Prolonged Use of IMF Resources Terms of Reference for an Evaluation by the Independent Evaluation Office (IEO)

This note sets out the main issues to be addressed in the evaluation of prolonged use of IMF resources. Section A provides some background on the issue: why prolonged use might be a problem, including the main criticisms by various external observers; a discussion of the extent of prolonged use; and a summary of the available empirical evidence on several of the issues related to prolonged use. Section B then lays out the "terms of reference" for the project—namely, the broad sets of questions to be addressed by the evaluation, along with a preliminary indication of the proposed methodology to address these questions. A final subsection (B.4) raises some questions about what, if anything, should be done to limit the prevalence of prolonged use. By examining the IMF's role in a group of countries which have had especially intractable adjustment problems, the evaluation should also help to cast light on broader questions of program design and the appropriate balance between lending arrangements and surveillance activities.

A. Background Information

1. Main arguments on the pros and cons of prolonged use of Fund resources (UFR)

Why might prolonged use be a problem? The prolonged use of Fund resources by certain countries has been criticized from a number of perspectives. Without taking any a priori view on their merits, the main criticisms, which should be taken into account in framing the evaluation, are as follows:

• prolonged UFR suggests a lack of effectiveness of IMF-supported programs, as the repeated need to make use of IMF resources indicates a persistence of the balance of payment difficulties which such programs are intended to solve (Meltzer). In recent years, some critics have argued that lack of ownership, leading to poor program implementation, and program design flaws are key factors underlying the lack of effectiveness;
• it may encourage over-indebtedness either because of insufficient attention to debt sustainability issues, which cannot be satisfactorily addressed by a succession of flow treatments, or because a prolonged "IMF seal of approval" encourages overlending. (Bandow, Sachs); in heavily indebted countries it could reflect a strategy of "defensive lending" by the IMF and other multilaterals to avoid default (Birdsall);
• it may hinder institutional development in the borrowing countries, in particular by giving the IMF a lasting influence over their policy-making process, thereby undermining national sovereignty and often compromising the development of responsible, democratic institutions that correct their own mistakes and respond to changes in external conditions. (Meltzer); and
• it is in contradiction with the IMF's mandate as set forth in the Articles of Agreements, which stress that IMF resources should be made available to members "temporarily"¹ to cope with balance of payments disequilibria (Bird, Meltzer). A related concern is that some prolonged users may be substituting IMF financing for other, more expensive, financing sources.

However, it can also be argued that, in some cases, frequent recourse to IMF arrangements may take place for good reasons and be fully compatible with both the IMF's mandate and a broadly defined sense of economic efficiency:

• This could be the case for countries subject to frequent external shocks and for countries where external imbalances have deep-seated structural causes, which cannot be overcome over the short term, or only at a great cost to economic prosperity—an outcome IMF programs are specifically intended to avoid. The latter category could include many low income and transition economies.
• Moreover, since the mid-1980s, most IMF arrangements intended to tackle the latter kind of balance of payments difficulties for low-income countries have been funded from special accounts, (i.e. the SAF, ESAF, and PRGF) and consequently have not exerted pressure on the revolving nature of general resources.² Indeed, it has been argued that the views of the official international community on the appropriate length of Fund financial involvement have changed over time, at least for the low-income group of countries and that the Fund has been given a mandate to provide support, via the PRGF, for programs intended to foster growth and poverty reduction over a longer time horizon.
• In a probabilistic sense, there may always be countries that, ex post, are the victim of a series of unanticipated exogenous shocks. In such an environment, a zero rate of "prolonged use" may not be a desirable objective for the institution.
• Finally, it has also been argued that prolonged use of IMF-supported programs may simply reflect the unwillingness of other lenders (be they private or public) to provide financing without the "seal of approval" they consider an IMF-supported program to be. This raises important questions with regard to the appropriate balance between the IMF's lending and surveillance activities.

back to top

2. Description of the phenomenon of prolonged use of Fund resources (UFR)³

a) Definition of prolonged use

The prolonged use of IMF resources can be characterized in several different ways. The most common concepts found in various IMF policy papers and in the literature are summarized in Annex I. The various definitions focus on one of two underlying concepts: either on the amount of time spent under IMF-supported programs—where the primary emphasis is on the length of coverage of Fund conditionality—or on the length of time IMF resources are outstanding. In practice, there is a high degree of correlation between the coverage of the different concepts.

In the current project, it is proposed to use mainly a definition based on time spent under IMF arrangements. More specifically, we would define as prolonged users countries which have been under (non-precautionary) arrangements⁴ for at least 7 out of any ten years period during 1971-2000. Forty four countries would be covered by this definition, of which 15 countries would have had especially prolonged use (i.e. at least 15 years under IMF arrangements during the 30-year period 1971-2000). However, the evaluation will take account of the distinction between users of general and concessional resources; 29 of the 44 prolonged users (and 13 of the "very prolonged users") are PRGF-eligible. A full list of countries is given in Annex III.

Unlike the definition used in previous IMF reviews, the proposed definition does not impose any threshold on the outstanding use of Fund resources at the end of the period in order not to exclude countries which have completed their lending cycle, albeit a very long one, and in that sense have "graduated" from IMF support. On the other hand, the minimum amount of time under programs (seven years out of any ten-year period) is set higher than in previous reviews of the phenomenon.⁵

In principle, a distinction could be made between "prolonged" use and "repeat" use. Repeat 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 BOP problems and recourse to IMF financing. Prolonged users would encounter few such episodes of IMF "abstenance", 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 among the group of extensive users: all such countries appear to have experienced interludes when their external position improved, followed by renewed difficulties.

More generally, besides the core definition presented above, the evaluation will need to take account of the following:

• prolonged users of general resources
• prolonged users of concessional resources
• countries which graduate from IMF support after a period of prolonged use
• repeat users of precautionary arrangements
• Transition economies (mainly of the former Soviet Union), which became members of the IMF at the beginning of the 1990s, thus making their experience as prolonged users a recent—hence relatively—short one.

b) Prevalence of prolonged use⁶

A preliminary analysis of IMF arrangements databases since 1971 suggests that while the incidence of prolonged use has been affected by developments in the international economic environment, it is neither a recent nor a rare phenomenon regardless of the definition employed—nor is it one that is becoming less important (see Chart 1 in Annex II).

Evidence collected in previous internal reviews indicate that prolonged use started to build up in the second half of the 1970s and accelerated sharply in the first half of the 1980s as a result of the debt crisis so that, by 1986, 25 member countries were prolonged users of the Fund's general resources according to the definition used in the 1991 review (see Annex I). Thereafter, the availability of concessional resources resulted in a large shift of prolonged users from the general to the concessional window, so that by 1990 the number of prolonged users of general resources had fallen dramatically. In reality, however, the overall number of prolonged users had not declined since by 1990 five of them had fallen into arrears and been declared ineligible for further borrowing from the IMF. Prolonged use rose again in the 1990s, partly due to the expansion of the IMF membership to former Soviet Union states.

Countries eligible for concessional facilities now account for by far the largest number of prolonged users, but the smaller group of prolonged users of general resources have a larger total exposure to the IMF (see Charts 3 and 4 in Annex II).

Another indication of the extent of prolonged use is found in Jeanne and Zettelmeyer (2001). They measure the timeframe in which IMF members actually eliminate outstanding obligations to the Fund—after taking account of new lending—which they call a "lending cycle".⁷ They find that, for all developing countries, a surprisingly high 54 percent of the lending cycles initiated since the creation of the IMF were not completed at end-2000, and that the average length of such "incomplete" cycles is 18 years. Not surprisingly, the proportion of uncompleted lending cycles and their length is even higher for HIPC countries. But they also find that 30 percent of emerging market countries⁸ which initiated their lending cycle between 1947 and 1991 have yet to complete them, and, for these countries, the average length of time with outstanding obligations to the Fund is about 21 years, compared with an average lending cycle of 8 years for other members of this group (see Table 2 of Annex II for details).

back to top

3) Insights from the existing empirical literature on prolonged users and the causes of program interruptions

Previous empirical work on issues involving prolonged users and the causes of program interruptions have employed a range of econometric and other techniques, of varying robustness, and have covered different time periods. However, they suggest the following broad messages:

• Countries with lower per capita incomes are more likely to be recurrent users of IMF resources, although this probably largely reflects the existence of the concessional facilities (at least since they were introduced in 1986). Countries that are more vulnerable because of their dependence on primary exports or their landlocked nature also tend to spend longer periods in Fund-supported programs (Bird et. al. (2000), and Joyce (2001)).
• Not surprisingly, frequent users typically have lower (gross) reserve levels, larger external current account deficits and higher debt service ratios than infrequent users. However, although there is some (weak) evidence that frequent users have larger fiscal deficits, this does not appear to reflect a bias toward "big" government: frequent users tend to have smaller government sectors, but they also score less well on various governance rankings. (Bird et. al. (2000)).
• The choice of exchange rate regime does not by itself appear to affect the frequency of borrowing from the IMF (Bird et.al. (2000)).
• Empirical studies of the effectiveness of IMF-supported programs in general and the IMF's own conditionality reviews have tended to find that programs as a whole achieve their external objectives (e.g., in the sense of improving the external reserves position), while the results on growth and inflation are more mixed.
• Preliminary work by Conway (2000) also suggests that the continuing reliance on IMF-supported programs lessens the "turnaround" effect on the external position—and that the cumulative effect of long-term participation in IMF-supported programs may be associated with an increased probability of entering into a new external crisis (defined in terms of a renewed deterioration in reserves or pick-up in inflation).
• In comparing program projections and outcomes, Musso and Phillips (2001) conclude that there is no systematic bias toward optimism or pessimism of projections for growth or capital flows for the programs in their sample (which does not include programs supported under the concessional facilities). Follow-on programs, however, are associated with a significantly lesser degree of overoptimism on the prospects for inflation than the general run of IMF-supported programs. They also find that there appears to be a bias toward optimism in projections of growth and the capital account in the "big" IMF-supported programs (in terms of size of access to IMF resources)—probably reflecting the greater risks of falling into "bad" multiple equilibrium in such cases.
• In terms of influences on program implementation, Ivanova et. al. (2001), whose work, however, is not focused on prolonged users per se, find no significant difference between the effort the IMF invests (i.e. the magnitude of staff resources or intensity of missions) in implemented and non-implemented programs. However, several political economy variables, including the level of political stability and the strength of special interests in parliaments, do appear to have a significant influence on program implementation.
• In a detailed examination of the causes of program interruption among SAF and ESAF cases, Mecagni (1999) concludes: (i) in a majority of cases, the main cause of the interruption was fiscal in nature; (ii) however, interrupted programs were not more "ambitious" in their targeted fiscal adjustment than other programs; (iii) there was no obvious correlation between the number of structural performance criteria/benchmarks and the rate of observance of such conditionality; and (iv) most program interruptions appear to have had their origins in a variety of political factors.

back to top

B. Terms of Reference: Key Questions to be Addressed by the Evaluation and Proposed Methodologies

This section sets out the main questions that the evaluation project proposes to address along with a preliminary indication (in italics) of the methodological approach that is proposed for each question. For a number of issues, the most suitable methodology is likely to evolve as the evaluation proceeds, but the following broad approaches are envisaged, depending on the particular question posed. External researchers who have done work on these issues are also invited to submit their findings.

• A major part of the evaluation will consist of detailed case studies of three country cases (Pakistan, Philippines, and Senegal) that have been among the most prolonged users and also represent a mix of users under the general resources and the concessional facilities. Case studies are needed to address in greater depth many of the more qualitative issues. The purpose of the studies is to examine the role of the IMF, not to undertake a comprehensive assessment of a particular country's development strategy. They will involve detailed reviews of Fund documents, interviews with relevant staff, management and Executive Directors and country visits to seek the views of the authorities and a wide range of other stakeholders, including representatives of the private sector, labor groups, civil society, and academics. Narrower desk reviews of specific issues in other country cases may also be undertaken—e.g., countries that appear to have "graduated" from IMF support.
• Empirical analysis of a wider range of prolonged users, taken both as a whole and in various relevant groupings.
• · Econometric analysis on a large sample of prolonged and temporary users. However, given resource and time constraints, it is not proposed to launch major new research or data collection initiatives.⁹ Such analysis would largely draw on and, where necessary, adapt the work of Barro and Lee (2001), Bird et al (2000), Conway (2000), Joyce (2001), and Musso and Phillips (2001).

It should be recognized at the outset that much of the evidence, especially on the broader questions, may not be amenable to rigorous hypothesis-testing. One central methodological problem is that the decision to enter into an IMF-supported program, or series of programs, is endogenous. This affects both statistical comparisons across groups of countries and the results from case studies (where it could be argued that any sample of cases drawn from the group of prolonged users may be biased against the effectiveness of IMF-supported programs). In many cases, therefore, the evaluation will need to rely upon more qualitative judgments.

back to top

1) What does repeat use of IMF resources imply for the effectiveness of IMF-supported programs?

a) What are the characteristics of the prolonged users?

Bearing in mind the distinction between repeat users of general (GRA) and concessional resources, the following sets of questions will be investigated:

• Which countries get prolonged access? How do they differ from temporary/infrequent users of IMF resources? Are there common economic characteristics among these countries? Do internal political characteristics (or external political connections) appear to have any systematic effect on such access?

[Empirical analysis of the set of prolonged users, perhaps supplemented by econometric analysis on a large sample of prolonged and temporary users drawing primarily on existing and ongoing work (e.g. by Bird, Conway, and Barro and Lee)].

• Do there appear to be "healthy" cases of prolonged use, eventually leading to graduation from UFR (i.e. where, after a substantial period of IMF support to cope with deep-seated, structurally-related balance of payments difficulties, countries have "graduated" from IMF support)? What are the characteristics of a sustained "exit" from prolonged use?

[Desk review of selected "graduators" from prolonged use]

b) Why does it take so long to achieve balance of payments sustainability in the prolonged users?

• Country related factors
  
  
  • Was it weak policy implementation, whether or not it led to program interruption? If so, what was the role of IMF conditionality?
  • Were there factors beyond the reach of IMF-supported programs at play, such as policy reversals or lack of sustained implementation beyond the term of the program?
  
  [Case studies supplemented where possible by an empirical analysis of a broader sample. For example, the latter would include an analysis of the use of waivers: are prolonged users also repeat requesters of waivers?]
  
  
  • Why do some countries consistently fail to comply with programs that are repeatedly negotiated? Can such a pattern be identified ex-ante, e.g. through political science analysis, as suggested by the work of Dollar and Svensson (2000) and Ivanova et al. (2001) on the causes of program failures?
    
    

  [Case studies, including political economy analysis of one or two program events drawn from one of the country cases—to assess how the tools used by political scientists might add to the IMF's ability to assess political feasibility and risks.]

• Program-design related factors
  

• Was the programs' design adequate: to what extent was the need for one or more successor programs anticipated and recognized ex-ante? Were there frequent or systematic over-ambitious targets and/or over-optimistic assumptions about GDP and export growth? Did the design of programs reflect learning from previous programs?
• How effective was conditionality in coping with "intractable" problems that continued from program to program? Was there an adequate ex ante assessment of a country's implementation capacity?
• Did adverse external shocks play a part? Could they have been better dealt with through more stringent sensitivity analysis and more concrete contingency planning within the program design?

[Case studies. To address the issue of how well prolonged program use has dealt with deep seated structural problems, the longer-term track record of programs in coping with one or more structural issues that are clearly in the IMF's area of primary responsibility will be investigated.]

• Institutional factors

  • To what extent did prolonged UFR reflect refinancing of Fund credit, as part of a pattern of "defensive lending"?¹⁰

  
  [Analysis of the pattern of IMF lending and outstanding credit in prolonged and temporary users. Review of internal briefing papers for prolonged users.]

  • To what extent did prolonged use reflect the need for a framework to facilitate the rollover or restructuring of other external debt or to ensure the involvement of other sources of financing (in the private or official sector)? In particular, has the need for Fund arrangements in the context of Paris Club rescheduling led to a push for programs for which there would otherwise have been little justification?
    
    

  [Examination of the links between prolonged Fund programs and Paris Club reschedulings; interviews with donors and financial market participants]¹¹

back to top

2) What are the implications of prolonged use for the borrower and the IMF?

• Do "repeat use" programs have a different impact than other Fund programs on growth, etc.?
  
  [Econometric analysis of impact on growth, and the BoP. While not seeking to undertake any major new econometric research in this area, the evaluation project would ascertain the effects of including variables measuring the frequency of previous UFR in recent research work.]
  
  
• Has repeat UFR encouraged over-indebtedness for lack of sufficient attention to overall debt sustainability issues? To what extent did repeat use reflect a lack of an adequate mechanism to deal with an existing but non-recognized debt sustainability issue?
  
  [Case studies. While it is beyond the scope of the evaluation to test, in a rigorous quantitative manner, for any significant link between over-indebtedness and prolonged use, an assessment of how medium-term debt sustainability was analyzed in Fund documents will be undertaken to see if there was a tendency toward over-optimistic projections of a return to external viability.]
  
  
• Given the scope of the prolonged use phenomenon, is it a significant constraint on the revolving nature of IMF financial resources?

  [Analysis of the share of Fund financial resources utilized by prolonged users, based on the IMF's financial records]

back to top

3) Is internal governance adequately structured to minimize ineffective repeat use of IMF resources?

• How have the IMF Executive Board and management addressed the issue of prolonged UFR in the past? With what conclusions and operational implications?
  
  [Review of IMF policy papers, internal operational guidelines, and record of Board discussions]
  
  
• Ex-ante, was there a systematic assessment of the likelihood of success of the programs involving repeat users, including political economy considerations?
  
  How did staff, management (and ultimately the Executive Board) respond when faced with requests for further support from a prolonged user? Did the nature of conditionality and program design change and with what effect?
  
  Was there a candid assessment in briefing papers and staff reports of the problem of prolonged use, including of the timeframe needed to attain balance of payments sustainability, or was there a tendency to project an overoptimistic adjustment path to preserve the appearance of a temporary involvement by the IMF? When the timeframe was known to exceed the duration of the program, what conclusions were drawn?
  
  What consideration has been given to the preparation of an "exit" strategy for prolonged use, including those under the PRGF?
  
  What information was given to the Executive Board about the past performance of prolonged users and about the likelihood of success of the new program in reports on new requests for UFR? What did the Board say about those cases when approving the requests (in particular about the realism of program targets and assumptions and the path toward external sustainability)?
  
  [Case studies and desk review of a broader set of prolonged users]
  
  
• Ex- post, was there any internal assessment of why repeat users' programs failed (i.e. either went off-track or failed to achieve balance of payments sustainability). What could be done to strengthen such ex-post assessments?
  
  [Case studies]
  
  
• To what extent does prolonged use signal deficiencies in the surveillance process? For example, when Article IV surveillance and program negotiations are closely linked in time, are there any indications that surveillance assessments were insufficiently frank, perhaps to avoid undermining the credibility of agreed upon programs?
  
  [Case studies]
  

back to top

4) What lessons and possible changes are suggested by the experiences of the prolonged users?

While it is premature to predict the lessons that will be derived from the evaluation, the following broad areas will be considered.

a) In the design of programs

• Is the program design process sufficiently realistic and candid on the risks and tradeoffs involved and with regard to the time-horizon contemplated for the restoration of balance of payments sustainability?
• With an emphasis on the role of conditionality and ownership:¹² Is greater selectivity (in terms of number of programs and types of conditions) needed? Should the IMF dedicate more resources to the analysis of the political economy context of its programs and to the potential adverse side-effects of the policies it advocates, in particular distributional effects?
  

b) In the internal IMF governance process and the respective roles of surveillance and lending arrangements

• What can be done to strengthen the role of surveillance in cases where the IMF has a prolonged involvement via programs?
• What changes in internal governance processes would be helpful—both ex ante (e.g., in the information provided to the Board and in the analysis of the economic and political risks to achieving a program's objectives) and ex post (e.g., through a more systematic internal evaluation process).

c) In the design and use of various facilities and relations with other donors

The 2000 review of facilities has led to the adoption of several measures which should contribute to reducing the extent of prolonged use, in particular the creation of the contingent credit line (CCL), the strengthening of post-program monitoring provisions, the introduction of repurchase expectations and the imposition of a surcharge on large outstanding obligations. However, a number of additional steps could be considered:

• Should the IMF adopt stricter rules regarding prolonged users' access to its resources or various incentives to discourage prolonged UFR? More generally, should the IMF be more selective in providing financing, or in interrupting this financing, in cases which a program is not achieving its objectives?
• Should there be a more explicit "exit strategy" for prolonged users of facilities involving concessional resources? What should be the IMF's long-term role in such cases?
• What should be the IMF's role in cases where there is no pressing balance of payments need but there is a demand (either from the authorities, aid donors or private financial markets) for the overall monitoring role of an IMF-supported program?
• Should precautionary arrangements be used more extensively? Is a prolonged use of precautionary arrangements desirable?

back to top

Reading List

Alesina, Alberto, and David Dollar, (March 2000), "Who Gives Foreign Aid to Whom and Why?". Journal of Economic Growth, 5: 33-63.

Bandow, Doug, (1994) "The IMF: A Record of Addition and Failure" in Perpetuating Poverty: The IMF, the World Bank and Developing Countries, Cato Institute.

Bandow, Doug, and Ian Vásquez, (1994) "The Dismal Legacy and False Promise of Multilateral Aid" in Perpetuating Poverty: The IMF, the World Bank and Developing Countries, Cato Institute.

Barro, Robert J., and Jong-Wha Lee, (November 2001), "IMF Programs: Who is Chosen and What are the Effects".

Bird, Graham, Mumtaz Hussain and Joseph P. Joyce, (August 2000), "Many Happy Returns? Recidivism and the IMF". Department of Economics, Wellesley College Working Paper 2000-04.

Bird, Graham and Dane Rowlands, (2001), "Catalysis or Direct Borrowing: The Role of the IMF in Mobilizing Private Capital". World Economy, Jan 2001.

Bird, Graham and Dane Rowlands, (2000), "The Catalyzing Role of Policy-Based Lending by the IMF and the World Bank: Fact or Fiction". Journal of International Development 12, 951-973.

Bird, Graham and Dane Rowlands, (2001), "IMF Lending: How is it Affected by Economic, Political and Institutional Factors". Policy Reform, Vol.4, 2001.

Birdsall, Nancy, Stijn Claessens, and Ishac Diwan, (2001), "Will HIPC Matter?: The Debt Game and Donor Behavior in Africa". Paper presented at the WIDER Debt Relief Conference, August 2001, Helsinki.

Bordo, Michael D, and Harold James, (June 2000), "The International Monetary Fund: Its Present Role in Historical Perspective". NBER Working Paper 7724.

Brooks, Ray et. al. (1998), "External Debt History of Ten Low-Income Developing Countries". IMF Working Paper 98/72.

Conway, Patrick, (1994), "IMF lending programs: Participation and impact". Journal of Development Economics, Vol. 45, 365-391.

Conway, Patrick, (January 2000), "IMF Programs and Economic Crisis: An Empirical Study of Transition". University of North Carolina, Department of Economics.

Devarajan, Shantayanan, David Dollar, and Torgny Holmgren, (December 1999), "Aid and Reform in Africa". The World Bank.

Dicks-Mireaux, Louis, Mauro Mecagni, and Susan Schadler, (2000), "Evaluating the effect of IMF lending to low-income countries". Journal of Development Economics, Vol. 61, pp. 495?526.

Dixit, Avinash, (2000), "IMF Programs as Incentive Mechanism". IMF Research Dept., mimeo.

Dollar, David, and Jakob Svensson, (October 2000), "What Explains the Success or Failure of Structural Adjustment Programmes?" The Economic Journal, 110, 894-917.

Haque, Nadeem Ul, and Mohsin S. Khan, (December 1998), "Do IMF-Supported Programs Work? A Survey of the Cross-Country Empirical Evidence". IMF Working Paper.

Ivanova, Anna, Wolfgang Mayer, Alex Mourmouras, and George Anayiotos, (November 2001), "What Determines the Success or Failure of Fund-Supported Programs?" IMF Working Paper.

Jeanne, Olivier and Jeromin Zettelmeyer (October 2001). "International Bailouts: The IMF's Role". Economic Policy.

Joyce, Joseph P., (March 2001), "Time Present and Time Past: A Duration Analysis of IMF Program Spells". Federal Reserve Bank of Boston Working Paper No. 01-2.

Joyce, Joseph P., (1992), "The Economic Characteristics of IMF Program Countries". Economic Letters, 38 (237-242)

Krueger, Anne O., (1998), "Whither the World Bank and the IMF?" Journal of Economic Literature, Vol.XXXVI, pp. 1983-2020.

Marchesi, Silvia, (2000), "Adoption of an IMF Programme and Debt Rescheduling: An Empirical Analysis". Centre for the Study of Globalization and Regionalization Working Paper No. 56/100.

Mecagni, Mauro, (1999) "The Causes of Program Interruptions", in Economic Adjustment and Reform in Low-Income Countries, studies by the Staff of the IMF, edited by H. Bredenkamp and S. Schadler.

Meltzer, Alan et. al. (2000). Report of the International Financial Institutions Advisory Commission (available at http://www.house.gov/jec/imf/ifiac.htm).

Mussa, Michael, and Miguel Savastano, (1999), "The IMF Approach to Economic Stabilization". IMF Working Paper.

Musso, Alberto, and Philipps, Steven (2001), "Comparing Projections and Outcomes in IMF Supported Programs". IMF working paper.

Przeworski, A and J.R. Vreeland (2000), "The Effects of IMF Programs on Economic Growth". Journal of Development Economics, Vol 62, No.2, pp385-421.

Tammen, Melanie S., (1994),"Fostering Aid Addiction in Eastern Europe" in Perpetuating Poverty: The IMF, the World Bank and Developing Countries, Cato Institute.

Thacker, F., (2000), "The High Politics of IMF Lending". World Politics, 2000.

Tsikata, Tsidi, (1999), "Progress Toward External Viability" in Economic Adjustment and Reform in Low-Income Countries, studies by the Staff of the IMF, edited by H. Bredenkamp and S. Schadler.

Willett, Thomas D., (December 1999), "The Need for a Political Economy Capability at the IMF". Claremont Colleges.

Willett, Thomas D., (October 2000), "A Soft-Core Public Choice Analysis of the International Monetary Fund". Claremont Colleges.

back to top

Annex 1: Possible Definitions of Prolonged Use

This annex describes various approaches to defining prolonged use that have been used in the IMF or elsewhere. The actual definition resulting from each approach can be made more or less restrictive by varying the threshold which separates prolonged users from "temporary" users of Fund resources.

(i) prolonged effective use of the IMF general resources¹³ (IMF, 2000)

This concept 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 excludes 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 1984, 1991)

This concept encompasses both programs funded from the general resources account and from concessional trusts. It also includes programs which are only partially drawn upon. It may or may not include precautionary arrangements. It does not include drawings on Fund resources not backed by programs (such as first credit tranche purchases).¹⁴

A slightly different version of this concept is used by Bird et al to characterize frequent users of IMF resources. His definition is based on the number of programs signed 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 multi-year timeframe, such a definition does not measure the time spent under IMF arrangements.

(iii) prolonged indebtedness to the IMF (IMF, 1986; Meltzer; Jeanne and Zettelmeyer)

This concept focuses on the length of periods of indebtedness to the IMF, regardless of the origin of the outstanding obligations.¹⁵ However, because IMF facilities have repayment periods varying from 2½ to 10 years, this definition does not distinguish between countries which had only a few programs with relatively long repayment periods and those which had a large number of programs 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.

As noted in the main text, the current evaluation project proposes to use a definition based on the amount of time spent under IMF-supported programs, whether or not a country was eligible to draw.

back to top

Annex 2: An Overview of Prolonged Use of IMF Resources During 1971-2000

This annex provides additional background information on the extent of prolonged use of IMF resources.

Chart 1 shows the evolution of prolonged use over the period according to various alternative definitions.¹⁶ Regardless of definition, prolonged use became important in the 1980s following the debt crisis and has remained significant since then. One, but not the only, factor of importance in the continued phenomenon was the protracted balance of payments problems faced by transition countries. Moreover, the introduction of concessional facilities in the second half of the 1980s has tended to amplify the extent of prolonged use rather than just change its nature.

Chart 2 summarizes in graphic form the history of lending arrangements of 50 countries which at some time during the last three decades made an intensive use of Fund programs. The vast majority of these countries were still under program at the end of the period. Of the 13 countries which were not, three are ineligible to use Fund resources and one no longer exists (the Socialist Federal Republic of Yugoslavia-SFRY).¹⁷

Charts 3 and 4 show the evolution of prolonged use by country and by outstanding exposure to the IMF, respectively, according to the definition proposed in the main text.

Table 1 provides additional information on this group of countries. It shows that nearly 40 percent of these countries made an intensive use of the Fund's general resources, in the sense that they had outstanding obligations to the general resources account (GRA) of over 100 percent of their quota for ten years or more. When both general and concessional resources are taken into account, more than half the sample had average outstanding liabilities over the thirty year period in excess of 100 percent of their quota. In other words, these are countries which made both long and large use of Fund resources.

Table 2 provides information on the average length of "completed" and "incomplete" cycles of debt to the IMF (i.e., the average number of years after a country initiates borrowing from the IMF before its obligations are again reduced to zero. It is not perhaps surprising that most HIPC and many PRGF-eligible countries have had obligations outstanding to the IMF for lengthy periods, given the longer repayment periods of the facilities involved and the expectation that several multi-year arrangements might be required to help a country overcome its intractable external adjustment problems, even in relatively successful cases. (For example, two consecutive ESAF/PRGF arrangements followed by a 10-year repayment period would generate a "lending cycle" of 16 years.) What is more surprising is that a number of emerging market countries have also effectively borrowed from the Fund for very long periods. The 15 "very prolonged users" shown in Annex III (which includes both PRGF-eligible and some emerging market countries), have had outstanding obligations to the IMF for an average of 27½ years.

back to top

Charts and Tables

Chart 1

Repeat users defined as any of the three definitions below:

* more than 1 arrangement every three years on average over the period (frequent users) * more than half of the period covered by IMF programs (prolonged users) * more than 50% of quota in outstanding obligations for over 2/3 of the period (prolonged debtors)

Severe Repeat users defined as any of the three definitions below:

* more than 1 arrangement every 2 years on average over the period (frequent users) * more than 2/3 of the period covered by IMF programs (prolonged users) * more than 2/3 of quota in outstanding obligations for over 3/4 of the period (prolonged debtors)

Chart 2: Fund Arrangements During 1976-2000 (pdf 441Kb)

Chart 3: Prolonged Use By Number of Countries

Chart 4: Prolonged Use by Outstanding Exposure to the IMF1,2

1/ Outstanding obligations are shown according to country PRGF eligibility not according to nature of resources at stake. Therefore, all of blend countries' obligations appear under PRGF-eligible, even though some of them relate to the GRA. 2/ Prolonged users outstanding obligations in each year are taken into account only if they qualify as prolonged user in that year.

Table 1: Intensity of Use of IMF Programs, 1971-2000 (pdf 12Kb) Table 2: Completed and Incomplete Debt Cycles for Borrowers of the IMF, 1947-2000 (pdf 5Kb)

List of Prolonged Users Over 1971-2000

This table gives the full list of "very prolonged users" and "prolonged users" of IMF resources according to the main definition proposed in the text.

Very prolonged users (15 or more years under program)* Other prolonged users (7 or more years under program over any 10 year period)* Bolivia Coted'Ivoire Guinea Guyana Jamaica Kenya Madagascar Malawi Mali Mauritania Pakistan Philippines Senegal Togo Uganda (15) Argentina Bangladesh Benin Bulgaria Burkina Faso Congo D.R. Ecuador Equatorial Guinea Gabon Gambia Ghana Haiti Honduras Jordan Kyrgyz Republic Mexico Mongolia Morocco Mozambique Nicaragua Niger Panama Peru Romania Somalia Turkey Tanzania Yugoslavia Zambia (29)

* Excludes precautionary arrangements. If they were included, the following countries would also appear on the list: Costa Rica, Egypt, El Salvador, Korea, Latvia, Liberia and Uruguay. In addition, Argentina, Haiti and Panama would be classified as very prolonged users.

back to top

Footnotes

¹ Such temporariness is dictated by the need to ensure the revolving nature of IMF resources.

² To the extent that the PRGF Trust is expected to be self-sustaining at some point, concessional resources would also need to be revolving.

³ For the purposes of this paper, the terms "prolonged" and "repeat" use will be used interchangeably.

⁴ Programs supported by precautionary arrangements are the same as other Fund-supported arrangements except that the authorities have indicated that they do not intend to draw on the arrangement. To the extent that they do not reflect any actual balance of payments need nor involve any actual use of Fund resources, it can be argued that precautionary arrangements should be excluded from the core definition. However, the commitment to treat an arrangement as precautionary is never binding for the member country, should its circumstances change.

⁵ For example, this threshold was chosen so that any country with just two three-year arrangements (EFF or ESAF/PRGF) in a decade would not be classified as a prolonged user. In fact, lowering or raising the minimum threshold by one year would not alter substantially the size of the group of "prolonged users."

⁶ The findings reported in this section are illustrated by a series of charts presented in Annex II.

⁷ In the absence of subsequent programs, the length of each lending cycle should be equal to the sum of the program and the repayment period.

⁸ Defined by countries whose sovereign bonds are tracked in the J.P. Morgan Emerging Market Global Bond Index, which is an indication that they would normally be expected to have access to private market financing.

⁹ Whenever possible, the evaluation will rely on existing IMF databases, including the internal database on program objectives, outcomes and conditionality (referred to as the MONA — monitoring of arrangements — database).

¹⁰ Econometric work on countries in Africa by Birdsall et al (2001) suggests that net disbursements by multilateral institutions, including the IMF, are less influenced by the strength of economic policies when total and multilateral debt is high than when it is low. They argue that this shows that "defense lending" (i.e., to avoid default) tends to crowd out selectivity in favor of countries implementing good policies in such situations.

¹¹ A related issue is the catalytic effect of IMF-supported programs and whether prolonged use undermines this catalytic effect. However, this is primarily a topic for further research and not within the comparative advantage of the evaluation office. Bird and Rowlands (2000) conclude that there is little empirical support for a systematic catalytic effect on private flows. They conclude that there is some support for a catalytic effect on official flows to low-income countries.

¹² Dollar and Svensson, in an analysis of World Bank conditionality, contend that donor-effort variables, such as the number of conditions attached to a loan, the time devoted by staff to preparation and supervision etc., have little or no impact on the probability of success or failure of the adjustment program supported by the loan and that the main determinant of success/failure is the political context. Empirical work undertaken by Fund staff also indicates that the number of prior actions set in programs have limited value in establishing a basis for successful program implementation.

¹³ i.e. purchases from the general resources account (GRA), which are typically associated with a stand-by arrangement (SBA) or an Extended Fund Arrangement (EFF). The specific operational definition used in the 2000 review of prolonged UFR characterized as prolonged users countries with an outstanding use of Fund credit over 100 percent of quota and either nine years or more of effective UFR in the previous 30 years, or five years of effective use in the previous 15 years.

¹⁴ The thresholds used in internal IMF definitions have varied overtime: 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 ten years. In all cases, an additional criterion was an outstanding Fund credit of over 100 percent of quota at the end of the period.

¹⁵ 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% of normal maximum for six years or more" in the previous 10 years.

¹⁶ Reflecting available data at this stage of the project, the time series shown in the chart reflects a slightly different definition than that which will be used in the main evaluation. However, as noted in the main text, the scope of prolonged use is not very sensitive to the precise definitions used.

¹⁷ The SFRY was succeeded by the Federal Republic of Yugoslavia in 2000. Use of Fund resources by countries born out of the break up of the SFRY (e.g., Slovenia and Bosnia-Herzegovina) has not been treated s further UFR by the SFRY.