—in particular, the pre-1994 programs were too optimistic about how effective the adjustment strategy being pursued would be in promoting growth and sustainable financial viability; and (v) a stop-go pattern of program implementation that weakened the effectiveness of programs and contributed to the continuing “need” for IMF arrangements.
14. Successfully implemented programs tended to be characterized by strong political commitment (regardless of the precise nature of conditionality), significant upfront adjustment measures, and adaptations of policies during program reviews when there were substantial actual or prospective deviations from targets (usually fiscal targets). In the cases where implementation was weak, contributory factors included election cycle pressures that led to delays in implementing measures or sometimes to policy reversals.
15. IMF policies aimed at containing the phenomenon of prolonged use—such as reduction in access levels in successive arrangements, comprehensive ex post assessment of programs, and the formulation of “exit” strategies—were not applied consistently to Senegal. The rules on declining access levels were applied, although the country enjoyed slightly above average access levels in its three ESAF/PRGF arrangements. Ex post assessments were undertaken to some extent, although the most recent example, a country strategy paper prepared in 1998, did not draw lessons for program design from the nonimplementation of reforms under previous arrangements. On the issue of an “exit strategy,” we found no evidence that the issue of “graduation” from IMF-supported programs has been discussed internally or with the authorities in any systematic manner.
16. Key lessons for improving the effectiveness of Senegal’s IMF-supported programs include (i) the need to foster greater ownership of structural reforms; (ii) capacity constraints and measures for improving implementation should be articulated more clearly; (iii) in order to help programs adapt better to uncertainty, the principal risks and how policies would adapt to them should be set out more explicitly; (iv) Bank-Fund collaboration needs to improve to address fundamental institutional/structural issues at the heart of prolonged use; and (v) surveillance discussions and reports should be used as an occasion to “step back” and reconsider the overall strategy, including stress-testing of major vulnerabilities and downside risks, and a discussion of alternative policy options and trade-offs.
Morocco and Jamaica
17. Morocco is one of the few recent prolonged users to have “graduated.” After nine arrangements starting in 1980, it has had no program since 1993. Progress, while eventually substantial, was not smooth. Programs were initially overoptimistic about the time frame needed to restore external viability. There were setbacks from exogenous shocks and periodic slippages in implementation, but policies on core issues continued to move in the right direction, although not always at the pace envisaged by programs, in part reflecting the authorities’ concerns over political feasibility. For the most part, there seems to have been little difference in the IMF’s approach in Morocco compared with other prolonged users. The real difference seems to have been made by strong domestic ownership and political stability, rather than by the structure of conditionality, which was generally exercised flexibly through program reviews. The IMF and the authorities appear to have taken a narrower view of what would justify further program engagement than in some other prolonged users—for example, it is not clear that Morocco’s situation differed greatly at the time of its graduation from that of the Philippines.
18. Jamaica is a former prolonged user that made a decision, after its last EFF expired in 1996, not to seek further financial support from the IMF, although at that time it still faced large adjustment challenges. In particular, the authorities strongly rejected the IMF’s policy advice at the time that a major adjustment of the exchange rate was necessary to restore competitiveness and favored instead a strategy based on running high fiscal surpluses to prevent the debt dynamics from deteriorating further. In the staff’s view, earlier programs had been mainly motivated by the need to obtain foreign financing and debt relief, and the weak ownership that resulted was a major reason why programs had not achieved their objectives. The authorities’ move away from a financial arrangement with the IMF was associated with much stronger domestic ownership of macroeconomic policies, which proved critical to firmer implementation of their chosen strategy. Jamaica remained able to access international capital markets at fairly low spreads, despite the absence of an IMF arrangement. In 2000, staff monitoring of the authorities’ economic program was agreed, and this has played a useful role, allowing room to combine a strong domestic policy formulation process with a healthier policy dialogue with the IMF—while providing a signal to donors and lenders, particularly the multilateral development banks, about the adequacy of the macro-economic framework.
Specifically, recommendations 4 and 9 concern existing procedures, and recommendations 6, 8, 10, and 11 are an elaboration of recent or ongoing initiatives.