Chapter 6: Influence of IMF Governance and Other Institutional Factors on Prolonged Use
- International Monetary Fund. Independent Evaluation Office
- Published Date:
- April 2003
1. In this chapter, we examine governance issues that are relevant to the phenomenon of prolonged use. These include institutional biases that can affect program design and contribute to prolonged use, including the issue of whether program activities have “crowded out” surveillance in these countries, and the implications of the IMF’s role in providing a “seal of approval” on policies.1
Impact of the IMF’s Institutional Culture on Program Design
2. Some of the problems of program design identified in the previous chapter as contributing to prolonged use have their roots in the institutional culture of the IMF. These institutional characteristics are obviously not specific to prolonged users: they constitute the background to all IMF-supported programs and there is no straightforward and probably no unique explanation of why these institutional features led to prolonged use in some cases and not in others. However, any discussion of the causes of prolonged use that ignored these factors would be incomplete. We note, however, that recent initiatives have begun to address a number of the problems discussed below, and their success would help to reduce the occurrence of prolonged use in the future.
The treatment of institutional reforms in program design
3. Two factors that are critical to the sustainability of any adjustment effort, namely institutional change and good governance, have received too little attention until recently. The approach toward structural reforms adopted from the late 1980s onward often led to an overload of the reform agenda—a phenomenon observed in the country case studies—that resulted in a de facto failure to focus on the reforms that were most critical from the perspective of long-term sustainable adjustment and that often depended critically upon institutional development. (See Chapter 5 for a more extensive discussion of this point.) Governance problems, which were also important in all three case studies, were not discussed openly until the late 1990s, when the Executive Board adopted guidelines setting out the role of the IMF in governance issues. The case of the Philippines toward the end of the Marcos era, when governance issues were explicitly raised, was an exception.2
4. Mindful of these problems, the IMF in 2001 launched a review of conditionality.3 Although the review is still under way at the time of this evaluation, it has already led to a strong drive toward streamlining conditionality, in particular through a stricter application of the macroeconomic criticality test. By narrowing the scope of structural conditionality, this initiative should pave the way for programs that incorporate a greater prioritization of structural reforms. We would, however, emphasize that streamlining by itself is no guarantee that the few reforms pursued in each program will truly be the most critical ones, nor that they will be tackled in sufficient depth. Meeting these two challenges requires a strengthening of the treatment in program design of the key structural reforms, and a much closer and more effective collaboration with the World Bank than has generally been the case until now.
5. All three case studies reveal that one factor contributing to prolonged use was the presence of a few seemingly intractable structural problems, including in the areas of tax administration, public enterprises, and administered prices, that hampered adjustment. The IMF frequently approached those issues from the narrow angle of their direct—often just fiscal—macroeconomic implications, while the World Bank concerned itself with the design and implementation of the broader structural reform.4 In practice, this joint involvement sometimes gave rise to coordination problems that in turn led to weaknesses in program design. Such problems included overlapping—and on occasion conflicting—policy advice, ex post inadequate sequencing of reforms, or de facto cross-conditionality, all of which contributed to the ineffectiveness of the reform process in these areas. For example, in Pakistan, at one point the World Bank and the IMF had different targets for power prices, the Bank’s target being geared to achieving a specific rate of return on investment, while that of the IMF was motivated by fiscal considerations. In Senegal, progress in the disengagement of the state from the groundnut sector, which was pursued in successive IMF-supported programs for fiscal reasons, was long hindered by the lack of progress of reforms targeted at the reorganization of the sector, even under World Bank sector and adjustment lending operations.
6. The different time frames on which the two institutions operate has often been blamed for these recurrent difficulties, along with various procedural differences. However, in the case of prolonged users, both the World Bank and the IMF have had de facto a long-term involvement, which makes time frame conflicts a not fully convincing explanation. While a full assessment of the causes of these problems is beyond the scope of the current evaluation, they appear to lie much deeper than a mere failure of the two staffs to coordinate adequately; they reflect different institutional cultures, program modalities, and objectives—and hence will be harder to resolve. Indeed, a number of the difficulties in tackling structural problems in a coordinated manner occurred in cases where the staff of both institutions characterized the working relationship as good.5
7. Several steps have been taken in the recent past to clarify further the division of labor between the two institutions and devise concrete mechanisms and procedures that should foster a more efficient collaboration, in particular the designation of a “lead agency” in each policy area of common interest, with close interactions and accountability built-in.6 Most of the prolonged users surveyed as part of the evaluation recognized that there had been some improvement, because of the streamlining initiative and the PRSP process, but they emphasized that much more was needed. Looking ahead, the key challenge, in order to achieve sustainable adjustment earlier and thereby minimize prolonged use, is to ensure that the two institutions and the authorities will be able, jointly, to identify the limited number of structural reforms which are key to long-term macroeconomic sustainability and sustainable growth; to collaborate effectively in designing the most appropriate strategy for implementing such reforms; and to monitor their implementation through appropriate, but parsimonious, conditionality. This will require a much more effective meshing of the priorities and work programs of the IMF and the World Bank than in the past.
Insufficient attention to assessments of political feasibility
8. The case studies make clear that ownership and implementation capacity were often the Achilles’ heel of IMF-supported programs in these countries. These were also the reasons most frequently cited by IMF mission chiefs participating in the IEO survey when asked to rank the reasons why the programs they had negotiated or overseen had been unsuccessful. Many country authorities in prolonged user countries also expressed the view that the underestimation by the IMF of the technical and political limits to implementation capacity, and the consequent overoptimism about the speed of success, was an important reason for their prolonged use of IMF resources.
9. Political feasibility is clearly closely related to ownership but is not identical to it, and in principle one could argue that the IMF should make a realistic assessment of political feasibility of any program as an essential aspect of ensuring credibility, whatever the level of ownership.7 It is certainly possible to en-visage situations where programs are owned by the government but are not politically feasible.
10. Surveys of authorities in prolonged users and IMF mission chiefs indicate that the latter are not insensitive to domestic feasibility constraints. Seventy percent of mission chiefs surveyed reported that, in their experience, the frequency with which final program design departed from the initial briefing in order to enhance ownership or take account of political constraints was more than occasional. Furthermore, most mission chiefs reported that they did not consider their performance assessments to be adversely affected when program negotiations departed from their brief. But the frequency of such departures seems to be positively correlated with the experience level of mission chiefs. About three quarters of mission chiefs with experience in three or more countries reported such departures “sometimes” or “frequently,” compared with less than half of chiefs with experience in less than three countries (Figure 6.1).
Figure 6.1.Consideration Given to Ownership in Program Design
Source: IEO survey of IMF mission chiefs.
11. Many officials and other stakeholders from the prolonged user countries acknowledged that there were wide variations in approaches and that the most effective mission chiefs did invest considerable time in understanding the political situation and consulting with a broad range of participants, but they said that in general too little attention was paid to such issues. Discussions with IMF staff and reviews of internal documents (e.g., back to office reports routinely prepared following missions) also suggest that much more mission time is now devoted to discussions with various political and social groups in an effort to gauge, and enhance, the degree of consensus behind particular reform agendas. Any meaningful reforms will always attract some opposition and, as a number of staff stressed in our discussions, it is important to distinguish between genuine concerns about political feasibility and narrower considerations of short-term political inconvenience that may lead governments to delay reforms when the prospective benefits are only likely over the long term. Several internal documents reviewed by the IEO in the context of case studies suggest that IMF staff often has a good understanding of the political economy of program implementation and the potential risks. However, these considerations often did not surface in staff reports to the Executive Board.
12. The extent to which the IMF could or should analyze political factors in a more explicit and systematic way as part of its decision-making procedures on programs is an important but difficult question to address. A paper prepared at the request of the IEO by a team of political scientists illustrates some of the tools that might, in principle, be used in such political feasibility assessments, although incomplete information obviously limits how much such tools could be applied ex ante in highly complex situations. Nevertheless, the paper suggests that the application of some of these analytical tools to the Pakistan programs of 1993/94 and 1997 would have raised significant doubts about the government’s ability to implement them.8 It is impossible to say with certainty whether such analyses would have led to different decisions regarding the IMF’s involvement or program design, especially since there was a strong and understandable tendency to give new governments the benefit of the doubt. However, such analysis might have enabled staff, management, and ultimately the Board to make better-informed judgments and could have influenced the content of programs. Chapter 8 discusses further how the IMF’s handling of political feasibility in its decision-making process might be improved in the future.
Political influences on IMF decision making
13. Another factor that may affect IMF decisions and underlie the phenomenon of prolonged use is “political” pressure from influential shareholders. Pakistan may be the clearest example of a case where the decision-making process of the IMF has come to be perceived as politically driven by many stakeholders in Pakistan.9 Such perceptions were also present (although to a lesser extent) in other country case studies. More generally, political pressures were the third most frequent factor cited by mission chiefs participating in the IEO survey in explaining less-than-successful outcomes in the programs they negotiated or oversaw.
14. The impact of political considerations in the IMF decision-making process cannot be measured rigorously, but the survey of mission chiefs suggests that they are neither rare nor widespread. Only 7 percent of mission chiefs surveyed reported that their technical judgment regarding support for a UFR request had been overridden by political pressures “frequently” or “always.” However, as many as 48 percent reported that they had experienced such strong pressures “occasionally” or “sometimes.”
15. It is neither practicable nor desirable to expect that political considerations could be completely removed in an institution whose decisions have to reflect the views of its shareholder governments. However, a problem arises if political considerations are seen to overwhelm technical considerations, leading the IMF to support programs that have a low probability of success. This would inevitably raise concerns that the principle of uniformity of treatment across countries may not be upheld and it could also encourage a lax approach to implementation by the borrowing country. In our view, the focus of action should be on introducing greater clarity and transparency about the nature of the judgments that need to be made and who should make them. The existing guidelines state that the Managing Director of the IMF will recommend that the Executive Board approve a member’s request for IMF resources “when it is his judgment that the program… will be carried out.”10 This presumes that an absolute judgment can be made on technical grounds. In practice, however, all such judgments—about the likelihood that a program will be implemented and whether it will achieve its objectives if implemented—can only be made in a probabilistic sense, based on a clear assessment of the risks and trade-offs for both the member country and the world community. The present procedure does not make a sufficiently sharp distinction between the technical assessment of the risks and the judgments involved in weighing those risks, which is where any political considerations should presumably enter. This situation has important drawbacks. First, there is no formal and transparent channel through which political judgments on this balancing of risks can be suitably fed into the process before the final stage of Board approval even though it is at earlier stages that political pressures may be greater.11 This was confirmed by the survey of mission chiefs, significantly more of whom indicated that political pressures occur at key times before the Board discussion rather than at the Board itself (Figure 6.2). Second, the line of accountability between staff, management, and the Board becomes blurred. Moreover, the lack of transparency can give rise to exaggerated perceptions of political pressures, which are likely to weaken the effectiveness of IMF-supported programs.12
Figure 6.2.Timing of Political Pressure
Source: IEO survey of IMF mission chiefs.
11: Drafting mission brief; 2: Clearing brief; 3: Negotiating program; 4: Preparing Board meeting; 5: Board meeting.
16. We discuss in Chapter 8 what might be done to deal with this problem. The aim should be to ensure that such political judgments—which are an inevitable part of decisions on whether or not to proceed with programs whose outcomes are uncertain—should be clearly distinguished from technical judgments and should be made in a transparent manner at the level of the Managing Director and the Executive Board, who are accountable for them.
Incentives to overpromise in programs
17. The IEO survey of mission chiefs suggests that internal incentives create a tendency to over-promise. As discussed in Chapter 3, until the late 1990s, the rules governing the use of IMF re-sources implied that all programs submitted for Board approval had to show at least substantial progress toward viability within the program period, along with net repayments by the member after the expiration of the program. In practice, the IMF was often placed in a situation where it found it desirable to enter into an arrangement with a member even though there were good reasons to doubt that these viability requirements would be met. In a number of cases, this appears to have led to projections of a return to external viability over the medium term that were overoptimistic not just ex post, as discussed in Chapter 5, but also ex ante. About 45 percent of mission chiefs surveyed by the IEO reported that the need to show balance of payments viability by the end of the medium-term projection period had led to ex ante overoptimistic projections or overambitious program objectives “frequently” or “always,” with another 28 percent reporting the same phenomenon “sometimes.” In contrast, only about a quarter reported this practice occurring “never” or only “occasionally” (Figure 6.3).
Figure 6.3.Incentives Toward Overoptimism
Source: IEO survey of IMF mission chiefs.
18. While optimistic forecasts may appear to be supportive of borrowing countries in need of assistance in the short run, since they enable a program to be approved, their long-term effect is likely to be adverse. As the case of Pakistan illustrates, they create conditions in which programs are virtually certain to go off-track when the optimistic projections do not materialize, leading to new programs that often meet the same fate. This creates a cumulative impression of poor implementation on the part of borrowing countries and also poor program design on the part of the IMF, eroding the credibility of both. More realistic projections would help to identify financing needs more accurately, both in terms of the volume of assistance needed and the length of time for which support may be necessary. If assistance on the scale required is not feasible, it would at least lead to more realistic assessments of likely outcomes in terms of performance, which would enable the international community to determine whether these are indeed acceptable or whether additional support can be mobilized.
19. A significant proportion of staff members interviewed in the course of the country studies were also of the view that the internal review process gives a premium to “toughness” over realism in the setting of program targets, leading to targets that were more likely to be missed. The survey of mission chiefs further indicated that close to 30 percent of survey participants felt somewhat strongly that their individual performance appraisal would be better if they were “tougher” in negotiations with the authorities, with only 17 percent in clear disagreement with that view (Figure 6.4).13
Figure 6.4.Incentives Toward “Toughness” in Program Design
Source: IEO survey of IMF mission chiefs.
20. The primary purpose of IMF-supported programs, as stated in the Articles of Agreement, is to “give confidence to members,” and this objective probably contributed to the tendency to underemphasize downside risks in documents presenting a UFR request or a program review to the Board. The case studies—especially Pakistan—show that even when those downside risks were identified at the stage of the interdepartmental review process, which was frequently the case, they were subsequently toned down in the final Board documents, as were any disagreements between functional and area departments on the appropriate adjustment path or program design. While the ability of the IMF to “speak with one voice” is one of the strengths of the institution, it should not be at the expense of full presentations of the risks to a program, an exploration of policy alternatives, and a frank assessment of implementation capabilities.
21. On the authorities’ side, the case studies suggest that there can be tactical incentives for knowingly agreeing to a program that overpromises in order to achieve a release of external financing. For example, in the particular case of fiscal revenue targets, ex ante overoptimism sometimes made it easier to reach a domestic consensus on the magnitude and design of fiscal adjustment (since overoptimistic revenue targets postponed difficult decisions on expenditure cuts).
The role of surveillance
22. In theory, the exercise of surveillance in program countries should provide an opportunity for a “reality check” to compensate for the incentives discussed above. However, the evidence from the case studies suggests that some important aspects of surveillance have been weakened in prolonged use cases. Since the early 1980s, Article IV consultations in program countries have generally been conducted jointly with UFR-related discussions. The case studies suggest that, compared to current guidelines, there was often less in-depth medium-term sensitivity and vulnerability analysis, less exploration of the possible trade-offs between different policy options, and a less candid discussion of divergences of views between the staff and the authorities (Box 6.1).14 This appears to stem from in-built concerns not to “rock the boat” by including any analysis that might undermine the credibility of the program and its desired catalytic effects.
23. In countries where programs are occasional events, the combination of Article IV consultations and UFR discussions may also reduce the quality of surveillance, but the impact is temporary. It is potentially much more significant for countries that have a long series of IMF-supported programs, as a weakening of surveillance in such cases could allow vulnerabilities to build up without appropriate warning signals being sent. The buildup of a large uncovered foreign exchange exposure through foreign currency deposits in Pakistan, which culminated with a deposit freeze in 1998, is one illustration of this risk.
The IMF’s learning culture could be improved, including through more systematic program assessments
24. Despite various policy requirements and frequent calls by the Board for more “stocktaking” opportunities, there are too few occasions when the institution steps back on country programs to re-consider its overall strategy. One attempt at creating systematic opportunities for staff to step back from routine program work and think strategically about the IMF’s involvement in the member country was made by management in 1992, with the creation of internal “country strategy papers” (CSPs). However, experience with that instrument proved unsatisfactory.15 In response, a revamped CSP strategy was prepared in 1997, which called for such papers to provide essentially two elements: (i) a critical and frank review of the latest arrangements, covering not only whether program targets were met, but also the appropriateness of program design; reasons for the success of, or problems in, program implementation; the progress made relative to the ultimate objectives of the country; the remaining issues to be tackled; and an assessment of the effectiveness of the collaboration with World Bank staff; and (ii) a forward-looking strategy for future IMF involvement, taking account of the lessons drawn in part (i) and focusing on the broad pattern of macroeconomic adjustment and key areas of structural reform (including a discussion of the rationale for their choice, phasing, and prioritization).
Box 6.1.Factors Used to Assess the Quality of Surveillance in Prolonged Use Cases
To assess the quality of surveillance in the country cal-case studies, we rated the performance of each surveillance report for nine functions viewed by the IEO as “key elements” of surveillance in a program context.1 These nine functions are (i) provision of realistic medium-term and alternative scenarios; (ii) provision of meaningful sensitivity analyses; (iii) discussion of risks to the assumptions and projections; (iv) discussion of the risks and impact of policy slippages and of vulnerabilities; (v) balanced reporting of the authorities’ views, including any significant differences with staff; (vi) cogent presentation of proposed policy course; (vii) discussion of policy alternatives and trade-offs; (viii) critical and frank review of previous UFR performance; and (ix) presentation of collaboration/interaction with the World Bank.
It must be recognized that, to some extent, such a comparison does involve judging previous surveillance exercises by current standards. More specifically, whereas all the functions highlighted above have been part of surveillance requirements since at least the early 1990s, and often earlier, the emphasis they have received has tended to increase over time. For instance, discussions of risks and vulnerabilities received a much stronger emphasis after the 1995 Mexican crisis, and an even stronger one since the 1997/98 Asian crisis. In contrast, the requirement that staff prepare an analysis and assessment of performance under IMF-supported programs in connection with Article IV reports, adopted in 1979, was never emphasized in sub-sequent surveillance guidelines.
As noted in the main text, the overall assessment is that many of these key functions of surveillance were not fully implemented in the prolonged use case studies. However, there is some indication that the quality of surveillance in the case study countries did improve over time—most clearly in the Philippines—with respect to some of the key functions identified here, particularly as regards the identification of vulnerabilities. Nevertheless, the main conclusions that some key surveillance functions have not been fully implemented in prolonged use cases remain valid.1 These nine functions draw on the “minimum requirements” of Article IV reports as identified on the PDR’s web-site (which reflects the guidance given by the Executive Board in successive biennial reviews of surveillance up to the 2000 exercise as well as in ad hoc discussions of specific aspects of surveillance policy), as complemented by three other sets of internal guidelines of particular relevance for program countries (the 1979 “Guidelines on Conditionality”—especially guideline 11; the 1990 note on “Assessments of a Member’s Ability to Repay the Fund” and the 1997 note on “Country Strategy Papers,” both issued by PDR) and the 1998 instruction note on Report Writing by the First Deputy Managing Director. These internal instructions add a qualitative dimension to the general guidelines, which themselves are primarily concerned with coverage.
25. The documents reviewed as part of the country case studies suggest that the quality of CSPs produced since this revamped approach did improve significantly. However, relatively few were produced (in total 24 full CSPs and 3 partial ones between mid-1997 and end 2001—that is, less than 6 a year on average), primarily due to workload constraints. Thus, most programs still do not benefit from any stocktaking exercise when they expire, and those that do are not necessarily the ones for which there would be the greatest need to learn from experience (i.e., in particular, programs that went irretrievably off-track). Moreover, those prepared for the case study countries suggest that the judgments in CSPs were often not fully reflected in subsequent program or surveillance documents.
26. In contrast with specific country programs, experience with IMF general policies has been subject to quite intensive and usually high-quality reviews. Many of these reviews have drawn clear lessons from their analyses that are of relevance to program design in general, including for prolonged users.16 However, the process often stopped at the diagnostic stage and some of the lessons identified subsequently took a long time to be actually incorporated into program practices. For instance, several staff papers from the mid/late 1980s identified the lesson that the authorities’ commitment was key to the success of any program and that, if it was lacking, efforts to optimize the program design were bound to be of limited effectiveness.17 Similarly, most of the ingredients of robust and realistic sustainability assessments mentioned in a paper discussed by the Executive Board in June 2002 that proposed a formal framework for mainstreaming “best practices” in that area, were actually identified and circulated as guidelines to area department staff as early as 1990, with a view “to avoid well-known traps—such as scenarios which seek to establish viability by the use of optimistic assumptions. There is a need to be open about the possible dangers.”18 Many of these reviews also drew useful lessons for program design that have yet to become fully operational—for example, the importance of an early focus on institutional reforms, particularly for sustainable fiscal adjustment.19
27. Thus, the IMF often seems to have been quite good at identifying lessons, but less effective at ensuring that they were fully absorbed into its everyday operations. The experience of the country case studies suggests that this slow absorption of lessons and broader policy guidance into actual operations on a systematic basis contributed to weaknesses in program effectiveness and hence to prolonged use. However, it is clearly not a phenomenon that is special to prolonged use cases.
28. These shortcomings in the IMF’s learning culture are compounded by two human resource management issues. First, as noted in Chapter 4, the internal turnover on country assignments is high. While not specific to prolonged users, this is a significant impediment to the development of an adequate learning culture, not least because many of the lessons to be learned from experience are country specific. Thus, even if the lessons from policy review were fully and swiftly incorporated into the IMF’s operational practices, an important gap would re-main. Second, high staff mobility limits the accountability of mission chiefs: when mission chiefs have a shorter tenure, on average, than the time span of most programs, their contribution to the outcome of a given program becomes harder to assess. Clearly, program success depends to a very large extent on factors beyond the mission chief’s control, but an effective mission chief should make a difference. Yet, results of the IEO survey indicate that most mission chiefs do not feel that their career progression depends significantly on whether the programs they negotiate and oversee achieve their objectives (Figure 6.5).
Figure 6.5.Mission Chiefs’ Career Progression and Program Outcomes
Source: IEO survey of IMF mission chiefs.
The “Gatekeeper” Role Assigned to the IMF with Respect to Many Other Sources of Official Financing Also Contributed to Prolonged Use20
29. In addition to the internal factors affecting program design that lead to ineffective programs and thus encourage prolonged use, there is a systemic factor that has the same effect: the gatekeeper role of the IMF in providing a signal for other resource flows. There is considerable empirical evidence that aid is most effective when a sound economic policy framework, including a sound macro-economic environment, is in place. The IMF is the international institution most suited to providing certification regarding the soundness of macroeconomic policies, and the main multilateral lenders and most major bilateral donors have come to rely upon the IMF to give this signal when making decisions about their program and adjustment lending. However, the expectation seems to be that the signal will be conveyed in the form of an IMF lending arrangement. Since many countries expect to rely upon bilateral and multilateral flows for an extended period, the requirement that there should be an IMF-supported program in such cases is a recipe for perpetuating prolonged use.
Is IMF financing necessary for a seal of approval?
30. It is appropriate to question whether creditors/donors should insist on an IMF-supported program as the only credible seal of approval or whether the same objective could be achieved in some other way, for example, through the surveillance process or staff’s assessments of countries’ PRSPs. In their responses to the IEO, official donors and creditors often justified their preference for a program on the grounds that a lending arrangement provides a clearer and more reliable assessment.21 For creditors—especially the Paris Club and the private sector—“burden sharing” issues may also be important: some of these groups have been reluctant to provide additional financing if large net repayments are being made to the IMF or if the absence of new IMF financing implies that other new sources of financing will not be forthcoming.22
31. These considerations are reinforced by the fact that, historically, the Executive Board has been rather reluctant to generalize the recourse to inter-mediate signaling tools, such as “enhanced surveillance” in the 1980s, staff-monitored programs (SMPs) and other “shadow programs.”23 It is also possible that the various instruments of “strengthened surveillance” were not regarded as providing sufficient assurances to donors because of the lack of clearly defined standards for such instruments. In some cases, these instruments were resorted to when the member’s program fell short of what could be supported by a lending arrangement; in other cases, the member’s program met or even exceeded UFR standards, but the member did not want to be perceived as being in need of IMF resources.
32. It is worth noting that, unlike the official sector, private creditors have become less dependent on IMF-supported programs as a “seal of approval.” Such a signal was needed to enable a debtor country to negotiate rescheduling or restructuring agreements with its private creditors in the decade or so after the debt crisis of the early 1980s and this contributed to prolonged use at the time, for example, in the Philippines. However, this factor has been much less important recently, owing to less widespread debt-servicing difficulties, together with an increasing tendency on the part of the private sector toward one-time restructurings of the debt stock or debt exchanges, rather than a succession of annual arrangements.
33. Nearly all the private financial institutions surveyed as part of this evaluation considered an IMF-supported program to be highly desirable for a debt restructuring or a coordinated debt rollover, but not absolutely necessary.24 They observed that, historically, a number of debt restructurings had been completed where an IMF-supported program was not possible. Examples include the reschedulings for a number of East European economies and for South Africa in the 1980s, and for the Islamic Republic of Iran in the early 1990s. In the view of some of the respondents, these reschedulings were “among the most successful.” Respondents also noted that the timing of restructurings by private creditors need not be tied as closely to IMF-supported programs as, for example, Paris Club reschedulings typically are. There are many examples of debt restructurings by private creditors that were completed when IMF-supported programs had gone off-track.25
34. For countries that are not facing immediate debt-servicing difficulties, many private investors and lenders take the country’s relationship with the IMF into account, directly or indirectly, in their investment and risk management decisions, but IMF financial involvement is not a necessary signal.26 Several alternative mechanisms are available for that purpose, for example, precautionary arrangements, staff-monitored programs, and IMF surveillance through Article IV consultations—whose signaling role has been enhanced in recent years as a result of improvements in transparency, with many such reports now being published. These alternatives vary in the degree to which they imply formal IMF approval of policies. The Executive Board has traditionally emphasized that, unlike the so-called enhanced surveillance procedure or precautionary arrangements, staff-monitored programs do not constitute IMF endorsement of the country’s policies. Internal reviews of experience have shown that these distinctions may not always be clear to private market participants. Other factors mentioned as limiting the usefulness of surveillance reports to the private sector include the considerable variability from country to country in terms of disclosure and publication of reports, a lack of timeliness, and limited frequency. In this respect, staff-monitored programs or precautionary arrangements can add value by providing more frequent assessments of developments, at the time of periodic program reviews. In general, most respondents to the IEO survey found IMF surveillance to be a useful input into their risk management assessments.
35. Jamaica’s experience—discussed in more detail in Chapter 12—is interesting since it was able to maintain access to private markets on relatively good terms despite publicly ruling out new lending arrangements with the IMF. Senior Jamaican officials interviewed indicated that a commitment to transparency—including publishing the IMF surveillance reports, even when they did not fully share the assessment, and explaining the reason for their different approach—had helped maintain good private market access.27
36. The key reason why considerations of an IMF “seal of approval” for the private sector is rarely the determining factor behind prolonged UFR is that most investors and lenders ultimately rely on their own judgments, for which the assessments and analyses of the IMF are only one source of information. The views of other institutions—including credit rating agencies—often also play an important role. Some institutions also expressed reservations about relying too much on the IMF’s judgment, in part because of misgivings about its impartiality, noting its vested interest in helping its members.
37. An important question is whether a prolonged IMF program involvement weakens the credibility of the signals provided by lending arrangements and hence reduces its catalytic effect. It was not possible, in the context of this project, to test rigorously this question, partly because available cross-country evidence on the existence of any catalytic effect is inconclusive.28 However, among the studies which found a positive correlation between IMF-supported programs and market access, those which tested for the impact of prolonged use found that the beneficial effect of IMF involvement declined in size with the length of past IMF involvement and that continuing IMF presence was associated with higher spreads.29
Consequences of the gatekeeper role for lending arrangements and surveillance
38. There is some evidence from the case studies that insistence on having the IMF’s seal of approval delivered exclusively in the form of a lending arrangement can compromise the quality of IMF-supported programs, and therefore the quality of the seal of approval. This is because the potential consequences for the member country, in terms of loss of financing, are so serious that pressures build—on both the authorities and the IMF—to agree to programs with a low probability of success.30 There are similar pressures to keep the programs formally on-track as long as possible, often through unsustainable efforts, even when several performance criteria are not met, under the assumption that officially calling the program off-track would trigger adverse donor responses. Such factors contributed to the problems with a number of Pakistan’s programs and have also been an issue in Senegal and the Philippines.
39. Whatever signaling device is used, it is clearly important to resist pressures to compromise quality. But there is a case for considering vehicles other than an IMF lending arrangement in some circumstances. Private sector creditors have already acknowledged this in their approach to lending to developing countries. Official donors surveyed by the IEO generally took a cautious stance, pointing out in particular that requiring an IMF “seal of approval” to target some of their aid flows had been instrumental in improving the effectiveness of their aid, so that any departure from this—sometimes hard-won—policy rule would imply a step backward in this respect. This premise notwithstanding, most acknowledged the above-discussed dangers of an exclusive reliance on IMF lending arrangements to deliver the requested seal of approval. Most were therefore open to considering alternatives, provided that (i) they suitably addressed the information needs of the donor community; (ii) they delivered sound and candid analyses of economic conditions and prospects; and (iii) they provided a clear assessment that the member’s policies are sufficiently strong to be supported. Some possibilities in this regard are explored in Chapter 8.
In addition to the country case studies and the questionnaire sent to the authorities of all prolonged users, the discussion draws upon a survey of IMF mission chiefs. The questionnaire sent to prolonged users is reproduced in Annex 5. The main components of the mission chiefs’ questionnaire and its results are shown in this chapter. We would like to thank the IMF’s Office of Internal Audit and Inspection (OIA) for the help they provided in implementing the latter survey.
Even though these guidelines were adopted following a declaration of the IMF’s Interim Committee on September 26, 1996 that attached particular importance to “promoting good governance in all its aspects,” their adoption was controversial enough for the Fund’s Legal Department to be asked to ascertain their consistency with the Articles of Agreement. The conclusion of consistency was reached based primarily on the fact that the guidelines did not assert that the IMF had a general mandate to promote good governance, but rather identified certain areas of IMF involvement that were referred to as governance. In practice, the guidelines limit the IMF’s involvement to the economic aspects of governance that could have a significant macroeconomic impact. They were reviewed and left unchanged by the Executive Board in early 2001.
However, there were important exceptions. For example, in the Philippines, IMF-supported programs took the lead in the dismantling of sugar, coconut, and other monopolies that were part of the “crony capitalism” legacy of the Marcos era.
The history of cooperation between the IMF and the World Bank is reviewed in more detail in an annex to IMF (2001e).
Additional initiatives are the creation in 1999 of joint Bank-Fund “products,” such as the FSAP and joint staff assessments of members’ PRSPs, along with the adoption of policies and operational guidelines to strengthen Bank-Fund collaboration in country programs and conditionality in 2001 and 2002.
The concept of “ownership” and its possible operational implications were rarely discussed until the last few years. This situation has changed dramatically since 2001, and no fewer than six staff papers directly concerned with that subject have been issued, of which two are research papers and four policy papers discussed by the Executive Board. The most recent papers are IMF (2001d) and Boughton and Mourmouras (2002).
Wimmer and others (2002). The purpose of the paper is to illustrate the analytical tools that are available to address such issues, not to undertake a full-fledged political economy analysis of particular cases. The paper is available on the IEO’s website at http://www.imf.org/ieo. Its main points are summarized in Appendix 1 to the Pakistan case study (Chapter 9).
Many of the staff members interviewed concurred that many of the decisions on the IMF’s involvement in Pakistan since the 1980s had been politically driven to a large extent.
1979 “Guidelines on Conditionality,” guideline 7. The revised guidelines that were under discussion by the Executive Board at the time of writing of this report contain identical language. Guideline 8 further entrusts the Managing Director with a responsibility to ensure the nondiscriminatory treatment of members in the application of UFR policies.
The only such process at present takes the form of “informal” Board discussions, convened in the most sensitive and high-profile cases. But since there is no official record of these discussions, they have little value-added in terms of transparency and accountability.
If only by providing a convenient excuse to all parties involved in the negotiation for the shortcomings of program design, and by blurring the signals sent by the program to economic agents, both within and outside the country.
There were no marked differences across responses according to mission chiefs’ exposure to prolonged users.
The 2002 Review of Surveillance, which looked at the implementation of surveillance in program countries during 2000–2001, came to similar conclusions. See “Biennial Review of the Implementation of the Fund’s Surveillance and of the 1977 Surveillance Decision: Surveillance in a Program Environment” (SM/02/82, Supplement 2, 3/15/2002).
Memorandum from the then Director of PDR to IMF management. The memorandum indicated that, by mid-1997, only 31 CSPs had been produced (of which only 12 concerned prolonged users), and noted that “it is not clear that a number of the CSPs generated value added commensurate with the efforts involved. In addition, several were lengthy and heavy on detail, providing little assessment of operational relevance or guidance in terms of priorities.”
There have been 58 such policy reviews since 1979, covering the entire spectrum of UFR-related issues, from theoretical and operational aspects of program design to access policies, external financing, and the design of various facilities.
See, for instance, EBS/86/211, “Program Design and Performance Criteria,” which noted that “it is important to bear in mind the fundamental limit to the efficacy of program monitoring imposed by the degree of commitment of the member to the policies that make up the program. The commitment of members to their program is essential to their success. While the Fund’s role in providing policy advice is important to the formulation and design of the program, it can only be effective if the member is committed to it and implements it with the consequent resolve.” Also, EBS/87/40, “External Adjustment, Financing and Growth—Issues in Conditionality,” which noted that “the politics of adjustment is often as important as the economics. With benefits of hindsight, it appears that in some particularly difficult situations the capacity to implement adjustment policies was not sufficient. In some instances time may be required for a strategy to emerge that is consistent with domestic political realities.”
See “Assessments of Member’s Capacity to Service its Financial Obligations to the Fund,” memorandum from the Head of PDR to area departments, July 5, 1990: “[Medium-term] scenarios should prove useful in identifying cases where external viability would be ensured only if strong policy actions continued in the post-program period, only if the external environment remained particularly favorable (say, relative to developments in the past), or if creditors were to provide large amounts of exceptional financing through the projection period. It is important that these scenarios not be presented in a way that fails to warn of potential dangers. . . . When an improved policy stance in the post-program period is mainly an extrapolation of progress expected to be made under the program, or if further major improvements are expected in the post-program period without specific policy underpinnings, these limitations should be made explicit. Similarly, an attempt should be made to assess the quality of the adjustment effort, in particular from the point of view of the sustainability of the program. Temporary or cyclical factors should be identified in evaluating the underlying strength of the adjustment.” The note further stressed “the importance of evaluating the risks of policy slippages and the implications of such slippages for the medium-term prospects of the balance of payments and the member’s capacity to repay the Fund. These evaluations could be based on an assessment of those key policy areas where the authorities have had difficulties with program implementation in the past. Also, if key policy elements have been proposed but have yet to be legislated or implemented, the analysis would assess the implication of a failure to adopt the required measures in a timely fashion. Whenever possible, specific contingency measures should be outlined for cases when legislatures fail to act on key policy elements in the program.”
Mackenzie and others (1997): “One basic lesson for successful growth-oriented fiscal adjustment is that, to the extent possible, it should emphasize administrative reforms right at the outset. [Experience] suggests that early reform is more likely to be substantial reform.”
The following sections draw upon a series of questionnaires to and interviews with the main bilateral donors and creditors, with the Paris Club Secretariat, and with a range of international banks and asset managers engaged in emerging market investments.
The latter point, in some cases, appeared to stem from the perception that the IMF would conduct its assessments more seriously if its resources were involved than otherwise.
In the case of the Paris Club, relying on the IMF has the added advantage of providing an objective framework to determine which countries should benefit from debt restructuring and on what terms. The multilateral and informal nature of the Paris Club also makes its reliance on the judgments of the IMF with regard to debt-service capabilities and the macroeconomic framework natural.
This reluctance was based on varied reasons, including a fear that making IMF assessments too public or explicit might turn the institution into a universal credit rating agency and concerns about marginalizing the catalytic role of the IMF’s resources.
In most cases this was because they viewed agreement on an IMF-supported program as providing some assurance that realistic economic policies and reforms would be implemented, eventually leading to the restoration of external viability and so enhancing the value of the debt to be restructured. In many cases, the direct financing implications of IMF involvement were also considered important.
For example, Pakistan’s 1999 debt rescheduling agreements with private creditors were concluded as the EFF/ESAF arrangement was off-track, with little prospect of renewed IMF support in the near future.
In a few cases, the relationship is even explicitly taken into account in quantitative risk management systems.
One senior Jamaican official said that, in his experience, sending appropriate signals to private financial markets did not require an IMF-supported program, but it did require that the authorities not be “fighting with the IMF.”
Several studies covering the 1970s and 1980 found evidence of a negative relationship between IMF involvement and the sub-sequent supply of new loans. Bird, Hussain, and Joyce (2000) find that the relationship is unstable but, when significant, strongly negative as far as private capital flows are concerned. Conway (2000) and Eichengreen and Mody (2001) find a positive relationship under particular specifications. Evidence is more consistent as regards official capital flows, for which there does appear to be a positive relationship between the presence of an IMF-supported program and those flows. However, the sustainability of aid flows in the post-program period is not established.
To be sure, countries with market access retain, to some extent, the flexibility of dispensing with an IMF lending arrangement if they feel the steps required from them to be granted the IMF’s seal of approval are not in their best interests. The case of Jamaica, discussed in Part II (Chapter 12), is a good illustration of that situation. On the other hand, countries that do not have access to private capital markets, either permanently or temporarily, have strong incentives to accept the IMF’s conditions even when they have misgivings about their feasibility or likely impact.