Chapter

Chapter 8: Conclusions and Recommendations

Author(s):
International Monetary Fund. Independent Evaluation Office
Published Date:
April 2003
Share
  • ShareShare
Show Summary Details

1. This chapter summarizes the main conclusions of this study on the extent of prolonged use, on the factors underlying it, and on its effects. We then make a number of specific recommendations designed to counter the ill effects associated with this phenomenon. An important caveat is in order at this stage: since the IEO’s mandate is to evaluate the IMF and not the policies followed by its member countries, the emphasis of both our conclusions and recommendations is on the IMF’s role. This focus may give the impression that the IMF should be able to solve all problems in its areas of expertise, if only its interventions and modus operandi could be perfected. Clearly this is not the case. There are obvious limits to what any external agency can achieve, and the primary responsibility for the successes and failures of economic policies necessarily lies with the governments of the countries concerned. This point was emphasized by many officials we met during the evaluation and is the essence of ownership.

Major Conclusions

The prevalence of prolonged use and the nature of prolonged users

2. Prolonged use of IMF resources, regardless of how it is defined, has consistently expanded since the 1970s among both low-income and middle-income countries, in terms of number of countries, share of the IMF’s membership, and the extent of financial exposure. In terms of the number of prolonged users, most of the expansion is accounted for by those eligible for the concessional facilities; however, in terms of financial exposure, prolonged use of the IMF’s general resources is much larger. Furthermore, prolonged use is persistent, with relatively few “graduators.” In addition, arrangements with prolonged users now represent half the total number of ongoing IMF-supported programs.

3. Although prolonged users have attracted a substantial share of both the IMF’s general and concessional resources, they were not a significant constraint on overall lending since the IMF’s liquidity position remained comfortable. However, since decisions on the size of access to IMF resources and on quota increases are endogenous, it is difficult to tell ex post whether prolonged use led to implicit rationing of resources to other users.

Factors underlying prolonged use

4. 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 but it is also partly related to program design and implementation issues.

Systemic factors associated with the role of the IMF

5. There are three major systemic factors that lead to an increase in prolonged use, though their impact in this respect has not been fully acknowledged.

6. A broadening of the rationale for IMF program involvement beyond achieving short-term balance of payments adjustment. The international community increasingly looks to the IMF to help developing countries—particularly the poorest—implement and maintain policies and institutions needed for the achievement of sustainable growth. It looks to the IMF for an assessment of whether policies and institutions are in place that can deliver a sustainable macroeconomic position, and to monitor the situation over time to check that these policies remain on track. In low-income countries, it is also looking for a broader assessment, and subsequent monitoring of progress, on policies to achieve balanced growth over the longer term; many of the aspects of such an assessment go beyond the traditional concept of the IMF as being responsible primarily for short-term stabilization. The fundamental objectives of the IMF, as set out in Article I of the Articles of Agreement, are sufficiently broad that they could encompass such an expanded role. However, this raises the basic question of where a legitimate adaptation of roles ends, and where inappropriate “mission creep” begins. This inevitably involves judgments on the most efficient allocation of responsibilities among institutions and also on whether lending arrangements are the most suitable tools to pursue the above objectives. Many of these questions go beyond the scope of this evaluation, but it is important to consider the potential consequences of the approach that has been adopted for the emergence of prolonged use and its possible adverse impact. We do not believe this has been done sufficiently.

7. IMF lending as a seal of approval and the boundaries between programs and surveillance. One of the factors underlying the expansion of prolonged use is that most official creditors/donors insist on an IMF-supported program as a seal of approval, which becomes a precondition for new adjustment loans and grants or for debt relief and restructuring. There is some evidence that such insistence compromises the quality of IMF-supported programs, and therefore the quality of the seal of approval. With so much riding on the decision, there are strong pressures to agree to a program even though the program may be deficient in several respects. The same tension would of course exist with any other form of seal of approval, but, for reasons discussed below, the association of the seal of approval with repeated programs is especially problematic. Moreover, the signal sent by a short- or even medium-term program may not be the type of seal of approval that official creditors and donors—who usually have a medium- to long-term perspective—should be seeking, especially if it does not ensure the strengthening of core institutions needed for good policies to stay in place beyond the term of the program. In whatever manner the seal is provided, it is important that its quality be maintained. This suggests the need to look for a mix of instruments that provide a seal of approval better suited to the needs of the global community.

8. Choices on where the boundary between programs and surveillance should lie, and on the scope and strength of surveillance processes, will have a major impact on the extent of prolonged use. The preeminence acquired by programs over surveillance in addressing the evolving needs of the international community in a number of circumstances appears to reflect a judgment that only an IMF financing arrangement provides a strong enough vehicle to achieve the desired results.1 However, the nature of the surveillance process itself can be adapted to meet this need. Some of the recent initiatives go some way in this direction. For example, recent efforts to make surveillance assessments more transparent, to sharpen their diagnosis on vulnerability issues, and to promote the observance of internationally agreed standards and codes already provide a potentially stronger instrument for monitoring a country’s progress than existed for much of the period covered by the evaluation. Such initiatives could be expanded further.

9. The expectations of the international community for some form of “seal of approval” signal by the IMF could be met through different combinations of enhanced surveillance, a series of programs, or precautionary arrangements. For low-income countries, this could also involve building on the PRSP process and the need for a positive joint IMF/World Bank staff assessment of each PRSP and PRSP review. It is for the IMF’s members to decide which route they want to take. However, if they wish to continue to rely primarily on a series of programs, then the result is likely to be continued prolonged use of IMF resources for a significant proportion of the membership. This should be acknowledged explicitly and, as discussed below, is likely to have implications for how the IMF organizes its work in such countries. This evaluation also suggests that such an outcome could involve some significant drawbacks. Changes in the nature and modalities of programs can help mitigate these drawbacks, but are unlikely to eliminate them completely.

Program-related factors

The evaluation suggests that a number of program-specific factors have also contributed importantly to prolonged use.

10. Some deep-seated adjustment problems take a long time to fix, even in a perfect world where programs are well designed and implementation is smooth. There is some evidence that the problems of countries that eventually became prolonged users were more severe at the start of their long program involvement. If the IMF is to continue to seek to help countries tackle these problems over an extended period then the challenge is to design programs that recognize from the outset that a longer time frame—with repeat programs and an appropriate division of labor among the IFIs—may be required to achieve lasting adjustment, while ensuring that such a time frame does not become a device to postpone action. Without such a recognition, our evaluation suggests that the short-term focus required by programs may lead to the adoption of approaches that are likely to be ineffective in tackling deep-seated problems because they are not complemented by effective implementation of core institutional and structural changes.

11. Some programs have suffered from design and implementation problems. The country case studies suggest a number of reasons why some programs have been less effective than initially expected in achieving their longer-term objectives. It is not possible to say definitively how much these problems have contributed to prolonged use, since the cross-country evidence discussed in Chapter 5 suggests that these problems also arise in “temporary” user programs. Nevertheless, they are worth noting because they have implications for program effectiveness.

  • Many programs suffer from an overoptimistic time frame—reflecting the difficulty of matching the short-term conditionality of a program to complex, structural and institutional changes that are central to sustained growth-oriented adjustment. There are also institutional incentives to “overpromise” on the speed at which core reforms can be implemented and longer-term sustainability attained.

  • A lack of sufficient emphasis, until very recently, on strong domestic ownership leading to the approval of programs to which governments are inadequately committed.

  • Structural conditionality that was insufficiently focused on key issues. The issue seems to be one of prioritization, rather than the number of structural conditions per se; indeed, programs with prolonged users were not, in general, more burdened with such conditions.

  • Insufficient priority to assessing and improving implementation capacity, and to reforming core institutions so as to ensure that adjustments are sustainable.

  • Insufficient assessment of the real economy responses to the program and to the sources of growth (e.g., leading, in a number of cases, to an overestimation of how rapidly private investment or exports would respond).

  • Absence of a strategy for responding to inevitable uncertainties about the economic environment, which sometimes led to ad hoc corrections that were inconsistent with longer-term objectives.

  • Insufficient opportunities to step back and re-consider the overall strategy pursued by programs while learning lessons from experience.

We are not suggesting that all programs suffered from these problems, or that there was no learning over time. In some of the case studies, IMF-supported programs were associated with significant improvements in the policy environment over time.

Many of the problems listed above have already been recognized within the IMF, and some important initiatives are under way to try to address them. Of particular importance are ongoing efforts to enhance ownership and to “streamline” structural conditionality, so as to concentrate the focus on medium-term, macro-critical structural issues, and to improve collaboration with the World Bank. For the low-income countries, the PRSP/PRGF process is the key initiative to try to embody genuine government commitment, through a broader consensus building process, into the formulation of a medium-term policy framework. However, the jury is still out on how much these initiatives have changed the way the IMF operates in practice and therefore how well they will address core problems.

Is prolonged use a problem?

12. Our evaluation suggests that prolonged use does present problems that were not sufficiently appreciated when decisions were made that were likely to encourage extended program involvement. These problems broadly fall into two categories: potential costs to the prolonged user countries; and adverse effects on the credibility of the IMF.

  • There is some qualitative evidence that prolonged use hinders the development of robust domestic policy formulation processes over time, which partly reflects insufficient attention to country ownership in past programs—al-though it is not possible to test the counterfactual of how institutions would have developed in the absence of lengthy IMF program involvement.

  • There is an inherent tension between the quasi-permanent conditionality implicit in prolonged use, and country “ownership,” in the sense of countries taking responsibility for the conduct of their economic policy, both by being in the driver’s seat and by facing the consequences of their decisions. Recent changes, including the PRSP process, may help mitigate these tensions but are unlikely to eliminate them.

  • Some of the case study experiences also suggest that the perception that IMF resources would be available over the long term, despite policy slippages, may have weakened incentives to take decisive action to deal with some problems.

  • If, as appears to be the case, prolonged use in some cases has resulted from pressures on the IMF to agree to a series of weak programs—for example, to open the way for donor support or debt rescheduling or because of political pressures—then the effectiveness of these programs will be weakened and the credibility of all IMF-supported programs may be adversely affected.

  • Some aspects of the IMF’s independent surveillance functions tend to be crowded out by program activities in the countries concerned, which might reduce the credibility of surveillance.

13. We recognize that some of the adjustment problems faced by IMF member countries, especially the poorest of them, do indeed take a long time to resolve and this justifies somewhat greater acceptance of prolonged use in these cases. Nevertheless, many of the potential costs mentioned above would also be relevant in such cases. Moreover, acceptance of a lengthy program involvement in a significant share of the IMF’s membership would have consequences for the IMF’s role that, in our view, have not been fully recognized. Widespread prolonged use is to some extent inconsistent with current internal operational procedures that are still largely built around the relatively short-term framework of programs. Consequently, there needs to be a clear understanding of what the IMF’s role is expected to be in such cases, so that its operational approach can match that role.

Recommendations

14. In our view, the drawbacks associated with prolonged use are sufficiently serious to warrant a greater effort to reduce its extent; to look for other and better ways to provide the seal of approval that the international community wants; and, where prolonged use still does occur, to look for ways to mitigate its drawbacks since even in cases with “good” reasons for such use, there can be undesirable side effects. Our recommendations seek to address these challenges. Some of them would be applicable only to actual prolonged users. However, any strategy aimed at reducing prolonged use that restricted itself to tackling the problem once it has already materialized would be of limited value. Therefore, many of the recommendations are of more general applicability to the IMF’s approach to programs. In that sense, they can be seen as elements of a preventive strategy for improving the effectiveness of the IMF’s operations and hence reducing the likelihood of prolonged use.

15. The recommendations concern three aspects of the IMF’s operations: (i) the rationale for IMF involvement and the use of its facilities; (ii) program design and implementation; and (iii) IMF internal governance issues.

Recommendations on the rationale for IMF involvement and the design and use of its facilities

16. We recommend that the Executive Board adopt an operational definition of prolonged use, as an essential requirement for evolving a strategy for reducing the likelihood of such use. The evaluation shows that although there are current internal guidelines approved by the Board for dealing with prolonged use cases—calling in particular for the systematic ex post evaluation of programs, for specific justification of IMF involvement, and for a progressive reduction in the size of access in such cases—these guidelines have not always been implemented in a systematic manner. This is partly because there is no formal definition that would identify countries for which prescribed procedures must apply. We fully recognize and strongly endorse the need for flexibility in the decision to provide IMF assistance in individual cases, but the adoption of a formal criterion to identify prolonged users would not eliminate this flexibility. Its purpose would be to trigger automatic due diligence procedures whenever a country meets the prolonged user criterion. The operational definition could be based on the criterion we have used in this study, or indeed some other criterion that is felt to be more useful. The criterion could also distinguish between general and concessional resources, in order to reflect the special circumstances of low-income countries and allow more extended involvement in their case.

17. We recommend that greater efforts be made at judging whether countries are ready to implement credible programs, and that the IMF should be more selective in extending financial support. Many countries may be in a position where they need to make adjustments for which financial support is needed, but they may not always be ready, for complex political and social reasons, to implement the necessary adjustment measures. In these circumstances, IMF-supported programs are likely to be unsuccessful, and the IMF may well need to hold back from providing finance until circumstances are more appropriate. A more rigorous approach to assessing the willingness and ability to undertake adjustment measures and associated reforms should help to reduce prolonged use by encouraging a stronger commitment toward implementation and therefore a more effective adjustment process with a higher probability of graduation. Greater selectivity does not mean that the IMF should play no role where the conditions are not yet ready for financial support. It should (i) actively help to create the conditions for an effective domestically owned program through a frank and transparent policy dialogue, through candid surveillance, and through technical assistance; and (ii) be ready to provide financial and technical support in a timely manner when circumstances are conducive to effective implementation.

18. We recognize that a decision to withhold IMF financial support in such circumstances involves very difficult judgments since it may worsen the economic situation for the country concerned, at least in the short term. We are not suggesting that these implications should be disregarded; they involve very difficult trade-offs between undesirable alternatives that need to be weighed carefully in each case. However, evidence from the case studies shows that there is a strong risk that programs approved under circumstances in which credible action is unlikely, for whatever reason, will only postpone the resolution of problems, perhaps even allowing them to get worse, without offering any good prospect of sustainable adjustment. In such cases, raising the threshold for what is required for IMF support, especially in terms of the probability of implementation, is likely to yield a better outcome in the long run. In calling for greater selectivity, we are not seeking programs that look “tougher” on paper; indeed, an essential counterpart of greater selectivity is that programs focus only on what is essential for longer-term sustainability and have a realistic time frame (see next section).

19. Operationally, this means that UFR proposals to the Board should contain an explicit and frank assessment of the readiness of potential borrowers to implement programs. Current guidelines already call for a judgment by management that the program will be carried out, but the evidence from the case studies indicates that the assessments on which this judgment is based are sometimes done in a perfunctory manner. Few such judgments are likely to be totally clear-cut, but the Board should be provided with a candid assessment of the risks.2

20. We recommend that the IMF aim to provide the international community with credible alternatives to the current situation where IMF lending arrangements have become a precondition for many other bilateral and multilateral flows. It is up to each donor and creditor to decide the conditions on which they will provide financing, and all legitimately want assurances that an appropriate policy framework is in place to make their financing effective. However, the requirements for effectiveness of different types of flows are different. For many longer-term flows these requirements could be met without always having an IMF-supported program if suitable alternatives are developed, such as greater use of strengthened surveillance, reliance on joint staff assessments of PRSPs, shadow programs, and precautionary arrangements. More generally, we recommend that the IMF should aim at developing a mix of tools that could serve to deliver a seal of approval in different ways, depending on the member’s circumstances (in particular, its eligibility for the PRGF and for the HIPC Initiative), donor/creditor requirements, and the strength of the member’s policies and institutions.3 Such forms of enhanced surveillance may even have an advantage over lending arrangements since they can have a longer-term and broader focus, covering all elements of a country’s economic strategy.

21. For PRGF-eligible countries, one possible approach could be to design a form of enhanced surveillance (once a degree of macroeconomic stability is restored), in order to provide a clear signal on the appropriateness of the macroeconomic framework and a monitoring of progress over time on both macroeconomic performance and on macro-critical structural reforms outlined in the PRSP. In effect, the IMF would start from the country’s own medium-term program as outlined in the PRSP, and assess it (in cooperation with the World Bank in aspects of policy and institutions that fall outside the IMF’s core areas). Thus, it should be made clear that not every PRSP would be accompanied by an IMF arrangement under the PRGF.4 In whatever manner this was done, it would clearly be necessary to incorporate into surveillance reports and Board summings-up an overall judgment that a country’s economic strategy is sustainable and has good prospects of achieving the desired objectives, to give a clear signal to the donor community.

22. We recommend that programs for identified prolonged users should include an explicit “exit strategy.” This is already called for in current guidelines, but is often not implemented, possibly because of the lack of a specific definition of prolonged use. The details of the exit strategy would vary from country to country—and especially between PRGF-eligible countries and others. An important feature of any exit strategy would be for the IMF to reduce progressively its own resource contribution (which again is already a feature of current guidelines) while remaining actively engaged, along the lines discussed above, in providing the “seal of approval” that may be needed for other donors and creditors to maintain their flows. Such an approach would enable the Board to be more proactive in identifying cases where a scaling back of IMF program involvement would be appropriate.5

23. We recommend the introduction of a differentiated rate of charge for prolonged users as a signaling device. We do not support formal restrictions on the duration of prolonged use because all member countries should have the ability to access IMF resources if the need is justified, and formal time limits would ignore both this right and the variability of country circumstances. However, there is a case for a differentiated rate of charge for prolonged use exceeding some limits. We recognize that there is no evidence that the cost of IMF resources has been a significant factor in determining prolonged use, but the introduction of such a charge could serve as a signal of excessive dependence on the IMF and possibly provide a political incentive to avoid such prolonged use.

Recommendations for program design and implementation

24. Program design and implementation issues have been extensively discussed in the Executive Board on many occasions, and several of the recommendations in this section essentially consist of a reaffirmation of what are supposed to be existing guidelines or “best practices,” but which our evaluation suggests have not always been implemented on a consistent basis. In the recommendations on internal governance processes in the subsequent section, we make a number of suggestions on how internal incentives and procedures could be improved to encourage improved implementation and strengthened learning processes. Many of the issues discussed here have also been the subject of recent initiatives by the IMF; in these cases, we make a number of additional suggestions that could help increase their effectiveness.

25. We recommend that specific operational procedures be developed that will ensure that program design places greater emphasis on the nature of the domestic policy formulation process, in order to maximize ownership. The IMF has already recognized the importance of promoting ownership,6 most notably in the PRSP process and in the ongoing re-view of conditionality, and this is the right direction in which to move. A number of steps could be taken to fulfill this objective:

  • (i) The IMF should modify its procedures to move toward a situation in which the authorities have the initial responsibility for proposing a reform program. Ideally, this could be done by having the initial request seeking an IMF arrangement take the form of a letter of transmittal of a domestic policy document outlining the broad approach of the authorities. This should be the starting point for negotiations. We are not suggesting that such an approach be an additional prerequisite for an arrangement, since differences in administrative capacity will affect the pace at which countries are able to take the lead in formulating programs. We understand that some countries already adopt an approach close to what is being proposed, whereas in many others the IMF staff takes the lead. Technical assistance could be provided to help build the capacity for policy formulation where needed.7 Clearly, the submission of a domestic policy document would only initiate the process. It would still be the IMF’s responsibility to assess the proposed program to determine whether it has a good chance of achieving its objectives and to negotiate strengthening the program where needed. The specific structure of conditionality would emerge from the negotiations but could then be viewed as concrete commitments under-taken within the broad framework proposed by the authorities. Such an approach would help to ensure greater ownership of the broad directions of the program. We recognize that it would require greater flexibility in scheduling missions, which should be tailored to the timetable of national policy agendas, and may mean that negotiations take longer to complete. If a balance of payments crisis is imminent, there may be less time for detailed policy formulation in advance of a mission, but most program missions for medium-term programs—especially in the prolonged users—do not take place in such an environment.

  • (ii) The IMF should, wherever possible, encourage a process whereby the core elements of a program are subject first to a domestic policy debate within the member country’s own policymaking institutions.8 While the nature of this policy debate will vary with the particular institutional circumstances of each country, two general messages are relevant: (a) high levels of political authority need to be fully engaged; and (b) the more transparent and participatory the process the better.

  • (iii) High-quality surveillance should help to create a better understanding of what would be expected of the authorities should a program become necessary. This would contribute to the perceived transparency of IMF policies in the design of programs.9 Surveillance reports should, therefore, actively seek to present alternative policy options and to analyze the trade-offs between them: this is already “best practice” but it is not general practice.

26. We recommend that programs give much greater emphasis to fostering key institutional changes and to strengthening implementation capacity. Our assessment based on the case studies and questionnaire responses from prolonged users’ authorities suggests that strengthening the institutional base for implementing reform is a much more important determinant of the long-term success of programs than the detailed structure of conditionality. Staff program documents should include an explicit assessment of the key institutional requirements for effective implementation and how these can be strengthened. As already called for in the most recent procedures, they should also be clear about the division of institutional responsibilities between the IMF and the World Bank (see below), and reporting the Bank’s assessment of institutional constraints in those areas where it is in the lead.

27. We recommend greater selectivity in program content along with further improvement of collaboration with the World Bank, a more differentiated use of conditionality, and a broadening of the time frame of program design. This recommendation is in line with ongoing initiatives to streamline conditionality, which we strongly welcome. In our view, the thrust of streamlining should not be primarily on the quantity of conditionality per se but on improving its pri-oritization and integration with program design, which should then be reflected in a more parsimonious recourse to waivers at the stage of implementation—in other words, picking the battles better and fighting them well.10 The experience of the case studies indicates that, if the overall volume of conditions exceeds implementation capacity, some conditions—not necessarily the most important ones—may be effectively implemented, but others will not. In keeping with the spirit of the draft revised guidelines on conditionality, the IMF should identify those issues that are truly critical to sustainable macroeconomic adjustment and then focus on them in depth.11 In our view, effective implementation of these principles would require a number of operational changes:

  • (i) Making a more differentiated use of the various modalities of conditionality. Conditionality should be seen primarily as a tool to help focus on critical areas that need concentrated attention, as well as an instrument of mutual accountability between the IMF and the authorities. Structural conditions should focus on aspects of the program that are critical for sustainable adjustment, but it is equally important that all such aspects be addressed by the program. This was not always so in the programs examined in the case studies. The mix of various modalities of conditionality should reflect the order of priority attached to each measure by the authorities and the IMF as well as the planned sequencing of reforms, both of which should be explained in staff reports to the Board.12 Furthermore, our analyses, especially in the case studies, also suggest that the substance of conditionality matters more than its formal structure. Particular efforts should be made, when negotiating conditionality with the authorities and when assessing compliance, to put the emphasis on actions that will ensure substantive progress toward meeting the program’s objectives rather than on formal compliance with narrowly defined conditions on a checklist. We recognize that this would require greater flexibility and judgment.

  • (ii) Making greater efforts to tailor the effective time frame of program design to the foreseeable length of the reform and adjustment process. This does not necessarily imply that the time frame of IMF arrangements should be further lengthened. Indeed, experience suggests that country authorities themselves are often reluctant to commit themselves firmly for long periods—partly reflecting political uncertainties. However, our study has shown that where a long-haul adjustment effort is required, it is at best ineffective and at worst counterproductive to try to force adjustment within a shorter, and essentially arbitrary, time frame. One way to proceed would be to design a medium-term strategy for IMF involvement, covering the full length of the required reform and adjustment process. The strategy would build on domestic documents (i.e., PRSPs where they exist, planning documents, programmatic laws, “lois cadres,” etc.) and indicate the nature of IMF involvement, including through successive programs, in different stages. The strategy should lay out several key elements: what the objectives of this involvement are; what combination of lending, including possible repeat programs, policy advice, and technical assistance are envisaged to achieve them; and also what exit strategy would be followed. Whether individual arrangements are signed and funds disbursed would continue to be guided by the same policies as before. We do not propose preparation of separate Board documents outlining the strategy: UFR request reports should be used to set out the proposed strategic framework, while surveillance reports or program reviews should update and monitor progress against them. Our proposal is therefore essentially a further strengthening of the approach that is supposed to be used with internal country strategy papers, but with the central elements of the proposed strategy—and sub-sequent assessments and possible reappraisals—conveyed to the Executive Board.13

  • (iii) Further strengthening collaboration with the World Bank. Initiatives such as the agreement on the “lead agency” concept are useful first steps, but ensuring that the new approach works effectively is likely to require deeper operational changes and sustained emphasis by management.14 Where the World Bank does not appear to be in a position to deliver the necessary complement to the program, staff reports should be candid about the issue and let the Board decide to what extent the IMF should concern itself with those issues. To en-courage such candor, the traditional appendix to staff reports describing the country’s relations with the World Bank—which at present is typically pro forma and adds little of substance—could be replaced by a more substantive discussion of the World Bank’s strategy in the country, showing how it complements that of the IMF and flagging any significant differences of view or areas where the two institutions’ strategies are not fully integrated. However, the case studies suggest that meshing the approaches and time frames of the two institutions will be an enormous challenge.

  • (iv) Systematically incorporating more in-depth analysis of real economy responses to the key policy elements of programs, and the sources of growth and devoting proportionately less attention—and staff time—to the fine-tuning of the financial programming exercises.15 Ideally, such analyses should be conducted regularly in the context of the IMF’s surveillance activities—drawing where appropriate, on the expertise of the World Bank and other institutions—and the accumulated knowledge should be used at the time of programs involving UFR. This will help to avoid building programs around unrealistic expectations, especially as regards exports and tax revenue growth. Moreover, using the surveillance process to flag candidly weaknesses in the statistical base would help reduce the enormous amount of time that negotiating missions spend addressing data problems.16

28. We recommend that programs include more explicit discussion of the major uncertainties they face and of how policies will be adapted if underlying assumptions turn out differently. These problems are not new and are also not exclusive to prolonged use cases, but they are specially relevant for these cases as they would help to reduce the risks of programs being repeatedly blown off-track by the materialization of downside risks. It is not possible to pre-specify ways of dealing with all contingencies. However, the program review process needs to be sufficiently flexible to adapt the program in a timely manner when circumstances change, and an early under-standing on the major uncertainties and proposed responses, even in general terms, can help this process. For example, in cases where other forecasts (e.g., private sector “consensus” forecasts) of growth or of exports differ significantly from those assumed in the program, staff reports should discuss in concrete terms how program design would be modified should these other forecasts prove to be more accurate. Such a requirement would also help reduce the risks to programs caused by overoptimistic forecasts.

Recommendations for internal IMF governance processes

29. We recommend that systematic ex post assessments of programs be undertaken, with priority given to identified prolonged users, and the key messages reported to the Board. Such assessments should be part of a broader effort to disseminate more effectively “best practices” and lessons learned, and to maximize the effectiveness of the review process. Internal assessments of each program that is completed or permanently interrupted would help to ensure that the lessons for program design are absorbed more quickly and systematically. The country case studies highlighted a number of occasions where potential problems with program design (or its implementation) that contributed to prolonged use were identified, but where the lessons were not fully absorbed in the design of subsequent programs. Indeed, as is the case for several of our other recommendations, this is another example where existing guidelines require such action but they have not been implemented. The discussion in Chapter 6 also suggested that in a number of areas the IMF has been quite good at identifying lessons, but less effective at ensuring that they were absorbed into everyday operations. Building the review process around systematic ex post assessments will foster better implementation of “best practices” and provide clear opportunities for reconsidering the overall strategy.

30. One of the reasons why previous calls for more systematic assessments of programs have not been implemented was excess work pressure, and we recognize that implementation of this recommendation will require some additional resources. If necessary, the requirement for such ex post assessments could be phased in, with priority given to existing prolonged use cases. We also think it is important that the key elements of any debate on program design options that takes place during this process is conveyed to the Executive Board. Clearly, care must be taken not to hamper the ability to deliver a consistent message to the authorities, but the credibility of that message will ultimately be strengthened if it is seen to be derived from a process that considers different options and is designed to learn from experience. In this regard, we also recommend that:

  • (i) Staff reports—especially those involving requests for new arrangements for prolonged users—should provide more of a perspective of the history of the IMF’s program involvement with a country. This should highlight what has been achieved and where previous strategies have fallen short of their objectives and why.

  • (ii) The MONA database, which is the key internal information system for tracking performance under programs, should be made more comprehensive, accurate, and up to date. At present, the MONA database does not include information on programs that go off-track—even though these are the very ones that should be followed most carefully.17 During the course of the evaluation, substantial errors and gaps in the database were also discovered, especially with regard to data on outcomes. Existing weaknesses in data on how programs have performed are an impediment to efforts to enhance the IMF’s ability to learn from experience and to monitor the implementation and impact of its own policies. We have discussed this issue with PDR staff and we understand that efforts along these lines have now been initiated. We also recommend that the MONA database be made accessible to outside researchers, in order to en-courage further analysis and feedback on program successes and failures.18

31. We recommend that steps be taken to further strengthen surveillance in prolonged use cases. Evidence from the case studies suggests that some important aspects of surveillance have been weakened when undertaken in a prolonged program context. In countries where programs are temporary events, such a “crowding out” of surveillance by program activities could be less important, but its consequences are potentially more significant for countries that have a long series of IMF arrangements. We recommend the following steps:

  • (i) The surveillance guidelines should be modified to clarify the expectations on the role of surveillance in program cases. The draft guidance note discussed by the Executive Board in July 2002 already takes a step in that direction by highlighting the need, in program countries, for surveillance to bring a fresh perspective by providing: “(i) a comprehensive assessment of economic developments, beyond the narrow focus of program targets; (ii) a candid analysis of short and medium-term prospects, including a thorough discussion of risks and vulnerabilities; (iii) a stock-taking of the policy strategy to date and the effectiveness of the measures implemented in pursuit of that strategy; and (iv) a candid account of the dialogue between the staff and the authorities on the key policy issues and the strategy looking ahead.”19 As part of the discussion of risks and vulnerabilities, a useful addition to these requirements, in prolonged use cases, would be the presentation of a policy slippage scenario, to illustrate what implications the country and the IMF would be faced with should the current program go off-track. Furthermore, as discussed above, the surveillance reports for prolonged users should be used as an opportunity to encourage a frank and open debate on the IMF’s overall strategy in a country. Surveillance reports in program countries should also make a special effort to incorporate the views of the World Bank on those segments of the policy agenda where it is in the lead, and to report candidly about the quality of Bank-Fund collaboration in the country.

  • (ii) There is a case for creating some greater institutional separation between programs and surveillance, especially in the context of prolonged use. The International Monetary and Financial Committee in its April 2002 communiqué called for a “fresh perspective and appropriate distance” in the conduct of surveillance. This is particularly important in prolonged use cases. Giving these expectations more emphasis in surveillance guidelines and, as discussed in the previous recommendation, embedding them in a systematic ex post assessment process, would go a considerable way in this direction.20 One additional step that has been suggested would be to have entirely separate teams to conduct the two activities, but this would involve substantial resource and coordination costs for both the IMF and the authorities, and signals on policies might also become confused. However, there is merit in taking some institutional initiatives to achieve greater separation between surveillance and program activities for prolonged users. At a minimum, surveillance reports should not be treated as offshoots of program activities. For instance, the internal review process should deal with surveillance reports for countries under program in exactly the same way as other surveillance reports, which is not the case at present.21 An option that might be considered in a limited number of cases, such as the most prolonged users, is for the chief of the surveillance mission to be chosen from outside the relevant area department.22 We recognize, however, that making the greater separation operational in practice involves delicate trade-offs, in particular in terms of continuity of the policy dialogue and country knowledge management. These trade-offs would have to be appropriately managed.

  • (iii) In the same spirit, 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. Discussions with the staff and internal documents reviewed by IEO make clear that there can be considerable debate within the IMF (as well as between the staff and the authorities) on key policy options, and that this debate often draws upon outside analysis, including through informal contacts and seminars. However, the analysis presented in final reports submitted to the Board is often designed to support the final, agreed position at the expense of understating the extent of trade-offs between different strategies. One possible approach to improve on current practices would be to experiment with including “second opinion” analysis from outside sources in the selected issues paper prepared for Article IV consultations, along with any staff response. The focus of such analysis would be on critical issues where there is a wide divergence of views on the appropriate approach.23

  • (iv) The precise frequency of Article IV consultations with program countries is less important than that they take place at an appropriate time— that is, when a “fresh look” would be most valuable. It is especially important to have timely consultations when programs are faced with unexpected challenges, when they go off-track, or before a new program is negotiated.24

32. We recommend strengthening the ability of IMF staff to analyze political economy issues so that a better understanding of the forces that are likely to block or enhance reforms can be taken into account in program design.25 Although it is widely recognized that ownership and political and social feasibility are crucial for effective implementation and sustainability of reform, too little attention is often paid to these aspects in program design. This is a complex area, and it would be unrealistic to expect the IMF, or for that matter any external agency, to do too much, since it is ultimately for governments to determine what sets of policies would be acceptable to their societies and to increase the acceptability of desirable reforms. Excessive involvement of external agencies in this area would itself be contrary to the whole idea of domestic ownership. The paper prepared as background to this evaluation sets out, by way of an illustration, a number of tools and proposes a series of basic questions that could be asked in trying to judge the political feasibility of a program.26 To a large extent, using such questions to guide a basic assessment before supporting a program would only bring discipline and consistency to analyses that are already carried out in best practice cases, though their conclusions are not always reported to the Executive Board. More systematic assessments might also be commissioned in cases where political feasibility has been a major obstacle to program implementation. If more in-depth assessments were undertaken, priority should be given to the most prolonged use cases.

33. Finally, there are two other internal governance issues that surfaced in the context of this study and, although not exclusive to prolonged users, deserve some consideration in view of their seriousness and potential aggravating effects on prolonged use.

  • We recommend that a review of (explicit and implicit) internal incentives facing staff be under-taken with a view to minimizing turnover of staff working on countries and to fostering increased candor and accountability. Excessive staff turnover—between departments but also between different country assignments within the same department—appears to be a widespread problem. Although not peculiar to prolonged users, such excessive turnover is particularly detrimental in their case. A revamping of internal personnel incentives to encourage greater stability is needed.27 The focus of these incentives should be tilted toward encouraging the development of a deeper familiarity with the problems of individual countries, and correspondingly increasing responsibility, through longer country assignments rather than just acquiring the minimum necessary experience and moving on.28 Furthermore, the questionnaire of mission chiefs, whose results have been discussed in Chapter 6, revealed that existing incentives, as perceived by mission chiefs, do not sufficiently encourage realism and candor, nor do they foster accountability. Efforts should be made to identify the source of these perceptions and, to the extent possible, correct them.

  • We recommend that procedures be evolved that will help avoid the appearance of political intervention in the IMF’s determination of whether programs are deserving of support. Political considerations are unavoidable in an institution governed by the votes of its shareholder governments. However, these considerations should be taken into account in a transparent manner—with decisions and accountability clearly at the level of the Executive Board. As discussed in Chapter 6, the process by which political considerations are currently handled in the IMF’s decision-making process is inadequate, and this could affect the credibility of programs and thereby occasionally contribute to prolonged use. While it is reasonable for the Managing Director to take account of shareholder concerns about systemic trade-offs when deciding what risks are acceptable, the present approach has two problems. First, there is no formal—hence no transparent—channel through which political judgments can be fed into the process before the final approval stage. Second, the line of accountability between staff, management, and the Board can, in practice, become blurred. The problems can be mitigated through greater transparency, to which two operational changes could contribute: (i) requiring all program presentations to the Executive Board to be prefaced by an explicit assessment of implementation risks and (ii) when management judged these risks to be high, giving the Board an opportunity to express—on the record—its own judgment on the trade-offs involved before the program was presented for approval, based on a candid assessment of these risks and of the implications of withholding IMF support.

34. Implementation of some of the recommendations would itself raise significant organizational issues. Where our evaluation has provided some insights as to how these implementation issues might be addressed, we make specific recommendations to that effect. However, we have not attempted to spell out operational details in all areas, and we recognize that further work would be needed to translate some of the recommendations into fully operational solutions.29

35. Several of the recommendations have resource implications. Some will clearly involve greater staff inputs, most notably those involving ex ante assessments of ownership and implementation capacity, ex post assessments of programs, and the provision of technical assistance. Others mostly involve a ratio-nalization of current practices which, through greater focus and selectivity, should contribute to staff resource savings. In addition, to the extent that all these recommendations succeed in reducing the scope of prolonged use, the ensuing decline in the size of UFR activities would also eventually reduce the current excess demands on IMF staff time. Although it is difficult to quantify the overall impact of the recommendations, we anticipate that they would probably involve an overall resource increase in the short term, with some reduction possible in the longer term as the scope of the IMF’s involvement in a number of prolonged users is reduced. But the most critical question from the perspective of long-term resource implications will be intensity of the IMF’s involvement in those countries where achievement of sustainable, growth-oriented adjustment is inevitably going to be a protracted process.

As noted in Chapter 6, some donor respondents to the questionnaire suggested that one of the reasons why such an endorsement was taken more seriously was precisely because it involved a financial commitment by the IMF. But such an approach would by its very nature imply prolonged use.

It is difficult to prespecify exact criteria that the Board should use in making such judgments, but the case studies suggest a number of examples where it would have been better if the IMF had been more restrained in entering into or extending programs—such as the Philippines during much of the Marcos and Estrada administrations (once the extent of governance-related problems become clear) and in Pakistan for parts of the 1990s.

One recent example of a development in this direction is the agreement concluded in July 2002 between Jordan and the Paris Club: in effect, Paris Club creditors agreed to a consolidation period more than twice as long as Jordan’s SBA and decided to rely on Executive Board discussions on post-program monitoring and Article IV consultations to assess Jordan’s performance after the expiration of the SBA as a basis for deciding on the entry-into-force of the annual phases of the debt rescheduling agreement. This example, along with donors’ responses to the IEO questionnaire on the topic (see Chapter 6), suggest that the scope for flexibility in the mix of tools to deliver the seal of approval should be explored further.

Joint IMF–World Bank staff assessments of PRSPs are typically concluded by an assessment of whether or not the PRSP in question constitutes an adequate basis for IMF and World Bank concessional lending. These assessments could conceivably be tailored to donors’ concerns in such a way as to allow them to extend the concluding judgment to concessional lending from other sources, thereby making an IMF lending arrangement redundant for seal of approval purposes.

The discussion that took place in the Executive Board at the time of the fifth review of the Philippines’ 1998 SBA—which triggered a wide-ranging internal discussion of future strategy—is one good example of such a proactive approach.

See, for example, IMF (2001d), available on the IMF’s website.

Many officials from the three country case studies and the questionnaire responses emphasized that IMF technical assistance had generally not been very effective in helping countries develop the capacity to design and, especially, implement economic policies—which are essential elements of ownership. Participants in the 2001 external consultation process on conditionality made a similar point.

A robust domestic policy formulation process does not necessarily mean near-universal consensus or a requirement to consult nongovernmental groups in a particular way; it just means that the main policy elements of a program would carry sufficient support in the core political institutions, including parliaments. The extent to which this is possible will depend, inter alia, on how urgent is the need for IMF financial support.

We recognize of course that in the event of unexpected developments leading to crises, programs may need to introduce measures not envisaged in surveillance. Nor should programs necessarily undertake to fix every single problem diagnosed in the surveillance process if there are higher priorities.

Indeed, the data discussed in Chapter 5 indicate that prolonged users’ programs on average had less extensive conditionality than “temporary” users. Yet they also had a higher proportion of waivers, which were often followed by serious program interruptions. Thus quantity is not the critical issue.

At the time this evaluation was completed, the Executive Board was in the process of completing the review of conditionality initiated in 2001, by approving a revised version of the “Guidelines on Conditionality” adopted in 1979.

For example, prior actions should generally be limited to measures whose absence at the start of the program would jeopardize its chances of success, and that can effectively be put in place in a short time frame. Any measure that does not meet either of these criteria would be more effectively dealt with through otherforms of conditionality. In particular, prior actions on measuresthat are meant primarily as tests of the authorities’ owner-ship but the adoption of which, in and of itself, has little macro-economic impact, should be avoided. If there are serious doubts about ownership, it is better to wait until some credible track record is established rather than devise tests through prior actions that are not critical to success. In intermediate situations, the actual implementations of existing guidelines on prolonged use related to the front-loading of the adjustment effort and the back-loading of disbursement could serve to mitigate implementation risks without jeopardizing ownership. (Data reported in Chapter 5 suggest that these guidelines were often not followed.)

See Chapter 6 for a discussion of the role of the country strategy papers.

Since 1989, there have been approximately ten reviews and progress reports on IMF–World Bank collaboration, all of which diagnosed room for improvement and put forward remedies. This record suggests that the underlying problems are complex and deeply rooted.

When we refer to the excessive amount of time devoted to the excessive fine-tuning of the financial programming exercise, we are not implying that IMF staff does this out of a misplaced sense of priorities, or that simple exhortation will correct the problem. Indeed, it is the staff themselves who have most emphasized this issue during our discussions. Rather, this issue is another example of the tensions between the short-term framework of a program and the more important, but often less precisely defined, longer-term goals. For example, the rather rigid formal framework for quantitative performance criteria—which requires Board approval of waivers whenever a deviation occurs, no matter how small—tends to raise the stakes for even minor deviations, since countries are often reluctant to be seen to request waivers. The recent heavy focus on cases of misreporting has added to these pressures. There is no easy solution to such problems, since a quantitative monitoring framework is justified, but it would help if Board papers were franker about the margins of uncertainty surrounding the details of program design.

This is already supposed to be current practice. Area departments are expected to work closely with the IMF’s Statistics Department in developing strategies to remedy data deficiencies and to enhance statistical capacity, including through a prioritized use of technical assistance.

Apparently, the reason why “off-track” programs are not monitored comprehensively in MONA is because it is the completion of the review that triggers the administrative process to update the database. No action is taken until the review is completed, so programs that are permanently interrupted cease to have their database updated.

We understand some researchers have already been given access to the database, on a case-by-case basis, with safeguards to protect confidential data.

See SM/02/184 of June 14, 2002, “Biennial Review of the Implementation of the Fund’s Surveillance and of the 1977 Surveillance Review—Follow-up.” As regards stock-taking, the 1997 guidelines on country strategy papers discussed in Chapter 6 constitute a good description of what “stepping back” should aim to achieve.

Indeed, a few examples from the country case studies suggest that the country teams under the existing arrangement do have the ability to step back and take a frank look when circumstances permit (e.g., when there is not a strong incentive to avoid “rocking the boat” on an already agreed program).

Thus, within PDR, the Surveillance Policy Division should have the primary role, not—as is currently the case—divisions in charge of reviewing programs.

Implementing this recommendation would require particular precautions to ensure the adequate preparation and follow-up of Article IV consultation missions, allowing a suitable feedback to program discussion. The suggestions made by the OIA in its 2001 report on the organization and management of country missions, if implemented in these cases, would go a long way toward addressing the most critical issues in that respect.

One example of where such an approach might have been fruitful is the debate between the Jamaican authorities and the staff on the appropriate exchange rate and monetary framework in the late 1990s.

This is the spirit of the decision approved by the Executive Board in July 2002, which shifts countries under program to a 24-month Article IV consultation cycle instead of the standard 12-month cycle. It is important that this decision be implemented faithfully to its spirit—that is, to ensure that surveillance takes place in a timely manner when needed—and not mechanically, as the latter might result in a further weakening of surveillance in program countries.

It has been suggested that the IMF should hire full-time political scientists to undertake such tasks, but a potential problem with such an approach is that a few such specialists would not be integrated into the negotiating process and would risk being marginalized. As a minimum, efforts should be made to enhance the training of IMF staff on the various political science tools that can be used to analyze the feasibility of policy reforms. In addition, as noted in Chapter 6, the views expressed in our interviews in the case study countries on the role of the resident representatives was very positive, with many officials and other stakeholders expressing the view that they should be provided with greater scope to provide input on the feasibility of particular proposals—an approach that is already supposed to be “best practice.”

See Appendix 1 to the Pakistan country study (Chapter 9) for a further discussion.

The only formal personnel requirement for intradepartmental mobility is that, to be promoted to the B-level (i.e., grades with greater management and supervisory responsibilities), an economist should have worked in at least two departments; otherwise, she or he can only be promoted to a B-level position outside their current department. However, there is ample anecdotal evidence that internal incentives strongly encourage mobility for IMF economists (see, for instance, “Review of Personnel Management in the Fund,” OIA, February 2000).

The internal “Economist Development Guide” recently prepared by the IMF’s Human Resources Department is a step in this direction.

In a few areas, recent reports by the IMF’s Office of Internal Audit and Inspection (OIA) touch upon some of these organizational issues and we have made reference to these reports where they appear relevant.

Annex 1: Possible Definitions of Prolonged Use

This annex describes various approaches to defining prolonged use that have been used previously in the IMF or elsewhere and presents more details on the evolution and persistence of prolonged use.

The precise definition resulting from each approach can be made more or less restrictive by varying the threshold that separates prolonged users from “temporary” users of IMF resources.

  • (i) Prolonged effective use of the IMF general resources1 (IMF, EBS/00/187)

This concept focuses on resources borrowed under stand-by and extended arrangements and excludes programs financed from concessional trust funds (SAF, ESAF, PRGF) for low-income countries, as well as programs in which the financing approved is not fully disbursed, either because they are off-track (i.e., the country is not eligible to borrow) or because they are treated as “precautionary” by the country’s authorities. This is the narrowest of the possible approaches and risks excluding important issues, such as the implications of failed/interrupted programs and the IMF’s role in low-income countries.

  • (ii) Prolonged time spent under IMF-supported programs (IMF, SM/84/91 and EBS/91/108

This concept encompasses programs funded both from the General Resources Account and from concessional trusts. It also includes programs that are only partially drawn upon. It may or may not include precautionary arrangements. It does not include drawings on IMF resources not backed by programs (such as first credit tranche purchases).2

A slightly different version of this concept is used by Bird, Hussain, and Joyce (2000) to characterize frequent users of IMF resources. Their definition is based on the number of programs adopted by a country during a particular period, regardless of the type of arrangement at stake, its treatment (i.e., precautionary or not), its duration, or its degree of completion. However, because many programs have a multiyear time frame, particularly those under the EFF and PRGF, such a definition does not measure the time spent under IMF arrangements.

This concept focuses on the length of periods of indebtedness to the IMF, regardless of the origin of the outstanding obligations.3 However, because IMF facilities have repayment periods varying from 2½ to 10 years, this definition does not distinguish between countries that had only a few arrangements with relatively long repayment periods and those that had a large number of arrangements with shorter maturities. An interesting application of this approach was used by Jeanne and Zettelmeyer (2001) to derive estimates of the length of “lending cycles” to particular countries (Annex Table 1.1).

Annex Table 1.1.Completed and Incomplete Debt Cycles for Borrowers from the IMF, 1947–2000
Average duration of

cycles (years)
Number of

countries
Incomplete

debt cycles
CompletedIncomplete
All countries186887.117.9
Industrial countries2504.7n.a.
Developing countries161887.617.9
Africa52386.122.7
Asia2913921.2
Europe282110.27.9
Middle East1426.59.5
Western Hemisphere37147.618.1
HIPC countries142386.123.5
Non-HIPC developing countries11950813.6
PRGF countries280589.320.6
Non-PRGF developing countries81308.212.7
Prolonged users (PU)344417.322.3
Non-PU developing countries11747914.1
EMBIG countries427157.813.8
Non-EMBIG developing countries134737.618.8
Memorandum Item: excluding cycles initiated after 1991
HIPC countries142356.124.9
Non-HIPC developing countries119228.223.3
PRGF countries280439.325.6
Non-PRGF developing countries81148.620.4
Prolonged users (PU)44357.324.7
Non-PU developing countries117229.523.7
EMBIG countries32787.920.6
Non-EMBIG developing countries134497.824.9
Source: Database assembled by Jeanne and Zettelmeyer.Note: This table is an adapted and expanded version of one shown in Jeanne and Zettelmeyer (2001). “Complete” and “incomplete” debt cycles refer to cases where a member has borrowed from the IMF and where the subsequent obligation to the IMF has eventually fallen to zero (“complete” cycle) or where further borrowing meant that the obligations to the IMF have not yet fallen to zero (“incomplete” cycle). The sum of complete and incomplete cycles exceeds the number of countries because each country may experience several lending cycles. The sum of HIPC, PRGF, PU, and EMBIG countries exceeds the total number of countries because these categories overlap in part.

As noted in the main text, the current evaluation project uses a definition based on the amount of time spent under IMF arrangements, whether or not a country was eligible to draw. In principle, a distinction could be made between continuous “prolonged” use and more episodic “repeat” use. These episodic users may have interludes when their balance of payments situation improves and they begin to repay the IMF, but such episodes are followed—perhaps as a result of intervening policy slippages—by further balance of payments problems and recourse to IMF financing. Prolonged users would encounter few such episodes of IMF “abstinence,” perhaps reflecting incomplete adjustment within the life of a program or longer-term debt sustainability problems that were not adequately addressed up front. In practice, however, it is not possible to make such a clear-cut distinction: all such countries appear to have experienced interludes when their external position improved, followed by renewed difficulties.

Annex Table 1.1 and Annex Figure 1.1 provide further details of the intensity of prolonged use and its evolution to supplement the discussion in Chapter 2.

Annex Figure 1.1.Frequency and Duration of Recourse to IMF-Supported Programs Across the Membership, 1992–01

Sources: IMF Policy Development and Review Department databases and IEO calculations.

Note: IMF members that did not enter into an arrangement with the IMF over the period are not represented in these figures. The number of prolonged users (PUs) and “temporary” users (TUs) correspond to the cumulative number of programs over a rolling 10-year time frame.

Notes

That is, purchases from the General Resources Account (GRA), which are typically associated with a Stand-By Arrangement (SBA) or an Extended Fund Facility (EFF) arrangement. The specific operational definition used in the 2000 review of prolonged UFR characterized as prolonged users countries with an outstanding use of IMF credit over 100 percent of quota and either 9 years or more of effective UFR in the previous 30 years, or 5 years of effective use in the previous 15 years.

The thresholds used in internal IMF definitions have varied over time: in 1984, it was set at four or more programs with purchases in the previous 10 years; in 1986 and 1991, it was raised to five annual arrangements in the previous 10 years. In all cases, an additional criterion was an outstanding IMF credit of over 100 percent of quota at the end of the period under review.

This concept was used to define prolonged users in a 1986 internal IMF review, with a threshold of “continuously outstanding credit tranche positions in excess of 25 percent of normal maximum for six years or more” in the previous 10 years.

Annex 2: Background Material on the Evolution of IMF Policies on Prolonged Use

This annex provides additional information to support the discussion in Chapter 3.

Evolution of the IMF’s Attitude to Prolonged Use

On several occasions in past decades, the Executive Board recognized that adjustment often required a longer time frame than implied by existing UFR policies1 and, in response, instituted new modalities of UFR. These new policies were initially conceived as short term and temporary, out of concern to preserve the monetary nature of the IMF and the revolving character of its resources, but they ended up being renewed year after year. Until the early 1990s, these guiding principles were thought to apply equally to the use of the IMF’s general and concessional resources.2 Thereafter, the policies applied to the two groups diverged, and there was a gradual acceptance of a greater degree of prolonged use of concessional resources while giving renewed emphasis to the revolving character of the IMF’s general resources. 3

Prolonged use of the IMF’s general resources

The official interpretation of the IMF’s mandate initially emphasized the temporary nature of the support that the IMF could provide to its members: “The authority to use the resources of the Fund is limited to use in accordance with its purposes to give temporary assistance in financing balance of payments deficits” and “the task of the Fund is to help members that need temporary help. The Fund’s attitude toward the position of each member should turn on whether the problem to be met is of temporary nature and whether the policies the member will pursue will be adequate to overcome the problem within such a period.”4

The creation of the EFF in the wake of the first oil shock marked the first important departure from the original conception. However, the wording of the decision made it clear that this departure was intended to be the exception, not the rule: the EFF was to be used in special circumstances, including where a member suffered serious payments imbalances relating to structural maladjustments and where it was expected that the needed improvement in the balance of payments could only be achieved over an extended period. Subsequent developments, which included an increasing use of series of one-year SBAs and the institution and prorogation until 1992 of the enlarged access policy, ensured that the use of the EFF indeed remained exceptional, although not the recourse to IMF resources for a more prolonged period of time than implied by the original interpretation of the IMF’s mandate.

The 2000 Executive Board discussion of the “Review of Fund Facilities” marked a sharp reversal of attitudes toward prolonged UFR. On this occasion, a number of Board members expressed concern “that some members may rely unduly on Fund financial assistance in place of seeking market financing, and saw a need to review the Fund’s policies in this connection.”5 These concerns led to the introduction of repurchase expectations6 and of surcharges on outstanding obligations to the IMF in excess of normal access (i.e., 100 percent and 300 percent of a member’s quota). While primarily aimed at providing an incentive against large use of IMF resources, this measure was also presented as an indirect incentive to avoid prolonged use, to the extent that it is associated with rising outstanding obligations.

Prolonged use of the IMF’s concessional resources

Beyond the concessionality of the loans attached to it, the main innovation brought about by the ESAF was the relaxation of the requirement that lending arrangements should solve entirely members’ balance of payments problems. Instead, programs supported by the ESAF were required only to “assure substantial progress during the three-year period toward an overall position and structure of the balance of payments that is consistent with orderly relations with creditors and a reduction in restrictions on trade and payments, while permitting the timely servicing of obligations to the Fund” (EBM/87/171). The ESAF being initially conceived as a one-off operation, the decision was ambiguous, to say the least, as to how the unfinished agenda should be tackled in the post-ESAF period.

Between 1990 and 1997, the ESAF was gradually transformed through a series of steps into a permanent facility without any restrictions on the number of arrangements that an eligible member could enter into. In late 1990, the ESAF Trust Instrument was amended so as to allow one additional annual arrangement at the expiration of the initial three-year ESAF arrangement, although only where performance had been satisfactory and within unchanged overall access limits. In 1992, the Board opened the possibility of renewing ESAF support through a single one- or two-year arrangement, when the three-year commitment period had expired with undrawn amounts. Then, in 1993 the Instrument was amended again to allow for a second three-year arrangement, which could itself be followed by a single annual arrangement. This option was to be available only for good performers with appropriately strong adjustment programs. In 1995, the ESAF became a self-sustaining facility, offering eligible members indefinite access to concessional resources, though each member would remain bound by the limits set in 1993 regarding the number of arrangements and the “good performance” test. In 1997, these last limits were lifted.

These successive extensions were agreed upon only after protracted negotiations, due to the reluctance of a minority of Directors to legitimize prolonged use of the IMF resources, even concessional ones. The need to reflect these different perspectives led the Board as a whole to emphasize that the purpose of these successive extensions was not to provide a source of continuous financing for individual countries, but rather to maintain the Fund’s ability to respond to members’ needs as they arise.7 Apart from the factors mentioned in Chapter 3, this decision also reflected a third, “defensive lending” motivation: ensuring a smooth repayment by the countries with the heaviest debt-service ratios to the IMF.8

Evolution of the Strategy Vis-à-Vis Prolonged Use

Program design elements

From 1984 to 1991, reviews of prolonged use put a strong emphasis on improvements in program design and implementation to address prolonged use, each of the reviews essentially building on the previous ones and increasing the specificity of its recommendations. By contrast, the 2000 review, which tended to downplay the importance of prolonged use, did not suggest any specific remedy related to program design, nor did it recall or call for the implementation of the measures endorsed in previous reviews.

Access to IMF resources

The majority view of the Executive Board regarding access has consistently been that it would not be appropriate to introduce strict rules limiting access based on the frequency or length of UFR, because even perfect implementation might fail to deliver the desired balance of payments outcome. However, the policies adopted in 1983/84 on the use of general resources made it clear that access should be reduced over time and that past performance in using the IMF’s resources should be taken into account in the determination of further access.

While these policies were not applicable ipso facto to concessional resources, the decisions adopted by the Board from 1990 onward left little doubt that the guiding principles of access policy were similar for both categories of resources. In 1993, the Board decided that “for repeat users, access would take into account the amount of the member’s outstanding use of Fund credit and its record in using Fund resources. . . . This would signal the need to phase out the reliance on exceptional balance of payments financing” and “ensure that even with continued availability of the ESAF, individual members would, over time, phase out their reliance on ESAF support.”9 In 1995, the Board further specified that “lower (or no) access may be appropriate in the case of… countries that have relatively weak track records and are not able to implement sufficiently strong policies…”

Our case studies suggest that the justification of the level of access proposed in staff reports was treated in a rather perfunctory manner. This eventually caused the Executive Board in July 2000 to ask for a revision of the operational guidelines calling on staff to provide more detailed justifications of access proposals. As regards the evolution of the level of access, only about one-fifth of prolonged users with more than one three-year ESAF/PRGF arrangement had a consistently diminishing access. A similar proportion had access that actually increased over time. The remainder had access that either was stable over time or diminished only between the first and second three-year arrangement, and remained broadly stable thereafter. Among GRA arrangements, since 1990, 43 percent of prolonged users had higher annual access in their most recent (or last) arrangement than in their first, and just over a fifth had a consistently diminishing annual access. Another way to capture the lack of consistent implementation of access guidelines is to look at the evolution of prolonged users’ outstanding obligations to the IMF over time (see Annex Figure 2.1). The general trend is fairly consistent both within and across groups: outstanding UFR declined sharply in the second half of the 1980s, but then remained fairly steady during the 1990s.10

Annex Figure 2.1.Average Outstanding Obligations of Prolonged Users to the IMF

(In percent of quota)

Sources: IMF Treasurer’s Department and IEO calculations.

Note: In this figure, prolonged users are treated as a fixed group, consisting of the countries listed in Chapter 2. However, the broad trends are not very sensitive to the precise composition. Not every country in this sample was a prolonged user in each year. The choice of the fixed rather than the dynamic definition in this case was dictated by concerns not to understate the decline in outstanding obligations of the group of prolonged users.VPU: very prolonged users.

Strengthened analytical and assessment efforts

In 1990, the Executive Board approved the proposal to include in any new UFR request a systematic review of experience under preceding arrangements. In 1995, the Board went a step further by recommending stock-taking, on a case-by-case basis, toward the end of the three-year arrangement, to reflect on what has been achieved and how to ensure strong performance in a subsequent arrangement (i.e., without necessarily waiting for a new UFR request to arise).11

Exit strategies

Evidence from the case studies again suggests that the recommendation that staff reports should provide medium-term balance of payments projections and attempt to foresee a reasonable timetable for the disengagement of the IMF was often not followed. For example, medium-term projections for the Philippines in the 1994 EFF projected financing gaps even after market access had been restored. Part of the problem was the lack of clear criteria for a balance of payments financing gap in cases where countries had access to private financial markets. By contrast, in Pakistan and Senegal, most medium-term projections showed no financing gap beyond the program period, but such projections proved unrealistic.

As concerns the use of strengthened surveillance in the post-program period, the emphasis put by the 2000 “Review of Fund Facilities” on post-program monitoring essentially just formalized a preexisting disposition. While all GRA arrangements have a consultation clause stating that, under certain conditions, members shall consult with the IMF after the expiry of the arrangement “at the request of the Managing Director,” the facilities review instituted a presumption that countries with obligations to the IMF exceeding 100 percent of their quota at the expiration of the program would undergo this procedure for as long as their outstanding liabilities to the IMF exceeded the threshold.

For users of concessional resources, the principle of post-program monitoring as a means of avoiding prolonged use of ESAF resources was formally established in the early 1990s. In considering operational details for an ESAF successor, the Board endorsed the suggestion of “post-ESAF enhanced consultations and program monitoring… on a limited transitional basis, in cases where the macroeconomic situation remains vulnerable and the authorities perceive benefits in a continued close policy dialogue with the IMF.12 Subsequently, it was also envisaged that one option for continued IMF support for the programs of former ESAF users that ceased to have a need for IMF financing would be through precautionary arrangements: “Directors considered that… a precautionary arrangement would signal the Fund’s approval of the country’s adjustment program, thereby catalyzing financial support from other sources, while providing assurances that Fund resources would be available should the country’s circumstances change. Directors were persuaded, however, by the arguments against granting precautionary ESAF arrangements. They broadly agreed that ESAF-eligible countries without a recurrent or prospective balance of payments need could instead request a precautionary extended arrangement, which could be replaced or supplemented by an ESAF arrangement in the event that a balance of payments need emerged.”13

The implications of this exit strategy for other creditors were spelled out rather bluntly in 1991, when a staff report noted that: “In cases where external viability is not in reasonable prospect… the Fund could provide support in the early stages of the adjustment process… to help ensure the establishment of an appropriate macroeconomic framework. However, other creditors may have to continue their contributions, in part to facilitate repayments to the Fund, and there would need to be a clear acknowledgement by creditors of the revolving character of the Fund’s resources.”14

Notes

Initially, IMF financing in the upper credit tranches was typically provided under a Stand-By Arrangement, whose normal period is one year. It may extend up to but not beyond three years in appropriate cases (Executive Board Decision No. 6056-(79/38) of March 2, 1979). Since 1978, obligations incurred under a Stand-By Arrangement must be repaid within a period of 3¼ to 5 years.

For example, EBS/91/108, “Selected Operational Issues Related with the Use of Fund Resources” explicitly notes that the principle of the revolving character of the IMF’s resources must be applied consistently to general and concessional resources, and does not distinguish between the two in the remedial actions it suggests to deal with prolonged use.

The Executive Board reviewed prolonged use on several occasions during the 1980s, starting in 1984. The last comprehensive review of prolonged UFR was discussed by the Board in 1991. Thereafter, the issue was not put on the Board’s agenda until 2000, where prolonged use was discussed only as a background issue to the review of IMF facilities.

Executive Board Decisions No. 71-2 of September 26, 1946 and No. 102-(52/11) of February 13, 1952, respectively. The “Guidelines on Conditionality” adopted in 1979 further stated that: “The normal period for a stand-by arrangement will be one year. If, however, a longer period is requested by a member and considered necessary by the Fund to enable the member to implement its adjustment program successfully, the stand-by arrangement may extend beyond the period of one year. This period in appropriate cases may extend up to but not beyond three years.” (Executive Board Decision No. 6056-(79/38) of March 2, 1979.)

See Chairman’s summing up (BUFF/00/41).

For purchases in the credit tranches and under the CFF, the expectation schedule starts one year in advance of the obligation schedule, beginning 2¼14; years after a purchase and ending after 4 years. For the EFF, the expectation schedule begins after 4½ years, as with the obligation schedule, but repurchases are to be doubled, such that the expectation schedule will end after 7 years rather than 10 years under the obligation schedule.

See Chairman’s summing up of EBM/97/5, EBM/97/8, and EBM/97/10.

The then Managing Director put the case in the following terms: “I would suggest that these few cases could appropriately be addressed through the continued availability of concessional ESAF resources on present terms” [as opposed to extending to them one further round of ESAF arrangements with a 20-year maturity, as proposed by the U.K. Chancellor of the Exchequer]. “Through this instrument, the Fund would have the possibility of tailoring its financing to the individual situation of each member, extending for the period needed—in a few cases through several successive ESAF arrangements—the concessional financing required. . . , while avoiding significant humps in net transfers from the member to the Fund.” (BUFF/95/31.)

EBS/93/32, “Operational Modalities and Funding Alternatives for an ESAF” and EBS/95/130, “Continued Financing and Adaptation of the ESAF.”

The step declines observed in 1981, 1993, and 1999 partly reflect the impact of general quota increases. It should be noted that if access is reduced very gradually, disbursements may exceed repayments for a relatively long period, especially under concessional facilities, thus causing outstanding obligations to increase for a while even though access itself is being reduced.

See BUFF/90/37 and BUFF/95/95 for the acting Chairman’s summing up of the relevant Board discussions.

“Operational Modalities and Funding Alternatives for an ESAF Successor—Preliminary Considerations” (EBS/93/32).

Chairman’s summing up of EBM/98/73 on “Distilling the Lessons from the ESAF Reviews.”

“Selected Operational Issues Related with the Use of Fund Resources” (EBS/91/108).

Annex 3: Characteristics of Prolonged Users: Further Details on the Evidence

This annex provides more details of the analyses discussed in Chapter 4.

Econometric Evidence on the Characteristics of Prolonged Users

We estimated a series of probit regressions to examine whether prolonged users had economic and institutional characteristics that were different from “temporary” users. The characteristics considered, which were drawn from the recent empirical literature on participation in IMF arrangements,1 were (i) per capita GDP; (ii) real GDP growth; (iii) current account balance (in relation to GDP); (iv) international reserves (in months of imports); (v) debt-service ratio (measured in relation to exports); (vi) openness of the economy (measured as the ratio of the sum of exports and imports to GDP); (vii) primary exports (as a share of total exports); and (viii) volatility in the terms of trade (standard deviation of the terms of trade index).

Two definitions of “prolonged use” were employed in these exercises—one “fixed” over time, and the other “dynamic” (i.e., time-specific).2 Using the fixed definition, which classified a country as a prolonged user if it had IMF arrangements in 7 out of any 10-year period during 1971–2000, and entire sample period average data, we found prolonged use to be associated with lower levels of international reserves, with higher debt-service ratios, and with lower real GDP growth. There was no statistically significant difference between prolonged and “temporary” users with respect to the other characteristics considered (column 1 in Annex Table 3.1).3 When the sample was limited to PRGF-eligible countries only, prolonged use was found to be associated with higher debt-service ratios and lower GDP per capita (column 2 in Annex Table 3.1).4

Annex Table 3.1.Characteristics of Prolonged Users of IMF Resources1
Fixed definition sample averagesDynamic definition five-year averages
AllPRGFnon-PRGFAllAllPRGFnon-PRGFAll
Marginal probabilities(1)(2)(3)(4)(5)(6)(7)(8)
GDP per capita−0.048−0.545−0.0390.0140.0121.184−0.057−0.078
(1.09)(1.73)*(0.66)(0.26)(0.08)(1.47)(0.33)(0.31)
Real GDP growth−0.058−0.020−0.077−0.031−0.020−0.039−0.021−0.025
(1.72)*(0.42)(1.45)(0.84)(1.29)(1.33)(0.95)(0.93)
Current account balance0.0060.024−0.0280.013−0.008−0.021−0.020−0.045
(0.45)(1.29)(0.71)(0.77)(0.93)(1.50)(1.35)(2.19)**
Foreign reserves−0.070−0.017−0.025−0.0630.0560.136−0.0040.072
(1.80)*(1.43)(0.42)(1.50)(2.07)**(2.75)***(0.10)(1.44)
Debt-service ratio0.0150.024−0.0000.011−0.004−0.006−0.008−0.003
(2.16)**(2.10)**(0.03)(1.28)(0.96)(0.72)(1.56)(0.35)
Openness−0.0010.005−0.0060.0020.001−0.0030.0070.008
(0.34)(1.20)(1.28)(0.76)(0.40)(0.57)(1.42)(1.53)
Primary exports−0.0010.001−0.007−0.0010.003−0.0030.0080.008
(0.42)(0.20)(1.23)(0.34)(0.63)(0.43)(1.36)(1.09)
Term of trade volatility−0.004−0.005−0.006−0.003−0.009−0.008−0.0150.005
(0.72)(0.70)(0.30)(0.66)(1.44)(0.84)(1.54)(0.40)
Lagged GDP per capita−0.000−0.0020.0000.000
(0.25)(1.83)*(0.50)(0.48)
Lagged real GDP growth−0.015−0.0400.0000.025
(1.10)(1.75)*(0.02)(1.05)
Lagged current account balance−0.0200.002−0.064−0.076
(2.07)**(0.18)(3.58)***(3.36)***
Lagged foreign reserves−0.070−0.153−0.031−0.146
(2.57)**(2.79)***(0.98)(2.54)**
Lagged debt-service ratio0.0120.0180.0100.015
(2.91)***(2.18)**(2.20)**(2.25)**
Lagged openness−0.0030.004−0.010−0.008
(0.81)(0.72)(1.96)*(1.41)
Lagged primary exports−0.0020.006−0.009−0.008
(0.43)(0.87)(1.40)(0.98)
Lagged terms of trade volatility−0.000−0.0010.001−0.004
(0.08)(0.26)(0.22)(0.54)
Bureaucracy quality−0.231−0.077
(2.06)**(1.03)
Observations83483565218105113124
Pseudo R-squared0.140.260.180.130.180.320.290.32
p-value0.03780.03020.36420.26030.00000.00010.00110.0000
Sources: IMF, WEO and MONA databases; ICGR database; and IEO calculations

For countries not eligible for the PRGF (i.e., middle- and high-income users of IMF resources), we found no statistically significant differences between prolonged users and “temporary” users for any of the variables (column 3 in Annex Table 3.1).

Introduction of an institutional variable—quality of government bureaucracy5—suggested that prolonged use was associated with lower quality of government bureaucracy, and that once this factor was taken into account, the differences in economic characteristics (i.e., growth, international reserves, and debt-service ratio) were no longer significantly different between prolonged and “temporary” users (column 4 in Annex Table 3.1).

In order to allow for some dynamics, a second set of exercises used a period-specific definition of “prolonged use,” based on five-year average panel data (columns 5–8 in Annex Table 3.1). A country was defined to be a prolonged user in a particular five-year period if it had IMF arrangements in seven or more years during that and the preceding five-year period. Prolonged use was found to be strongly associated with (i) lower international reserves in the preceding five-year period but higher reserves in the current five-year period; (ii) lower current account balances in the preceding five-year period; and (iii) higher debt service in the preceding five-year period. No statistically significant difference was found in the quality of bureaucracy.

When the sample was limited to only PRGF-eligible countries, prolonged use was again associated with lower levels of reserves in the previous period but higher levels in the current period; and with higher debt-service ratios in the preceding period. Prolonged use was also associated with lower GDP growth for this group of countries. For countries not eligible for the PRGF, taking account of both contemporaneous and lagged effects, prolonged use was found to be associated with larger current account deficits, larger debt-service ratios, and less open economies.

Cross-Section Evidence on Comparison Between Prolonged and “Temporary” Users

Starting conditions

To compare the “starting conditions” of prolonged users at the beginning of their episode of prolonged use with those of contemporaneous “temporary” users, we identified two subperiods (1976–79 and 1988–91) during which a large proportion of the episodes of prolonged use that we studied were initiated and looked at economic conditions in the three years preceding the first program of the prolonged use series for the two groups of prolonged users thus identified. We then identified two control groups of “temporary” users, consisting of all the countries that entered into an IMF arrangement during the same periods. Starting conditions were appraised by looking at five measures of potential macro imbalances: public debt, external debt, current account balance, overall fiscal balance, and inflation. The results of the comparison are shown in Annex Table 3.2.

Annex Table 3.2.Comparison of Starting Conditions for Groups of Prolonged and “Temporary” Users(In percent of GDP, unless otherwise indicated)
Public debtExternal debtCurrent account balanceOverall budget balanceInflation (in percent)
1976–79
Prolonged users31.937.7−6.6−6.723.7
“Temporary” users58.915.2−3.8−6.019.0
Statistical significancens**nsnsns
1988–91
Prolonged users116.1157.1−4.8−10.19.8
“Temporary” users45.357.1−2.8−5.324.5
Statistical significancens*nsnsns
Sources: IMF, WEO database; and IEO calculations.Note: ** and * indicate statistical significance at the 1 percent and 10 percent levels, respectively; ns indicates no significance.

Economic performance and macroeconomic adjustment6

As regards GDP growth, a comparison between prolonged and “temporary” users of IMF resources over the last three decades suggests that, in most periods, prolonged users grew at a slower pace than “temporary” users, the exceptions being the early 1970s and early 1990s for middle-income countries (i.e., the times when there were few debt crises) and the 1990s for low-income countries (Annex Figure 3.1).

Annex Figure 3.1.GDP Growth

(Five-year average annual change; in percent)

Sources: IMF, WEO database; and IEO calculations.

Export growth was generally much weaker, on average, in the group of prolonged users than in the “temporary users” group as far as low-income countries are concerned. For middle-income countries, the opposite was generally true, but differences were less pronounced (Annex Figure 3.2).7

Annex Figure 3.2.Export Growth

(Average annual change; in percent)

Sources: IMF, WEO database; and IEO calculations.

The analysis of adjustment performance, as measured by trends in inflation and reductions in current account deficits, does not show any clear or consistent differences between the two groups, although there are large variations within each group.8 As regards fiscal deficits, in both middle- and low-income countries, prolonged users had higher deficits in the late 1970s, but they adjusted faster thereafter and thus had lower deficits than “temporary” users in subsequent periods (see Annex Figure 3.3).

Annex Figure 3.3.Evolution of Overall Fiscal Deficit

(Five-year annual average; in percent of GDP)

Sources: IMF, GFS database; and IEO calculations.

Key fiscal characteristics

Prolonged users have lower and more rigid government expenditure

Among middle-income countries, the expenditure to GDP ratio of prolonged users was consistently and markedly lower than for “temporary” users over 1971–2000. In other words, the prolonged users are not necessarily those with a tendency toward “big” government—indeed the reverse; as will be seen below, the most obvious distinguishing characteristic appears to be a weak tax base. The differences were less marked for the PRGF-eligible countries.

In both low-income and middle-income countries, the government expenditure to GDP ratio expanded significantly less over the last three decades in prolonged user countries than in “temporary” user countries, which might reflect either the fiscal discipline imposed by the successive IMF-supported programs entered into by prolonged users, or simply their generally poor ability to increase revenue collection, or some combination of the two (see below). The likely impact of IMF-supported programs is suggested by the pattern of government expenditure in low-income countries, which exhibits a clear downside break in the mid-1980s, when most of these countries started making extensive use of IMF resources, under newly created concessional facilities (Annex Figure 3.4).

Annex Figure 3.4.Government Expenditure

(Five-year annual average; in percent of GDP)

Sources: IMF, GFS database; and IEO calculations.

The analysis of the composition of government expenditure further reveals that, regardless of the income group, prolonged users had higher interest and defense expenditure (as a proportion of total ex-penditure) in all periods since 1970, and especially in the 1980s, largely reflecting a buildup in debt problems (see below). Other things being equal, these differences would result in a more rigid structure of expenditure in prolonged user countries, which might account for a more protracted adjustment process (Annex Figure 3.5).

Annex Figure 3.5.Interest and Defense Expenditure

(Five-year annual average; in percent of total government expenditure)

Sources: IMF, WEO database; and IEO calculations.

Middle-income prolonged users collect less tax revenue

Differences related to the tax revenue to GDP ratio are particularly pronounced among middle-income countries: over 1971–2000, prolonged users in that category have consistently had lower tax to GDP ratios than “temporary” users. Both prolonged and “temporary” users have registered increases in that ratio over time, but that increase was faster for “temporary” users up to the 1990s. By contrast, among low-income countries, there was no sustained increase in the tax revenue to GDP ratio over time, and the gap between prolonged and “temporary” users, which prevailed until the mid-1980s, was eliminated in later periods only owing to a decline in “temporary” users’ tax revenues (Annex Figure 3.6).

Annex Figure 3.6.Tax Revenues to GDP Ratio

(Five-year annual average)

Sources: IMF, GFS database; and IEO calculations.

Prolonged users faced a higher public debt burden for most of the period

Among PRGF-eligible countries, prolonged users’ stock of public debt (relative to GDP) was three times as large as that of “temporary” users at the beginning of the period. However, differences rapidly diminished from the mid-1980s onward, as the “temporary” users borrowed at a much faster pace than the prolonged users, perhaps reflecting the fact that many prolonged users had already encountered debt problems.

Among middle-income countries, prolonged users initially had a substantially lower public debt stock (relative to GDP) but debt levels for the group built up rapidly during the 1980s (Annex Figure 3.7).

Annex Figure 3.7.Public Debt Stock

(Five-year annual average; in percent of GDP)

Sources: IMF, GFS database; and IEO calculations.

External sector

Trade

For both low- and middle-income countries, but particularly for the latter, terms of trade shocks9 were, on average, of greater magnitude in prolonged user countries.

As regards trade openness, there is a marked difference among middle-income countries: prolonged users were continuously less open than “temporary” users, in the sense that their trade to GDP ratio was consistently lower—by 10 to 15 percentage points—over 1971–2000, even though for both groups that ratio increased over the period (Annex Figure 3.8). By contrast, there is no significant difference between “temporary” and prolonged users as far as PRGF-eligible countries are concerned.

Annex Figure 3.8.Trade Openness and Concentration

Sources: IMF, WEO database; and IEO calculations.

With respect to the composition of exports, prolonged users in both PRGF-eligible and non-PRGF-eligible groups had a higher share of primary exports than “temporary” users, and that gap tended to increase over time. The concentration of exports on primary commodities also declined faster in “temporary” users, which may be related to their greater openness to trade (see Annex Figure 3.8).

In keeping with the findings of previous studies on the determinants of repeat UFR,10 both groups of prolonged users on average had markedly lower gross reserves (in relation to their external debt) than “temporary” users. However, data on imports coverage by gross international reserves unexpectedly indicate that prolonged users have had a slightly higher coverage of imports than “temporary” users throughout the 1971–2000 period, and the difference, although small, is statistically significant (Annex Table 3.3). Once again, this may reflect the generally lower trade openness of the prolonged users.

Annex Table 3.3.Comparison of Prolonged and “Temporary” Users1(In percent, unless otherwise specified)
Economic Performance
1971–751976–801981–851986–901991–951996–20001971–801981–901991–2000
GDP growth(Period geometric mean)
PRGF-eligible users
Prolonged3.03.20.72.31.65.53.11.53.5
Temporary3.32.73.22.80.53.93.23.01.7
t test significance1nsns**nsnsnsns*ns
Non-PRGF-eligible users
Prolonged6.34.72.02.42.32.25.02.22.2
Temporary5.55.12.22.71.03.05.22.71.6
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan3.26.26.85.84.82.84.76.33.9
Philippines5.86.1−1.34.72.23.35.91.72.7
Senegal2.41.03.03.21.55.31.73.13.2
Per capita GDP growth
PRGF-eligible users
Prolonged0.50.4−1.40.0−0.92.70.4−0.70.7
Temporary2.31.9−0.30.6−2.71.52.1−0.2−0.6
t test significance***nsnsnsns**nsns
Non-PRGF-eligible users
Prolonged3.12.2−0.51.30.91.22.60.41.1
Temporary4.12.91.02.3−0.32.13.41.71.1
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan0.04.23.53.42.20.61.53.51.4
Philippines2.91.9−3.50.8−0.11.53.1−0.70.7
Senegal−0.6−0.3−0.8−0.7−1.02.5−1.20.30.7
Inflation
PRGF-eligible users
Prolonged12.019.255.2167.7158.077.115.364.515.7
Temporary10.711.113.213.618.633.410.714.123.3
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged19.518.927.855.177.648.321.132.122.4
Temporary19.517.223.042.858.447.318.326.418.6
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan15.78.85.29.411.27.312.27.09.2
Philippines17.012.914.312.710.07.114.713.08.5
Senegal13.58.97.5−0.66.81.410.15.84.1
Growth of exports
PRGF-eligible users
Prolonged3.25.7−1.32.94.67.44.41.46.2
Temporary4.94.72.44.64.19.14.33.24.4
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged8.37.34.25.57.26.87.74.87.0
Temporary5.16.73.27.13.64.85.15.24.3
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan−4.011.212.210.89.00.31.68.14.5
Philippines4.48.62.47.29.43.39.73.66.3
Senegal1.2−2.70.51.20.85.0−1.44.52.9
Fiscal Characteristics
Overall budget deficit (percent of GDP)(Period average)
PRGF-eligible users
Prolonged−4.0−6.0−7.9−4.5−4.1−3.1−5.2−6.8−3.4
Temporary−3.5−5.6−8.1−6.1−5.1−3.8−4.4−4.3−7.2
t test significancensnsnsnsnsnsnsns**
Non-PRGF-eligible users
Prolonged−3.1−4.8−4.5−2.6−1.2−2.2−4.1−3.8−1.5
Temporary−4.1−4.2−5.0−3.2−2.7−1.9−3.9−4.0−2.3
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan−7.6−8.0−6.1−7.3−7.6−6.5−7.9−6.7−7.1
Philippines−1.0−1.3−2.9−3.2−0.6−1.3−1.2−3.1−0.9
Senegal−1.1−0.7−5.9n.a.n.a.n.a.−0.9−5.9n.a.
Tax revenues (percent of GDP)
PRGF-eligible users
Prolonged13.814.916.014.314.314.116.115.215.3
Temporary20.220.520.513.014.613.920.317.613.8
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged13.115.616.315.018.718.714.316.318.5
Temporary18.421.624.124.025.125.020.224.224.9
t test significance******************
Memorandum
Pakistan10.312.310.612.415.616.011.411.515.8
Philippines15.018.718.1n.a.n.a.n.a.16.918.1n.a.
Senegal11.011.812.913.412.713.111.513.112.9
Government expenditure (percent of GDP)
PRGF-eligible users
Prolonged21.123.828.523.523.222.121.526.023.1
Temporary17.423.027.525.725.223.221.326.023.9
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged22.926.125.022.523.524.024.124.523.6
Temporary25.828.632.631.231.430.527.331.630.9
t test significancensns********ns****
Memorandum
Pakistan16.917.419.023.323.622.217.221.223.0
Philippines13.913.812.016.618.719.113.814.318.9
Senegal18.220.327.1n.a.2n.a.n.a.19.227.1n.a.
Of which interest (percent of expenditure)
PRGF-eligible users
Prolonged5.16.210.810.014.916.85.210.314.7
Temporary3.24.15.98.611.814.13.76.611.9
t test significance******nsnsnsns**ns
Non-PRGF-eligible users
Prolonged4.76.312.715.712.211.95.314.912.0
Temporary4.75.79.613.011.310.35.111.510.9
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan9.810.714.419.923.929.110.416.926.2
Philippines3.75.613.832.427.218.44.823.123.3
Senegal2.36.27.5n.a.n.a.n.a.4.37.5n.a.
Of which defense (percent of expenditure)(Period average)
PRGF-eligible users
Prolongedn.a.n.a.n.a.14.410.110.5n.a.14.59.9
Temporaryn.a.n.a.n.a.20.616.417.1n.a.20.516.5
t test significancen.a.n.a.n.a.ns****n.a.ns**
Non-PRGF-eligible users
Prolongedn.a.n.a.n.a.18.111.29.7n.a.17.211.2
Temporaryn.a.n.a.n.a.10.59.39.0n.a.10.59.1
t test significancen.an.an.a.**nsnsn.a.**ns
Memorandum
Pakistann.a.n.a.28.127.626.624.0n.a.27.725.9
Philippinesn.a.n.a.9.511.210.18.4n.a.10.99.6
Senegaln.a.n.a.8.86.610.38.7n.a.7.19.8
Public debt stock (percent of GDP)
PRGF-eligible users
Prolonged30.646.790.290.892.094.938.686.281.9
Temporary44.421.047.156.081.395.042.453.984.8
t test significancens******nsnsns**ns
Non-PRGF-eligible users
Prolonged43.437.945.755.244.140.540.554.542.1
Temporary29.236.151.748.445.443.332.647.643.3
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan66.956.854.473.876.579.161.864.177.2
Philippines43.530.229.651.958.160.336.840.858.9
Senegal13.9n.a.60.0n.a.n.a.n.a.8.060.0n.a.
Public debt service
PRGF-eligible users
Prolonged10.014.222.626.222.117.013.624.219.8
Temporary8.17.715.619.312.412.57.717.112.6
t test significancens**nsns**ns******
Non-PRGF-eligible users
Prolonged5.425.128.426.118.619.824.527.319.1
Temporary14.315.521.923.414.015.515.022.714.8
t test significance******************
Memorandum
Pakistan20.919.919.424.926.227.920.422.226.9
Philippinesn.a.23.435.530.621.612.223.433.117.4
Senegal6.214.615.827.715.518.412.221.816.8
PPG debt service (percent revenue)
PRGF-eligible users
Prolonged14.215.519.723.224.619.414.821.023.5
Temporary6.96.110.614.315.618.46.312.217.3
t test significancens****************
Non-PRGF-eligible users
Prolonged14.822.632.027.520.222.319.129.421.4
Temporary8.410.315.723.815.514.59.419.215.0
t test significance******nsns********
Memorandum
Pakistan19.115.918.018.823.621.317.118.422.6
Philippines13.014.928.043.235.730.914.135.633.6
Senegal10.024.118.1n.a.n.a.n.a.17.018.1n.a.
Stock of external debt (percent of GDP)(Period average)
PRGF-eligible users
Prolonged29.031.135.640.746.051.956.763.070.2
Temporary12.427.345.178.3109.898.719.959.4111.1
t test significance**nsns********nsns
Non-PRGF-eligible users
Prolonged28.741.763.978.761.753.135.271.358.3
Temporary20.028.143.054.147.042.524.149.747.6
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan52.247.040.047.850.351.548.543.950.8
Philippines29.145.372.179.462.561.237.275.861.9
Senegal17.036.761.059.766.275.430.570.861.2
Balance of Payments Characteristics
Current account deficit (percent of GDP)(Period average)
PRGF-eligible users
Prolonged−4.4−3.2−7.4−5.7−7.7−9.0−3.5−6.1−8.3
Temporary−4.2−4.5−7.4−6.0−7.9−9.1−4.7−6.4−8.6
t test significancens**nsnsnsns**nsns
Non-PRGF-eligible users
Prolonged−2.8−3.4−3.9−2.3−2.8−2.7−3.4−2.9−2.8
Temporary−4.5−2.6−5.6−1.8−1.9−3.0−2.9−3.5−2.3
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan−4.7−4.6−2.7−2.6−3.6−4.8−4.6−2.6−4.1
Philippinesn.a.−5.0−5.4−1.7−3.40.7−5.0−3.6−1.6
Senegal−4.6−8.3−13.4−8.2−6.4−4.2−7.2−10.8−5.4
Gross international reserves (months of imports)
PRGF-eligible users
Prolonged4.34.74.54.75.86.64.74.76.2
Temporary5.03.42.62.52.73.43.42.53.1
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged5.58.57.47.88.58.88.57.58.6
Temporary4.34.73.13.23.33.54.53.23.3
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan3.62.93.01.82.01.33.22.41.7
Philippinesn.a.4.31.72.43.13.24.32.03.1
Senegal0.30.30.20.20.72.60.30.21.6
Gross international reserves (billions of U.S. dollars)
PRGF-eligible users
Prolonged0.10.20.20.30.70.10.20.30.8
Temporary0.10.40.80.91.74.80.30.83.0
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged0.61.51.52.55.14.21.02.07.0
Temporary0.12.42.52.95.78.31.72.76.9
t test significance**nsnsnsns**nsnsns
Memorandum
Pakistan0.51.01.81.32.20.20.71.51.9
Philippines1.02.51.52.36.14.91.81.98.3
Senegal0.00.00.00.00.10.00.00.00.2
Gross international reserves (percent external debt)(Period average)
PRGF-eligible users
Prolonged35.914.16.26.09.111.524.86.110.2
Temporary103.364.033.620.418.223.984.725.221.0
t test significancens****************
Non-PRGF-eligible users
Prolonged60.433.314.717.228.825.444.115.926.3
Temporary93.677.830.124.746.839.984.727.443.2
t test significancensns**nsns********
Memorandum
Pakistan9.911.515.17.58.35.610.711.37.0
Philippines32.823.06.37.917.025.827.97.121.4
Senegal11.53.91.00.72.810.67.70.86.7
Terms of trade
PRGF-eligible users
Prolonged141.7146.6119.9113.096.6105.3144.1116.5100.9
Temporary137.0146.9143.1134.9115.4118.8142.0139.0117.1
t test significancensnsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged97.7105.3102.197.097.596.9101.599.597.5
Temporary99.5103.7106.1103.699.5100.2101.6104.899.8
t test significancensnsnsnsnsnsnsns**
Memorandum
Pakistan132.5133.6124.2120.297.1118.7133.0122.2107.9
Philippines124.996.892.8106.7105.392.0110.999.8100.5
Senegal100.6105.3103.6112.7100.7100.3102.9108.1100.5
Trade (percent of GDP)
PRGF-eligible users
Prolonged58.665.260.860.168.473.761.561.170.7
Temporary41.268.362.760.371.975.865.660.674.7
t test significance**nsnsnsnsnsnsnsns
Non-PRGF-eligible users
Prolonged48.259.861.060.866.473.458.961.969.5
Temporary61.972.069.971.180.785.269.570.982.7
t test significance****ns************
Memorandum
Pakistan29.231.233.835.739.037.830.234.738.5
Philippines45.047.148.555.370.2102.446.151.984.5
Senegal69.576.379.756.162.972.872.967.967.3
Share of primary exports (percent of merchandise exports)
PRGF-eligible users
Prolonged89.388.284.576.376.873.788.283.677.7
Temporary86.582.176.568.069.459.584.875.667.1
t test significancensnsnsnsns**nsnsns
Non-PRGF-eligible users
Prolonged73.169.366.758.253.349.971.261.051.2
Temporary72.567.865.456.247.845.969.761.547.0
t test significancensnsnsnsnsnsnsnsns
Memorandum
Pakistan44.144.839.829.517.615.544.434.616.7
Philippines90.281.075.565.952.424.585.670.740.0
Senegal81.988.087.275.965.649.084.680.758.3
Sources: IMF, WEO, IFS, and GFS databases; and IEO calculations.

Prolonged users generally faced a heavier external debt and debt-service burden

As far as the stock of external debt is concerned, prolonged users had a significantly larger debt/GDP ratio than “temporary” users until the late 1980s, after which the relationship reversed itself, even though the external debt of PRGF-eligible prolonged users kept rising in relation to their GDP. However, the debt-service burden, as measured by the external debt service to exports ratio, was significantly higher for prolonged users than for “temporary” users throughout 1975–2000.

Political characteristics

The literature on the effectiveness of structural adjustment programs has emphasized the importance of political economy variables in determining the outcome of these programs.11 It was not possible in the context of this project to collect data on the relevant variables over the entire period under review. However, based on the database used by Ivanova and others (2001),12 there appear to be few consistent differences between prolonged and “temporary” users as far as political characteristics are concerned.

The one important exception is the measure of political instability, which appears to be higher among prolonged than among “temporary” users for both middle- and low-income countries.13 Prolonged users as a whole also appear to suffer from ethnic fractionalization to a greater extent than “temporary” users, but this is true only for middle-income countries.

Finally, while several authors have found a relationship between IMF-supported programs (related either to their presence or to their design) and the closeness of the relationship between the member country and the IMF’s major shareholders, the comparison between prolonged users and “temporary” users in terms of their closeness to G-7 countries found no major differences.14

Notes

Due to data limitations for several users of IMF resources during the period covered by the evaluation (1971–2000), a maximum of only 83 countries were covered in the regressions. Also, because data for 1971–75 and for 2000 were missing for many variables for many countries, the annual time series data used spanned 1976–99. Among users of IMF resources that were excluded were countries that either did not exist in 1976 or had missing data for several variables during most of 1976–99.

The list of distinguishing characteristics here is much shorter than that reported in Bird, Hussain, and Joyce (2000), in which the authors found that repeated participation in programs (“recidivism”) was associated with: (i) lower levels of international reserves; (ii) larger current account deficits; (iii) lower and less volatile terms of trade; (iv) larger debt-service ratios; (v) larger capital outflows; (vi) lower per capita income; (vii) lower investment rates; and (viii) weaker governance. Differences in methodology may account for the different results. Bird, Hussain, and Joyce do not predefine a threshold for “recidivism”; rather they regress the number of arrangements and the number of program years on a range of variables using Poisson and negative binomial models.

These estimates do not take account of the likely strong endogeneity between growth and the likelihood that a country will request an IMF arrangement—for example, because exogenous shocks that worsen the balance of payments also harm growth. In Annex 4, when this endogeneity is taken into account, the negative association between growth on prolonged use disappears for PRGF-eligible countries.

The institutional variable used is the “Bureaucracy quality” index calculated by the International Country Risk Guide. It is designed to provide an indication of the policy environment, especially the extent to which policy formulation and day-to-day administrative functions are able to withstand political changes.

See Annex Table 3.3 for detailed figures and statistical significance of the comparisons. In the results presented here, the groups “prolonged users” and “temporary users” are both fixed populations (the former group consisting of the countries listed in Chapter 2), that is, we are looking at the characteristics of a broad group of countries that, at some point in the overall period encountered episodes of prolonged use against other countries which, at some point in the same period, entered into an IMF-supported program, but which did not become prolonged users. However, because the population of prolonged users does not change much over time, the results would not be substantially altered if a “dynamic” definition of prolonged use were used.

These results are statistically significant only for the PRGF-eligible group over the 1980s.

The definition of terms of trade shocks used here is the same as in Ivanova and others (2001) and Dollar and Svensson (2000), namely the difference between the change in the price of exports weighted by the share of exports in GDP and the change in the price of imports weighed by the share of imports in GDP (see Annex Table 3.3).

This database covers the countries that entered into the approximately 170 arrangements with the IMF between 1992 and 1998.

In contrast, measures of political cohesion and of quality of the bureaucracy suggest that prolonged users have a higher degree of political cohesion and a better bureaucracy than “temporary” users. Interestingly, the power of vested interests appears to be identical, on average, in all four country groupings.

See for instance Bird and Rowlands (2001b), Thacker (1999), Barro and Lee (2002), and Ivanova and others (2001). The variable used here to test for the influence of proximity to G-7 countries was the share of G-7 bilateral aid.

Annex 4: Effects of Prolonged Use on Growth: Details of the Econometric Results

This annex provides details of the econometric results discussed in Chapter 5.

Based on empirical analysis of a panel data set spanning five five-year periods (1975–99) for 130 countries, Barro and Lee (2002) found that when they did not control for endogeneity, their results suggested that participation in IMF arrangements was associated with contemporaneously lower per capita growth. However, after controlling for endogeneity of participation in IMF arrangements and for other determinants of growth, IMF arrangements had no statistically significant contemporaneous impact on per capita GDP growth, but rather a lagged negative effect. The authors employed an instrumental variables approach to control for endogeneity of participation in IMF arrangements.1 Specifically, they used the following as instruments for participation: (i) size of quota; (ii) political and economic proximity to IMF major shareholders (the United States, France, and the United Kingdom);2 and (iii) national staff (economists) at the IMF.

For the purposes of this evaluation, one of the coauthors of Barro and Lee (2002), Professor Jong-Wha Lee, extended the analysis in that study to consider whether “prolonged use” has an effect on growth that is distinguishable from that associated with “temporary use.” The rest of this section reports on the findings of several exercises undertaken by Professor Lee, using panel data for 82 users of IMF resources (GRA and concessional) over five five-year periods (1975–79, 1980–84, 1985–89, 1990–94, and 1995–99). The determinants of long-run per capita income growth used encompassed: (i) initial income; (ii) human resources (educational attainment, life expectancy, and fertility); (iii) investment rate; (iv) exogenous shocks (changes in the terms of trade); and (v) policy and institutional variables (government consumption, rule of law, openness, and inflation). Participation in IMF arrangements was measured by loan size.3

A first set of exercises estimated the effects of participation in IMF arrangements, without controlling for the endogeneity of such participation. The results suggested that after controlling for other determinants of growth, IMF arrangements were associated with lower growth contemporaneously and with a lag (equation 1, Annex Table 4.1). Incorporation of contemporaneous and lagged interactive terms to distinguish between “temporary” and prolonged participants in IMF arrangements yielded statistically significant coefficients on the interactive terms, suggesting significantly more adverse effects on growth for prolonged users than for “temporary” users (equation 2, Annex Table 4.1).4

Annex Table 4.1.Effects of “Prolonged Use” of IMF Programs on Economic Growth
Actual values of IMF loan sizeIMF quotas and staff,

political and economic proximity

to the United States and Europe
Instruments for IMF loan(1)(2)(3)(4)
Log (per capita GDP)–0.0271–0.0260–0.0269–0.0279
(5.988)***(6.037)***(6.042)***(6.469)***
Male upper-level schooling0.00360.00300.00350.0034
(1.875)*(1.653)*(1.877)*(1.896)*
Log (life expectancy)0.0360.0400.0420.054
(1.841)*(2.148)**(2.171)**(2.807)***
Log (total fertility rate)–0.0281–0.0300–0.0273–0.0303
(4.372)***(4.891)***(4.300)***(4.918)***
Investment/GDP0.00010.01280.00170.0122
(0.004)(0.406)-(0.260)(0.398)
Government consumption/GDP–0.092–0.069–0.068–0.049
(3.528)***(2.735)***(2.655)***(2.057)**
Rule-of-law index0.01110.00230.01300.0064
(1.374)(0.300)(1.638)(0.822)
Openness measure0.01360.01490.01410.0159
(3.046)***(3.500)***(3.266)***(3.771)***
Inflation rate–0.0212–0.0263–0.0191–0.0192
(2.644)***(3.641)***(2.838)***(3.406)***
Growth rate of terms of trade0.0690.0520.0720.062
(2.594)***(1.998)**(2.706)***(2.410)**
Contemporaneous IMF loan–0.185–0.183–0.178–0.071
(3.000)***(2.846)***(2.008)**(0.789)
Lagged IMF loan–0.1170.099–0.2140.074
(1.715)*(1.323)(2.027)**(0.818)
Contemporaneous IMF loan*–0.328–0.390
prolonged user(2.899)***(3.062)***
Lagged IMF loan*–0.528–0.517
prolonged user(4.663)***(4.416)***
p-value
(a)0.0020.0110.0070.536
(b)0.0000.000
Sources: IMF, WEO database; ICGR database; World Bank, WDR database; and IEO calculations.*, **, and *** indicate significance at the 10 percent, 5 percent, and 1 percent levels, respectively.

A second set of exercises controlled for the endogeneity of participation in IMF arrangements, using the set of instrumental variables employed in Barro and Lee (2002). There was little difference in results when no distinction was made between prolonged and “temporary” users (compare equations 3 and 1 in Annex Table 4.1); the effects of IMF lending on growth were found to be still negative and significant.5 This result contrasts with the finding in Barro and Lee (2002) that after controlling for endogeneity of participation in IMF arrangements, the contemporaneous effect on growth becomes insignificant. A likely source of the difference in results is the difference in coverage of IMF arrangements, demonstrating the sensitivity of findings of such cross-country regression exercises to sample coverage and size.

When a distinction was made between prolonged and “temporary” users, the main change in results was with respect to the estimated coefficient on the contemporaneous IMF loan size. The estimated coefficient was no longer significantly different from zero. The coefficients on lagged IMF lending and the interactive terms between IMF lending and the prolonged use dummy did not change much.

A third set of exercises examined whether the effects of IMF arrangements on growth differed between arrangements supported by general resources (i.e., SBAs and EFFs) and those supported by concessional resources (SAF/ESAF/PRGF). The results indicate significant differences (Annex Table 4.2). When the sample was limited to only SBAs and EFFs, strongly negative contemporaneous and lagged effects on growth were found in prolonged users but not in “temporary” users. When only concessional facility arrangements were considered, there was a negative contemporaneous effect on growth which was more than offset by a positive lagged effect in prolonged users, and no significant effect on “temporary” users.

Annex Table 4.2.Alternative Specifications of Equation (4) in Annex Table 4.1
SBAs and EFFsSAFs, ESAFs, and PRGFs
(1)(2)
Contemporaneous IMF loan0.043–0.043
(0.326)(0.415)
Lagged IMF loan0.0820.328
(0.888)(1.116)
Contemporaneous IMF loan * prolonged user–0.542–0.677
(3.250)***(1.913)*
Lagged IMF loan *–0.5840.853
Prolonged user(4.761)***(1.760)*
p-value
(a)0.8560.497
(b)0.0000.086
Sources: IMF, WEO database; ICGR database; World Bank, WDR database; and IEO calculations.Note: The estimation is based on the basic specification of equation (4) of Annex Table 4.1 with the specific change indicated in each column.*, **, and *** indicate significance at the 10 percent, 5 percent, and 1 percent levels, respectively.

Sample size limitations imposed by available data constrained the scope of the exercises undertaken by Professor Lee. As noted above, the results from such cross-country regression exercises can be sensitive to changes in the composition and size of the sample being studied. Bearing in mind these inevitable limitations the main findings were:

  • After controlling for endogeneity of participation in IMF arrangements, IMF lending was found to have negative effects on growth, over the contemporaneous as well as subsequent five-year period, in prolonged users.

  • For “temporary” users, the effects on growth of contemporaneous and lagged IMF lending are statistically insignificant.

  • The adverse consequences for growth of prolonged use appear to be concentrated in programs supported under general resources, and not in those under concessional facilities.

Notes

The authors argue that the generalized evaluation estimator approach, characterized by Haque and Khan (1998) as the “estimator of choice” for evaluating the effects of IMF-supported programs, does not adequately correct for selection bias (e.g., by reliance on fragile assumptions about the distribution of error terms for identification). They propose a set of political and institutional variables for use as instruments to control for the endogeneity of participation in IMF arrangements.

Political proximity is measured by voting record at the United Nations, and economic proximity by the ratio of bilateral trade to GDP.

In the broader sample used by Barro and Lee (2002), other measures such as program approval, or program participation (the fraction of time that a country operated under an IMF program during the five-year period) do not seem to have a significant impact on growth independently of loan size.

The definition of “prolonged users” was the same as that used in the “dynamic” definition in Annex 3, section on “Econometric Evidence on the Characteristics of Prolonged Users.” An alternative approach to exploring distinctions between prolonged and “temporary” users would have been to separate the data into two samples and estimate separate regressions for each group. The sample size for prolonged users was too small to implement this approach.

Barro and Lee (2002) considered only Stand-By (SBA) and Extended Fund Facility (EFF) arrangements, while the current exercise also includes arrangements under the IMF’s concessional facilities (i.e., Structural Adjustment Facility (SAF), Enhanced Structural Adjustment Facility (ESAF), and Poverty Reduction and Growth Facility (PRGF) arrangements).

Annex 5: Questionnaire Sent to Authorities of Prolonged User Countries

To check how representative of the broader group of prolonged users the findings of the country case studies were, the IEO sent a questionnaire to the authorities of all the countries identified as prolonged users in this study, as listed in Chapter 2.

Responses were received from the following 21 countries: Bulgaria, Costa Rica, Egypt, Ghana, Jamaica, Jordan, the Kyrgyz Republic, Malawi, Mali, Mexico, Mongolia, Morocco, Nicaragua, Pakistan, Peru, the Philippines, Senegal, Tanzania, Turkey, Uganda, and Zambia.

A copy of the questionnaire is reproduced below. Most respondents indicated that they did not want to be quoted directly, but the thrust of the views expressed are reflected in the main report, especially in Chapter 5.

Overview

  • 1. What is your general assessment of your country’s relations with the IMF over the long term?

  • 2. What do you see as the major factors that explain why your country made extended use of IMF resources? Could or should this have been avoided? What should the IMF have done differently? What has your country learnt from the experience of repeated programs?

  • 3. To what extent were IMF-supported programs for your country motivated by the need for a “seal of approval” in order to mobilize funds from other sources, rather than a need for IMF financing, per se? Would it have been feasible or preferable to provide such a “seal of approval” in some other way?

Program formulation and negotiation

  • 4. Did the IMF pay sufficient attention to the concerns of the authorities and other groups in the formulation and negotiation of programs? Were any disagreements on policies generally concerned with their substance, or on the pace and sequencing of measures or to potential difficulties in implementing programs?

  • 5. Was the IMF realistic about the political and social environment of programs and the constraints involved?

  • 6. How did the IMF’s prolonged involvement affect the development of economic institutions—including those involved in policy formulation and technical analysis—in the country?

Program design

  • 7. What, in your view, were the major strengths and weaknesses in the design of IMF-supported programs? Were IMF-supported programs too ambitious or overoptimistic? Did IMF-supported programs have an appropriate time-horizon? Did they pay sufficient attention to debt sustainability issues? Did programs make sufficient allowance for exogenous shocks?

  • 8. Did IMF-supported programs put the emphasis on the right structural reforms and prioritize appropriately? Was there an appropriate division of labor between IMF and the World Bank with regard to structural reform?

  • 9. Did the IMF learn from experience in designing successive programs?

Post-program experience

  • 10. In those cases where countries no longer use IMF resources, has the internal political dynamic altered since there has been no lending arrangement? Has the process of policy-making and related technical analysis process altered?

  • 11. For those countries that have made repeated use of precautionary arrangements, what are the main reasons for such an approach? What advantages do you see for a precautionary lending arrangement over regular IMF surveillance?

  • 12. Are there any other issues you would like to bring to our attention?

Annex 6: Data on Staff Inputs and Staff Turnover in Prolonged and “Temporary” Users

This annex provides additional information to support the discussion in Chapter 5, section on “Results from Cross-Sectional Evidence.”

Annex Table 6.1 provides details of the extent of staff inputs, as measured by the number of missions and mission days, in program countries. The results indicate such inputs were actually higher for “temporary” users. The difference is particularly marked for ESAF arrangements, which involved on average 51 mission days (41 percent) more for “temporary” users than for prolonged users. Likewise, the total staff resources invested by the IMF in programs with prolonged users were, on average, smaller than in “temporary” users’ programs: in both ESAF and GRA arrangements, the IMF’s effort, measured by the personnel costs of its UFR and TA missions, was over 40 percent higher in programs with “temporary” users.

Annex Table 6.1.Data on IMF Effort1
Number of mission daysNumber of missionsAdministrative costs (In millions of U.S. dollars)
During programIncluding three months before program approvalDuring programIncluding three months before program approvalDuring programIncluding three months before program approval
All arrangements
Prolonged users1261449111.51.6
Nonprolonged users16318612141.92.1
ESAF arrangements
Prolonged users1221409101.51.7
Nonprolonged users17318914152.22.5
GRA arrangements
Prolonged users1171389101.21.4
Nonprolonged users16018512141.72.0

Excessive turnover of mission chiefs appears to be a problem for many program countries, but has not been worse among the prolonged users (Annex Table 6.2).

Annex Table 6.2.Mission Chiefs Per Member Country, FY1996–2001
No UFRTU/PRGFTU/GRAPU/PRGFPU/GRA
(Number of mission chiefs)
Mission chiefs per member country
Average3.44.24.13.94.2
High671078
Low12221
Share of member countries with five or more

mission chiefs during six-year period (in percent)
1334321931
Source: Internal data compiled by the IMF’s Office of Internal Audit and Inspection at the IEO’s request, based on data collected for its review of mission organization and management.

As regards mission team staffing, continuity has also been low across all country groups—in most cases, less than half of mission members were involved in the same country in the two previous years—but it has been slightly better in prolonged user countries than in “temporary” user countries (Annex Table 6.3).

Annex Table 6.3.Mission Staff Continuity, FY1996–2001(In percent of total mission staffing)
Current fiscal year area department staff active in prior two fiscal yearsCurrent fiscal year FAD staff active in prior two fiscal yearsCurrent fiscal year PDR staff active in prior two fiscal years
199819992000200119981999200020011998199920002001
No UFR414639400025203306733
TU / PRGF474439423140333936464828
PU / PRGF525548474742413923324041
TU / GRA504351454740474025404042
PU / GRA495259555033505723365036
All countries484747454338414127384536
Source: Internal data compiled by the IMF’s Office of Internal Audit and Inspection at the IEO’s request, based on data collected for its review of mission organization and management.Note: FAD: Fiscal Affairs Department; PDR: Policy Development and Review Department.

    Other Resources Citing This Publication