Reform of the Policy on Public Debt Limits in Fund-Supported Programs
Author: International Monetary Fund
Publication Date:
14 Nov 2014
eISBN:
9781498342568
ISBN:
9781498342568
Language:
English
Keywords:
PP; debt limit; debt; present value; financing; loan; conditionality framework; debt risk rating; debt limits policy; financing assumption
Download PDF (1.5 MB)

The reform of the Fund’s policy on the use of conditionality on public external debt in Fund-supported programs (the “debt limits policy”) has been under discussion since March 2013. The discussion has taken place against a backdrop where lower income countries are seeking to boost growth through higher public investment levels, targeted in particular at large infrastructure gaps, while facing both a wider range of external financing opportunities and limits on the supply of traditional concessional financing. The reform of the Fund’s policy on debt conditionality in 2009 was a first step to accommodate these new realities: experience with the 2009 reforms has pointed to the need for more fundamental reforms to provide countries with greater flexibility to finance productive investments while containing risks to medium-term debt sustainability. The reforms proposed here build on the Board review of the debt limits policy in March 2013, ensuing informal Board discussions in January and May 2014, discussions at an informal seminar in September 2014, and various stakeholder consultations. In developing this reform proposal, staff has sought to first specify a robust set of principles to guide the use of public debt conditionality in all Fund arrangements and then examine how these principles should apply in the specific circumstances of countries that normally rely on official external concessional financing.

Abstract

The reform of the Fund’s policy on the use of conditionality on public external debt in Fund-supported programs (the “debt limits policy”) has been under discussion since March 2013. The discussion has taken place against a backdrop where lower income countries are seeking to boost growth through higher public investment levels, targeted in particular at large infrastructure gaps, while facing both a wider range of external financing opportunities and limits on the supply of traditional concessional financing. The reform of the Fund’s policy on debt conditionality in 2009 was a first step to accommodate these new realities: experience with the 2009 reforms has pointed to the need for more fundamental reforms to provide countries with greater flexibility to finance productive investments while containing risks to medium-term debt sustainability. The reforms proposed here build on the Board review of the debt limits policy in March 2013, ensuing informal Board discussions in January and May 2014, discussions at an informal seminar in September 2014, and various stakeholder consultations. In developing this reform proposal, staff has sought to first specify a robust set of principles to guide the use of public debt conditionality in all Fund arrangements and then examine how these principles should apply in the specific circumstances of countries that normally rely on official external concessional financing.

I. Introduction

1. This paper proposes reforms to the Fund’s debt limits policy (DLP)—the policy that currently governs the specification of conditionality on external public debt accumulation in Fund-supported programs.2 The case for reform was discussed in a formal Board meeting on March 22, 2013, where it was agreed that the key area in which reforms were needed related to the policies applied in the case of countries to whom concessional financing would normally be available. Specific reform proposals were discussed at that Board meeting, in ensuing informal Board sessions in January and May 2014, and at an informal seminar in September 2014.3

2. The reform proposed here builds on these earlier discussions. To provide a strong conceptual underpinning for the reform proposal, the paper first develops a set of general principles that would guide the use of public debt conditionality in Fund-supported programs, including identification of the circumstances under which it would be appropriate to specify separate limits on external and domestic debt accumulation. It then explains how these principles would be applied in countries that normally rely on official sector external financing provided on concessional terms.4 Reflecting this broader approach of focusing on public debt, rather than public external debt, the paper proposes adoption of a new set of “Guidelines on the Use of Public Debt Conditionality in Fund Arrangements” to replace the current “Guidelines on Performance Criteria with respect to External Debt in Fund Arrangements”.

3. The paper is organized as follows. Section II discusses the case and context for the reform proposal. Section III outlines a common architecture that would guide the use and design of debt limits under Fund-supported programs. Section IV examines how these principles would be interpreted in the case of countries that normally rely on official external financing on concessional terms to meet their external financing needs. Drawing on the preceding sections, Section V identifies key elements of the proposed new guidelines on debt conditionality; Section VI proposes transitional arrangements, and Section VII contains issues for discussion. A proposed Board decision with the text of the guidelines will be circulated separately.

II. Background and Rationale for Reform

4. The debt limits policy was last modified in 2009, with the changes introduced relating primarily to the policy as it applied to countries for whom concessional financing is normally available. The 2009 reforms sought to take account of improved macroeconomic management and performance in many of these countries and the widening array of financing options available to borrowers, the latter attributable both to the expanded activities of nontraditional official creditors and the improved credit-worthiness of borrowers, facilitated by international debt relief initiatives. The new policy provided countries with enhanced flexibility to tap non-concessional resources, within a well-specified framework where flexibility was linked to the significance of debt vulnerabilities and to a unified assessment of debt management and public investment management capacity.

5. The experience with the 2009 policy reform has been uneven: a systematic process for handling non-concessional borrowing was put in place for countries for whom concessional financing is normally available, but there were significant implementation challenges.5 The policy had the unintended effect of focusing policy discussions and program dialogue on individual projects and/or one segment of external borrowing, distracting attention from the broader fiscal framework and the financing strategy. The sharp dichotomy between concessional and semi-concessional loans is difficult to justify on economic grounds, not least because it failed to recognize that the two forms of financing were not equally available for all forms of investment projects. Finally, the systematic exclusion of concessional financing from debt limits left overall borrowing unconstrained in the majority of Fund-supported programs.

6. Debt conditionality has played different roles, and taken different forms, in GRA and PRGT-supported programs.6 In GRA programs, debt accumulation has typically been contained through a performance criterion on the fiscal balance, with debt limits, where used, deployed to allow for wider coverage of debt creating activities than is captured in the fiscal accounts; in the majority of these programs, debt limits have covered total public debt rather than external public debt alone.7 By contrast, debt limits in PGRT-supported programs have been near-universal and have focused on controlling the composition of external borrowing (i.e., the breakdown between CB and NCB) rather than the aggregate level of external borrowing.8

7. The reforms proposed here are being developed at a time where economic conditions in the majority of countries traditionally reliant on external concessional financing has further improved and where many of these countries now have some access to market funding:

  • For the preponderance of these countries, economic performance has been strong over the past decade; debt burdens are now much lower, with almost all countries eligible for HIPC/MDRI debt relief having received comprehensive external debt relief; and debt risk ratings in the majority of countries have either improved or remained stable since 2007 (see Annex II). That said, a number of countries have recorded significant increases in public debt and/or public external debt levels in recent years.

  • Given infrastructure and other obstacles to growth, the need for financing for public investment remains pressing, while the supply of concessional financing (including grants) is typically insufficient to meet these needs—and often comes with a strong sectoral bias (e.g., towards social sector projects).

  • The supply of project finance available to many lower income countries on semi-concessional or market terms has increased substantially over time, aided by expanded lending by emerging creditors and the countries’ own improved credit-worthiness. Separately, helped both by global financial market conditions and domestic credit-worthiness, many countries formerly excluded from tapping international credit markets have, in recent years, succeeded in launching one or more sovereign bond issues.

8. The reform proposed here seeks to address shortcomings of the existing debt limits policy, while taking account of ongoing changes in the external financing landscape that countries traditionally reliant on concessional financing now face. The overarching objective is to provide countries with greater flexibility to finance productive investments while containing risks to medium-term debt sustainability. Given the merits of grounding the treatment of countries traditionally reliant on concessional financing within a broader framework governing debt conditionality in all Fund-supported programs, staff has sought to first specify a robust set of principles to guide the use of public debt conditionality in all Fund arrangements and then examine how these principles should apply in the specific circumstances of countries that normally rely on official external concessional financing. In developing the proposed reform, the paper takes into account key concerns expressed by Executive Directors and other stakeholders in a series of discussions and consultations over the past eighteen months. These included: (i) ensuring even-handedness in the application of the policy; (ii) introducing unified debt limits, covering both concessional and non-concessional external borrowing, where relevant; and (iii) preserving incentives to encourage creditors to provide, and borrowers to seek, financing on concessional terms.

9. Reflecting increased financial integration, it is proposed to broaden the focus of the policy to encompass total public debt. This is already the de facto practice in most programs supported from the General Resources Account (GRA), where debt limits are typically set on total public debt, rather than on public external debt. However, as argued below, for countries where there is significant segmentation between domestic and external sources of financing, there is an economically sound case for continuing to specify debt limits on public external debt. That said, the appropriate form of any debt limit will depend on specific country circumstances.

10. The current policy contains specific features that apply to “countries to whom concessional financing is normally available”. To enhance precision, it is proposed to replace this usage with the concept of “countries that normally rely on concessional (external) financing”, used as a shorthand for “countries that normally rely on official external financing provided on concessional or near concessional terms.” The original concept lacks precision, in the sense of failing to distinguish between countries that have some, but relatively, limited access to concessional financing and countries that obtain the bulk of their external financing on concessional terms. We opt for the concept of “normally rely on” to focus attention on countries for whom concessional financing is a key source of public external financing, employing the term “normally” to reflect the fact that, from time to time, large non-concessional project loans or a sizeable sovereign bond issue could account for the bulk of public external financing in any one year.

III. The Use of Public Debt Conditionality in Fund-Supported Programs: General Principles

11. This section describes the broad principles that would guide the use of debt limits in all Fund-supported programs.9 As specified here, the use and form of debt limits would depend on: the macroeconomic circumstances of member countries, including the extent and type of debt vulnerabilities; the composition of public sector financing; the quality and coverage of fiscal data produced by national accounting and budget processes; and the objectives of the program.

A. When is the use of debt conditionality justified?

12. Public borrowing plans are one component of the fiscal program being supported under a Fund arrangement. The use of limits on debt accumulation is one potential tool to be deployed in designing fiscal conditionality, with the circumstances under which debt limits should feature as part of fiscal conditionality depending on country conditions and program objectives.10

13. The use of limits (or targeted sub-limits) on debt accumulation is normally warranted when a country has significant debt vulnerabilities. Debt sustainability depends not only on the level and expected trajectory of debt, but also on such features as the maturity of loans (including grace periods), the interest rate (e.g., market versus concessional) on these loans, the currency composition (domestic or foreign) of debt, and the creditor base.11 Where debt sustainability depends on ensuring that new debt has specific financing characteristics (e.g., long maturities to avoid bunching of repayments; concessional terms to limit the debt burden), these can be addressed via appropriately targeted conditionality.

14. The use of debt limits would also normally be warranted when the quality and coverage of the fiscal statistics produced by the national system of fiscal accounting and budgeting favor the use of debt conditionality instead of, or as a complement to, “above-the-line” fiscal conditionality. Specifically:

  • There would be merits in setting quantitative fiscal conditionality on measures of fiscal financing flows (“below-the-line” data), rather than on measures such as the fiscal balance (“above-the-line” data), if the quality and timeliness of the financing data is significantly better than the data on “above-the-line” flows.12

  • There could be merits in using a public debt limit as a complement to fiscal budgetary targets in cases where important public debt-creating activities are not adequately captured in the fiscal accounts (e.g., bank recapitalization, issuances of government guarantees, noncommercial state-owned enterprises, and other agencies outside the budgetary framework) and these activities pose a threat to the overall fiscal position.

15. Debt sustainability analysis (DSA) is the primary tool for identifying the extent of debt vulnerabilities. DSAs provide useful information not only on the extent of debt vulnerabilities, but also the type of debt vulnerabilities. For countries that rely primarily on market-based financing, debt vulnerabilities are assessed using the IMF’s Market Access Country DSA (MAC DSA).13 For countries that normally rely on concessional financing, including from the international financial institutions, debt vulnerabilities are generally assessed using the joint Bank-Fund LIC Debt Sustainability Framework (LIC DSF). The analysis provided in the joint Bank-Fund Medium Term Debt Strategy (MTDS) Framework, where available, can also help identify vulnerabilities in the debt structure.

16. For countries where the MAC DSA is used, the assessment of debt vulnerabilities is informed by the set of standard indicators in the MAC DSA, including the heat map, the tools to assess the realism of baseline assumptions, the fan charts, and other indicators. Typically, heat map indicators exceeding their upper benchmarks (either for the debt levels, gross financing needs, or the debt profile) would signal significant debt vulnerabilities. But analysis of the key drivers contributing to these vulnerabilities, as well as the magnitude of breaches, is warranted to inform conclusions, while allowance needs to be made for country-specific factors not adequately captured in the heat map (e.g., risks from public guarantees not included in the baseline).

17. For countries using the LIC DSF, the assessment of debt vulnerabilities is informed by the assessed risk of external debt distress or, where relevant, by the assessed overall risk of debt distress.14 In cases where there are significant vulnerabilities related to public domestic debt, the assessment would be made on the basis of the overall risk of debt distress as described in the LIC DSF (see IMF, “Staff Guidance Note on the Application of the Joint Bank-Fund Debt Sustainability Framework for Low-Income Countries,” November 5, 2013).

18. An assessed rating of moderate or high risk of debt distress using the LIC DSF would typically signal the presence of significant debt vulnerabilities; a low risk rating would signal the absence of significant debt vulnerabilities. The form of debt conditionality deployed in countries at high risk of debt distress would differ significantly from that deployed in countries at moderate risk of debt distress, given the elevated concerns regarding debt levels in the former group.15

19. The absence of debt conditionality in countries where debt vulnerabilities are not significant does not imply that a rapid build-up of the debt burden in such cases would be ignored in program design. A rapid build-up of debt would warrant a careful diagnostic of the factors at work. Should this diagnostic point to a need for policy corrections, these changes would need to be factored into the design and quantitative specification of the ensuing Fund-supported fiscal program; in some circumstances, this could involve appropriately targeted conditionality.16

B. What form should debt conditionality take?

20. Where use of limits on public debt is warranted, the specification of these limits will need to appropriately reflect country circumstances. The design of debt limits would thus be expected to vary, depending on both country conditions and program objectives. Limits might be set on either external debt or total public debt; target debt of specific maturities; be set as limits on the debt stock or on the contracting of new debt; be set in nominal or present value (PV) terms; and extend to any other transaction that may be debt-creating. Debt limits would normally continue to cover public and publicly guaranteed debt (PPG debt), or targeted sub-components of such debt (e.g., guarantees).

21. Some general principles regarding the specification of debt limits can be outlined as follows:

  • For countries with an open capital account and close integration with international financial markets, limits would typically not distinguish between domestically-held and externally-held debt;17 a limit on debt accumulation would typically cover total public debt.

  • For countries where there is significant segmentation between domestic and external sources of financing, either because of the presence of significant capital account restrictions or because the country receives substantial amounts of official external financing on nonmarket terms, there is an economically sound case for specifying distinct limits on external financing and domestic financing. That said, the appropriate form of any debt limit will depend on specific country circumstances.

  • In cases where much of the external financing takes the form of project loans that are disbursed over an extended period, it may be more appropriate, including for accuracy of monitoring, to specify the limit on external financing in the form of contracting of new debt, rather than on the disbursement of new debt.18

  • For countries where debt vulnerabilities are specific in nature, rather than linked to aggregate debt levels (e.g., bunching of maturities; weak controls on issuance of guarantees), the debt limit should be appropriately targeted on the specific areas of vulnerability.

  • The specification of debt limits needs to make appropriate allowance for debt management and monitoring capacity constraints, including a realistic assessment of the pace at which capacity building efforts in this area are likely to yield results.

  • The form of debt limit chosen should not inhibit countries from undertaking active debt management operations in response to changing market conditions (e.g., the prefinancing of future financing needs); the specification of adjustors can help address this issue.

22. The form of debt conditionality would continue to be guided by the Fund’s guidelines on program conditionality. Where the use of debt conditionality is so critical for achieving program objectives or monitoring implementation that interruption of disbursements under a Fund arrangement would be warranted in case of nonobservance, limits on debt should take the form of performance criteria (PCs). Where debt conditionality is critical to achieve objectives or monitor program implementation, but not so critical as to warrant interruption of disbursements, debt limits could take the form of indicative targets (ITs).

23. Given the varied options for debt limit design, the staff report accompanying the program request would contain an explanation for the specific selections made. In particular, the role of such factors as data quality and coverage and debt management capacity in influencing the specification of debt limits should be clearly laid out. Efforts underway to strengthen capacity in these areas should also be flagged.19

C. How should quantitative debt limits be set?

24. A borrowing (or financing) plan is an integral component of a country’s fiscal program, and hence of the planned macroeconomic policy framework.20 Assessment of this borrowing plan is thus one part of the overall assessment of the country’s macroeconomic plans, rather than a stand-alone element; program quantitative targets for debt accumulation are but one element in an agreed macroeconomic policy framework that is being supported under a Fund arrangement.

25. Several factors play a role in determining the appropriateness of the borrowing plan in a Fund-supported program. Compatibility of the borrowing plan with maintaining debt sustainability over the medium-term is a key concern; the borrowing space available would depend on the extent and nature of the country’s debt vulnerabilities. Other important aspects to be assessed include the feasibility of achieving planned borrowing levels at the envisaged terms and the implications of planned borrowing for debt composition and structure.

26. The level of borrowing to be accommodated in a program also reflects a wider assessment of the proposed macroeconomic policy framework: relevant factors include the appropriateness of the fiscal deficit from a demand management perspective, the envisaged trajectory of public investment and savings, the composition of public spending, the feasibility of implementing new spending programs (including investment programs) given capacity constraints, and so forth.

27. When country plans envisage significant increases in borrowing levels, fiscal and debt sustainability prospects need to be assessed with particular care. Where such plans reflect a significant expansion of public investment, the plausibility of the growth projections and their consistency with (a) the planned level and composition of investment and (b) public investment management capacity warrants close scrutiny.21 In assessing the expected growth payoff of investment, Fund staff can use a variety of approaches, including model-based analysis and available third-party assessments of individual large-scale projects.22 In assessing public investment management capacity, Fund staff can draw on a range of sources, including various standardized capacity assessments and relevant technical assistance reports.

28. In conclusion, the specification of quantitative limits on debt accumulation “drops down” from the features of the agreed fiscal program and macroeconomic policy framework. Debt sustainability analysis is one element of the assessment of the adequacy of this program, as is an evaluation of the borrowing plan and its feasibility. Macroeconomic programs predicated on a large scaling up of public investment in pursuit of a large growth dividend warrant careful assessment, particularly when implementation capacity is weak.

D. How should implementation of debt policies be assessed?

29. The starting point for assessing program implementation in regard to policies on debt accumulation is reviewing the observance of specified program quantitative targets. Observance of limits specified as performance criteria are of critical importance, requiring waivers if a linked disbursement is to proceed and/or a review to be completed. Nonobservance of limits specified as indicative targets warrant careful investigation to establish whether this poses a threat to program success that would warrant remedial action.

30. Assessment of program implementation should also give due attention to examining whether the realized pattern of debt accumulation was aligned with program expectations. An important area of inquiry relates to the realized financing mix, including the currency composition, terms, and maturities of loans contracted: was the realized financing mix broadly as envisaged, and, if not, were the observed deviations a cause for concern? In cases where borrowing plans were predicated on the pursuit of specific large-scale projects or of major bond issues, did these plans materialize as expected—and, if not, were the deviations observed a particular cause for concern? In cases where borrowing plans were linked to specific sectoral investment strategies (e.g., public infrastructure provision), was this strategy implemented as envisaged?

31. Program documentation needs to include sufficient detail on the key features of the envisaged borrowing plan to provide the basis for proper assessment in ensuing program reviews.23 The key features of the borrowing plan will depend on country circumstances: as examples, a) the maturity structure of new borrowing could be key in countries facing large rollover needs; b) the currency composition of new borrowing could be key in countries facing significant exchange rate pressures; c) tapping distinctive “market niches” (such as nonresident citizens) could be key in countries where significant net portfolio outflows by mainstream investors are expected; d) the contracting of specific project loans could be key in countries where the size of these projects is large relative to national output; and e) the sectoral focus of project financing could be key in countries contracting substantial amounts of new loans to “scale-up” public investment.

IV. Public Debt Conditionality in Countries That Normally Rely on Official Concessional External Financing

32. We consider here the role of debt conditionality in countries that normally rely on the provision of official concessional external financing; this was the area that Executive Directors and external stakeholders agreed should be the priority focus of any DLP reform proposal. Countries are deemed to normally rely on official concessional external financing if such funding usually accounts for the bulk of public external financing. By construction, countries that normally rely on the provision of official concessional external financing are characterized by significant segmentation between this form of external financing and (market-based) financing from domestic sources.

A. When is the use of debt conditionality justified?

33. As discussed in Section III.A, the use of debt conditionality may be justified on the basis of (a) the presence of significant debt vulnerabilities or (b) fiscal data quality or coverage concerns that favor the use of debt measures for specifying quantitative conditionality. To illustrate with some concrete examples:

  • In countries with significant debt vulnerabilities (as reflected in a moderate-to-high risk of debt distress), conditionality on the accumulation of debt would generally be warranted.

  • In countries where the quality and timeliness of the data produced by the budgetary accounting system is poor, basing fiscal conditionality on the evolution of the fiscal deficit may not be feasible or appropriate. In such cases, limits on debt accumulation—split into limits on (a) domestic credit to government and (b) the accumulation of public external debt—has typically represented the most effective specification of fiscal conditionality.

  • In countries where debt sustainability is not a significant concern (as reflected in a low risk of debt distress) and where the quality and coverage of fiscal data justifies the use of ”above-the line” fiscal conditionality (e.g., on the fiscal balance), the use of debt conditionality would generally not be warranted.

B. What form should debt conditionality take?

34. As discussed in Section III.B, debt conditionality, where warranted, would typically (but not necessarily) take the form of separate limits on public external and domestic debt accumulation, given the segmentation of financing sources:

  • For countries that normally rely on official concessional financing, there will typically be sizeable differences between the nominal and the present value of loans; to accurately capture the burden of the new debt being incurred, the limit on external debt accumulation should be specified in present value terms.24

  • As noted in Section III.B, when a large share of new external debt takes the form of project loans disbursing over several years, the limit on external debt would likely take the form of contracting or guaranteeing of new debt rather than on the disbursement-based incurring of new debt.

  • In situations where a significant share of debt in local currency is in fact external financing by foreign portfolio investors, the segmentation between external and domestic financing may be more apparent than real: in these cases, there may be a case for setting the debt limit on total public debt accumulation.

35. In countries where the capacity to monitor the evolution of debt is weak, the specification of debt limits needs to make appropriate allowance for these capacity limitations. The most significant weaknesses in debt monitoring capacity are likely to lie in the area of adequately capturing and tracking the contracting and disbursement of new external loans; quantifying the present value of a loan whose terms are known using a uniform discount rate should create fewer difficulties. In such cases, specification of debt limits along the following lines would be appropriate:

  • A performance criterion, specified in nominal terms, on the contracting of non-concessional external borrowing (the current approach), supplemented by a limit, again specified in nominal terms, on the contracting of new concessional debt. This limit would be explicitly specified in program documentation and included as a memorandum item in the standard quantitative conditionality table.25

  • As debt management and monitoring capacity is strengthened, the specification of these limits would be modified over time to converge to a present value limit on all new public external debt.

36. In assessing capacity to manage/monitor debt, staff should draw on the various sources of information available, including technical assistance reports, ratings and sub-scores provided under the Debt Management Performance Assessment (DeMPA), the Public Expenditure and Financial Assessment (PEFA), and Country Policy and Institutional Assessment (CPIA), etc. (see Annex III). Staff would collaborate closely with World Bank staff in forming an assessment of debt monitoring capacity, as is currently the practice in assessing debt and public investment management capacity in implementing the current debt limits policy.

37. When a country’s capacity to adequately capture and track the contracting and disbursement of external loans is deemed to be weak—a serious deficiency in public financial management capability—development of effective capacity in this area should be given high priority. Strengthening debt monitoring capacity should be an explicit objective of the Fund-supported program, with sufficient external technical assistance, including from the Fund, being mobilized to aid the authorities in building capacity in a timely manner. The expectation would be that such capacity-building efforts would bear fruit over the course of a three-year period, sufficient to allow the use of a comprehensive debt limit in any ensuing program.26 27

C. How should quantitative debt limits be set?

38. As discussed in Section III.C, quantitative limits on debt accumulation are derived from the agreed fiscal framework; the factors influencing the level of borrowing that can be accommodated are discussed in paragraphs 25–28. In this fundamental respect, the approach to quantification of debt limits would be similar across all program cases, whether financed from the GRA or the PRGT. That said, countries that have access to significant amounts of concessional financing are typically low income countries where public investment management capacity and debt management capacity are likely to be weaker than in higher income economies: the influence of such factors in assessing the proposed fiscal framework, including public investment and external borrowing levels would typically figure more significantly in staff assessments than would be the case in higher income countries.28

39. The levels of debt that can be accommodated within the fiscal program depend both on the assessed risk of debt distress and, within a risk category, on the scale of the existing debt burden. Thus, within the “moderate risk” category, countries with debt burdens closer to the levels consistent with achieving a “low risk” assessment would have larger potential borrowing space than would those countries with debt burdens approaching the levels that would trigger a “high risk” assessment.

40. Staff assessment of the borrowing plan needs to take account of the fact that loans contracted on non-concessional terms and loans contracted on concessional terms are typically imperfect substitutes for one another; the former should not be stigmatized or deemed as inherently inferior to the latter. All other things being equal, a loan on concessional terms is clearly preferable to a loan on non-concessional terms. In practice, the concessional and non-concessional loans potentially available to a borrower usually differ significantly from one another, whether it be in terms of the types of expenditure they finance (e.g., infrastructure versus social sector projects) or the conditionality to which use of the loans is subject (e.g., timeliness of decision-making on the part of the lender). Comparisons across alternative loans need to take full account of these differences, and of the borrowing country’s developmental priorities.

D. How should implementation of debt policies be assessed?

41. As discussed in Section III.D, assessment of implementation in program reviews would examine outcomes in regards to (a) observance of specified program quantitative targets and (b) consistency with the programmed borrowing plan. In this context, assessment of the implementation of the borrowing plan would need to cover both the realized financing mix (currency composition, terms and maturities, concessionality mix) and the extent to which high profile components of the borrowing plan (e.g., sovereign bond issues, large project loans) had evolved as anticipated. Nonobservance of a performance criterion would, of course, require investigation and an assessment as to whether, on the basis of the relevant circumstances of the nonobservance or any planned remedial actions, staff would propose the granting of a waiver for nonobservance. Assessment of the implementation of other components of the borrowing plan would be judgment-based and could point to a need for modifications to the program, depending on the specific circumstances.

42. Assessment of the implementation of borrowing plans in program reviews would necessitate specification of the key features of the borrowing plan in program documentation. Key features would include, among others: a) a breakdown of sources of new borrowings across different categories of concessionality, along the lines contained in Table 1; b) identification of any large scale loans envisaged as part of the borrowing plan (e.g., large external bond issues, large project loans); and c) the provision of aggregated information on the planned use of external financing across sectors, along the lines contained in Table 1, in those countries where project loans account for the bulk of external public financing.29 It should be noted that Table 1 includes program targets for both concessional and non-concessional borrowing levels—the latter number being a key element in implementation of the World Bank’s Non-Concessional Borrowing Policy (NCBP). The NCBP’s objectives and operational features are discussed in Annex IV.

Table 1.

Summary Table on External Borrowing Program: Illustration

article image

Contracting and guaranteeing of new debt. The present value of debt is calculated using the terms of individual loans and applying the 5 percent program discount rate.

Debt with a grant element that exceeds a minimum threshold. This minimum is typically 35 percent, but could be established at a higher level.

Debt with a positive grant element which does not meet the minimum grant element.

Debt without a positive grant element. For commercial debt, the present value would be defined as the nominal/face value.

43. The documentation of the borrowing plan would need to preserve country negotiating flexibility and the confidential nature of information, where relevant. To avoid hampering the country’s ability to press for the most favorable possible terms on future credits, program assumptions should not be provided for the terms (grant element) of specific loans. These concerns would be less significant in an ex post context, enabling staff to provide more detail, if warranted, on debt contracted—albeit still with appropriate attention being given to any confidentiality concerns.30

44. The provision of information on the borrowing plan will serve other purposes, aside from monitoring program implementation. At a technical level, information on expected levels of concessionality will make clear the relationship between the nominal debt, financing assumptions in the fiscal accounts and balance of payments with the corresponding present value or average grant element figures captured in the debt limits. At the same time, information on the financing strategy, including the projected concessionality element, will provide official creditors with a degree of confidence regarding the country’s debt management goals and practices and capacity to repay. The provision of details on the balance between concessional and non-concessional financing also helps address burden-sharing concerns, thereby helping ensure continuing access to financing on concessional terms.

E. Special Cases

45. Debt limits would be more tightly framed for countries at high risk of debt distress (or in debt distress). For these countries, it is expected that non-concessional borrowing would be exceptional—but not precluded, given the imperfect substitutability between available concessional and non-concessional loans. Consistent with this approach, programs in such cases would include a performance criterion on non-concessional borrowing, coupled with a performance criterion or indicative target on the contracting of concessional debt. In these country cases, where containing debt accumulation is critical, there could also be grounds (conditional on support from the country’s key bilateral donors) for defining concessionality as entailing a higher grant element than the 35 percent level typically employed in Bank-Fund operational work.

46. There may be circumstances in which the likely terms of a large loan (or loans) are sufficiently uncertain as to generate a significant margin of uncertainty on the projected present value of the new debt to be contracted in the borrowing plan. In such situations, the use of appropriately specified adjustors would be warranted to accommodate this uncertainty.31 The uncertainty regarding terms would need to involve large loans (if the present value target is to be significantly impacted by the uncertainty) provided on below-market terms; the logic for the adjustor should be explained in program documentation.32

F. On Concessional Lending to Low Income Countries

47. Staff will continue to advocate for the provision of financing on fully concessional terms to low income countries (LICs), defined here as all countries eligible to obtain concessional financing from the Fund. The existing debt limits policy sent a clear message to official lenders on the importance that the Fund attached to ensuring that LICs have access to substantial external financing on concessional terms—in effect, providing a significant grant element to loans to enhance the net impact on growth and poverty reduction. This message has not changed. Ensuring that official lenders continue to provide adequate volumes of concessional financing may require reaching understandings among creditors that they will collectively avoid a competitive erosion of concessionality; such understandings may be more easily reached in the context of a wider dialogue between official lenders and low-income country borrowers on “pro-development” lending practices than via a dialogue limited to creditors.

V. Guidelines on Public Debt Conditionality: Key Elements

48. We describe here the key elements of the proposed Guidelines on public debt conditionality: further details and implementation issues shall be fleshed out in a Guidance Note for staff in due course. The discussion here focuses only on the circumstances under which debt conditionality should be used and the form that it should take: determination of the quantitative targets for debt accumulation in an individual case will depend on country conditions (discussed earlier), including the availability of borrowing space as determined by the DSA.

49. The authorities’ borrowing plan is one component of the fiscal program being supported under a Fund arrangement. The use of limits on debt accumulation is a tool to be deployed in designing fiscal conditionality; the circumstances under which debt limits should feature as part of fiscal conditionality depend on country conditions and program objectives.

50. The use of limits (or targeted sub-limits) on debt accumulation is normally warranted when: a) a country has significant debt vulnerabilities or b) when the quality and coverage of fiscal statistics produced by the national system of fiscal accounting and budgeting favor the use of debt conditionality instead of, or as a complement to, “above-the-line” fiscal conditionality.

51. The approach taken to setting quantitative debt limits is similar across countries: evaluation of proposed borrowing plans, informed by the assessment of debt vulnerabilities in the DSA, is one component of the assessment of the fiscal program of the government. Program targets for debt accumulation are set as a “drop-down” from the agreed fiscal program.

52. For those countries that do not normally rely on official external concessional financing, the following Guidelines are envisaged.

  • Debt conditionality would be expected to be deployed in cases where countries are judged to have significant debt vulnerabilities, as assessed using the MAC DSA.

  • Limits on public debt would be specified in terms of the nominal value of debt, with the precise specification depending on country circumstances and data availability. Depending on the extent of financial market integration, separate limits on external-issued debt and domestically-issued debt could be justified.

  • Limits on public debt would ideally cover all public and publicly guaranteed debt, but institutional circumstances and data availability could justify the use of narrower coverage of the public sector.33 Targeted conditionality, focusing on specific components of public debt (e.g., short-term external debt, issuance of public guarantees) could be justified on the basis of specific debt vulnerabilities and institutional weaknesses.

53. For those countries that normally rely on official external concessional financing, the following Guidelines are envisaged:34

  • Program conditionality for countries assessed as facing a low risk of debt distress need not include limits on public external borrowing.35

  • Program conditionality for countries facing a moderate risk of debt distress, would include a performance criterion on the accumulation of external debt. The PC would cover all forms of public external borrowing (i.e., both concessional and non-concessional) and would be specified in net present value (NPV) terms.

  • For countries identified as being at high risk of debt distress (or in distress), the current debt conditionality framework would not change significantly: non-concessional external borrowing would be allowed only under exceptional circumstances and there would be a performance criterion set on the allowed nominal level of non-concessional external borrowing. In addition, a limit on the accumulation of concessional external debt would be specified, either in the form of a performance criterion or an indicative target.

  • In cases where limits on external borrowing are an integral component of fiscal conditionality, the use of external debt limits would be accompanied by appropriately specified limits on the accumulation of domestic debt. In cases where limits on external borrowing are a supplement to “above-the-line” fiscal conditionality, limits on domestic debt accumulation may not be needed.

54. Exceptions to the guidance specified in paragraph 53 would be accommodated in the following circumstances:

  • In countries with an open capital account and significant financial integration into international markets, it may be more appropriate to set a limit on total public debt accumulation rather than on externally-issued debt.

  • In countries where the use of debt conditionality is warranted but the capacity to capture and monitor the contracting of debt is weak, the performance criterion would take the form of a limit on the contracting of non-concessional external debt.36 To enhance the focus on ensuring centralized control and monitoring of concessional debt, a target for the contracting of concessional external debt would be specified and included as a memorandum item in the conditionality table.

Table 2 contains a listing of 61 PRGT-eligible countries, classified on the basis of their current debt distress rating.

Table 2.

PRGT-Eligible Countries, Classified by Most Recent Debt Rating 1/2/

article image
Source: Fund Staff LIC DSAs.

Ratings as of July 31, 2014 and excludes countries that have never had a Fund arrangement or cases where the DSA was not available (cases where no DSA was conducted). Five countries without a published debt sustainability analysis in over one year are not shown in the table.

Countries that are in bold currently have a Fund-supported program or staff monitored program.

Zimbabwe is not PRGT-eligible due to its removal from the PRGT-eligibility list by a Board decision in connection with its overdue obligations to the PRGT. It would be expected to become PRGT-eligible if the remedial measure were to be lifted.

VI. Transitional Arrangements

55. Given the significance of the reforms proposed, it is expected that the policy will be introduced on a gradual basis. It is proposed that the policy would take effect at end-June 2015. Conditionality in pre-existing Fund-supported programs would be modified only when understandings on such modifications have been reached between staff and the member’s authorities and the modifications have been approved by the Executive Board. Staff would be expected to discuss such modifications in the context of the discussions for the first program review following the entry into effect of the new debt limit guidelines.

56. A review of experience in implementing the policy would be conducted once a sufficient body of evidence has been accumulated to allow proper assessment, but no later than three years after the entrance into effect of the new policy. This period will allow for adequate time to take stock of experience with implementation and inform whether some aspects of the policy would need to be further refined.

VII. Issues for Discussion

  • Do Directors agree that the use of debt conditionality should be guided by: a) the extent of a country’s debt vulnerabilities; and b) the relative technical merits (in terms of ease of monitoring, comprehensiveness of coverage, etc.) of using “below-the-line” debt measures versus “above-the-line” fiscal balance measures in specifying fiscal conditionality?

  • Do Directors agree on broadening the scope of the current guidelines to cover public debt (both external and domestic) from its current focus on public external debt, recognizing that in many cases there will be grounds for specifying separate limits on domestic borrowing and on external borrowing?

  • Do Directors agree on broadening the scope of the debt conditionality guidelines used in countries that normally rely on concessional financing to cover all public external debt, rather than only non-concessional external debt?

  • Do Directors agree that program conditionality for countries assessed as facing a low risk of debt distress need not include limits on public external borrowing?

  • Do Directors agree that debt limits, where warranted, should be specified in net present value terms in the case of countries that normally rely on concessional external financing, recognizing that there may be specific circumstances (as specified in Section V) where this guidance would not apply?

  • Do Directors support the proposal that a country’s capacity to manage/monitor debt for program monitoring purposes be assessed along the general lines described in Annex III

  • Do Directors agree that program documentation should include a description of the key features of the authorities’ borrowing plans?

Annexe I. Debt Limits in Recent GRA Programs

1. In recent GRA programs the design of debt conditionality reflected concerns about debt vulnerabilities, as well as interaction between fiscal conditionality and program objectives. Table AI summarizes assessment of debt vulnerabilities, as well as design and coverage of fiscal and debt conditionality for a sample of six GRA programs covering the period from 2010–14. Where debt level was a key concern, program conditionality generally included overall debt limits on the stock of debt (Cyprus and Ireland), except where fiscal conditionality was sufficiently broad to capture most debt-creating transactions (Jordan). The specification of the debt conditionality varied depending on its criticality for meeting program objectives: in the case of Cyprus, the debt limit was specified as a PC; in the case of Ireland, the debt limit was set as an IT, supplementing fiscal conditionality by covering extra-budgetary debt creating operations. Where debt levels were low, debt limits were sometimes included to address risks relating to the creation of debt outside the fiscal framework itself (e.g., public guarantees (Ukraine), the composition of debt (Bosnia and Herzegovina).

Table IA.

Debt Limits and Fiscal Conditionality in Selected GRA Programs

article image

Assessed either under the current MAC DSA if program is ongoing or recently ended, or by the level of debt-to-GDP at the time of the program (if it expired or ended before the new MAC DSA framework was in place).

The relevant benchmarks for the debt-to-GDP ratio in the MAC DSA are: 70 percent for EMs, 85 percent for AEs.

GFN stands for gross financing needs.

Annex II. Debt Developments in PRGT-Eligible Countries since 2007

Public debt levels (as a share of GDP) have, on average, been stable in PRGT-eligible countries between 2007 and 2013, but the average masks significant differences both across and within country groupings. Public debt levels have increased in 15 of the 20”early HIPC” (defined below), with two-thirds of the increase accounted for by higher external debt levels; there is no clear trend observable for “non-HIPCs” (countries that did not receive HIPC/MDRI debt relief). In a number of countries, debt levels have increased quite sharply, although typically not to levels that have pushed countries into a higher debt distress risk category; in these cases, much of the debt increase has come in the form of higher levels of domestically-issued debt.

1. This annex examines the evolution of debt risk ratings and debt levels in PRGT-eligible countries (henceforth, LICs) between 2007 and 2013. The analysis focuses on countries that have a) had a program with the Fund at some point in time and b) have been the subject of at least one Debt Sustainability Analysis (DSA)—61 countries in all (Appendix Table 1). The group can be usefully divided into three subgroups for analytical purposes: a) “early HIPCs”—countries that had reached the HIPC completion point by 2007; b) “late HIPCs”—countries that have reached the HIPC completion point since 2007; and c) “non-HIPCs”—countries either not eligible for the HIPC initiative or eligible but not yet having reached the completion point.

Appendix Table 1.

Low Income Countries Considered for Debt Level Analysis 1/

article image

Countries that are in bold currently have a Fund-supported program or staff monitored program. Eleven LICs (Bhutan, Eritrea, Kiribati, Marshall Islands, Micronesia, South Sudan, St. Lucia, Timor-Leste, Tonga, Tuvalu, and Vanuatu) that have never had a Fund program or have only had assistance under the Rapid Credit Facility, and two LICs (Somalia and Uzbekistan) that have never had a LIC DSA, have been excluded from the analysis.

Early HIPCs refers to HIPCs that reached the completion point before January 1, 2007; late HIPCs refer to HIPCs that reached the completion point after January 1, 2007.

Eligible for HIPC; Chad has already reached the decision point under the HIPC Initiative.

Zimbabwe is not PRGT-eligible due to its removal from the PRGT-eligibility list by a Board decision in connection with its overdue obligations to the PRGT. It is expected to become PRGT-eligible if the remedial measure were lifted.

2. External borrowing by PRGT-eligible countries is on average highly concessional in nature, albeit with significant cross-country variation (Figure AII1). Staff estimates indicate that the average grant element of external loans contracted,1 weighted by loan size, was about 39 percent for PRGT-eligible countries during 2009–11, although the grant element varied markedly across countries. For example, in five countries in our sample the average grant element of new loans was below 20 percent. In some cases, the issuance of sovereign bonds contributed to the low grant element (e.g., Senegal).

Figure AII1.
Figure AII1.

LICs: Distribution of Countries by Average Grant Element of External Loans Contracted in 2009–11

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

Sources: World Bank WDI database; and Fund staff calculations.1/ Grant element calculations are based on 5 percent discount rate. Sample includes 60 LIC countries (excludes Zimbabwe as data is not avalable).

A. Evolution of the Risk of External Debt Distress in the LIC DSF

3. As of July 2014, one third of the countries in the sample are at low risk of debt distress while over two-fifths are in moderate risk of debt distress (Table AII1). Countries at high risk of debt distress include (i) countries in conflict or post-conflict situations (such as Afghanistan and the Central African Republic); (ii) countries that suffered from natural disasters (such as Haiti and Samoa); and (iii) several small island economies with large fiscal deficits and vulnerabilities to external shocks. Grenada, Sudan and Zimbabwe are in debt distress.

Table AII1.

PRGT-Eligible Countries, Classified by Most Recent Debt Rating 1/2/

article image
Source: Fund Staff LIC DSAs.

Ratings as of July 31, 2014 and excludes countries that have never had a Fund arrangement or cases where the DSA was not available (cases where no DSA was conducted). Five countries without a published debt sustainability analysis in over one year are not shown in the table.

Countries that are in bold currently have a Fund-supported program or staff monitored program.

Zimbabwe is not PRGT-eligible due to its removal from the PRGT-eligibility list by a Board decision in connection with its overdue obligations to the PRGT. It would be expected to become PRGT-eligible if the remedial measure were to be lifted.

4. Six countries have experienced a downgrade in their risk rating since 2007; while thirteen countries (excluding the late-HIPCs) have experienced an upgrade over the period (Table AII2). Seven countries experienced both an upgrade and downgrade in their risk ratings during this period, with no net change since 2007.

Table AII2.

LICs: Evolution of DSA Risk Ratings, January 2007–July 2014

article image
Sources: Fund Staff Reports.

DSA Risk rating changes happened at a time when the country is under a Fund program.

Including 7 countries that had both an upgrade and a downgrade between 2007 and 2014.

5. Risk rating downgrades result from a range of factors including macroeconomic, fiscal, or external shocks. Fourteen countries have had a risk rating downgrade at least once between 2007–13, of which six of them currently have a poorer risk rating than in 2007 (Table AII3).2 Common reasons for downgrades have been: a) weaker than projected macroeconomic outlook; b) weak fiscal performance; c) the contracting of significant amount of external debt; and d) the impact of changes in discount rate or CPIA ratings.

Table AII3.

Deterioration of Risk Ratings*1/2/

article image

Three countries whose recent DSAs are not published are not shown in the table.

As of July 31, 2014.

Early HIPC refers to HIPCs reached completion point before January 1, 2007; late HIPCs refer to HIPCs reached completion point after January 1, 2007.

Eligible for HIPC; Chad has already reached the decision point under the HIPC Initiative.

  • Weaker macroeconomic outlook. At least six episodes of downgrades (including Central Africa Republic, Mali, Mongolia, Samoa, and Sao Tome and Principe) involved a downward revision to macroeconomic variables and/or exogenous shocks. The causes of revision include lower natural resource production and exports (Mali, Mongolia, and Sao Tome and Principe), conflict (Central African Republic) and natural disasters (Samoa).

  • Fiscal path. Some episodes of downgrades were driven by lower fiscal revenue (Cape Verde), or slower-than-projected fiscal adjustment.

  • Contracting of external debt. The downgrades of three early HIPCs (Cameroon, Mozambique, and Niger) and three non-HPICs (Cape Verde, Chad, and Mongolia) were associated with contracting of a significant amount of debt to finance public investment in infrastructure or natural resource projects.3

  • Changes in discount rate or CPIA ratings. Changes in the discount rate were a contributing factor to downgrades in two cases (Mongolia and Mozambique); the downgrading of Burkina Faso in 2008 (subsequently reversed in 2012) resulted from a decline in its CPIA score, which resulted in lower debt thresholds.

B. Trends in Debt Levels

6. Public debt levels in LICs have, on average, been broadly stable between 2007 and 2013 (Figure AII2).4 But the aggregate data masks important differences across the different country groups (Figure AII3). Average debt levels have declined in late HIPCs—given the impact of debt relief—while rising in most early HIPCs and about half of the non-HIPC grouping. The variation in experience across countries is wide, as reflected in Figure AII4.

Figure AII2.
Figure AII2.

LICs: Public Debt, 2007–13

(weighted average, in percent of GDP)

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

Sources: WEO; and Fund staff estimates.
Figure AII3.
Figure AII3.

LICs: Public Debt by Country Groups, 2007-13

(weighted average, in percent of GDP)

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

Sources: WEO; and Fund staff estimates.
Figure AII4.
Figure AII4.

Changes in Public Debt Level in LICs, 2007 vs. 2013

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

Source: WEO.1/ Afghanistan, Madagascar, Mongolia, and Samoa are excluded due to missing public debt data.2/ Bolivia, Nicaragua, Papua New Guinea, Samoa, and Zimbabwe are excluded due to missing public external debt data.

7. Number of countriesMost of the early HIPCs have seen rising debt levels since 2007 (Figure AII5). Public debt (as a share of GDP) increased in 15 of the 20 early-HIPCs between 2007 and 2013, with about two-thirds of the increase being accounted for by external debt. For those early HIPCs whose debt increased, the public debt-to-GDP ratio increased by an average of 12 percentage points, of which 7 percentage points reflected an increase in public external debt. The largest increase in public external debt was observed in Senegal and Tanzania, largely reflecting a scaling-up of public investment, and in Honduras, stemming from a large deterioration of the fiscal position. In countries with the largest increase in total public debt (Ghana and Malawi), the bulk of the increase was accounted for by domestic borrowing. In addition to borrowing from official creditors, a number of early HIPCs (such as Bolivia, Ghana, and Senegal) tapped international capital markets through the issuance of sovereign bonds.

8. No clear pattern emerges in reviewing the evolution of debt in non-HIPCs. Debt increased in about half of the non-HIPCs and declined in the other half. The main reasons for the increase in debt vary across countries: in some non-HIPCs, it reflected the results of sustained low growth and poor fiscal performance (such as Grenada and Maldives). Other countries carried out large public investment programs financed with external or domestic debt (such as Mongolia). A few countries under prolonged debt overhang (such as Sudan and Zimbabwe) also saw an increase in debt owing in part to accrued late interest and penalty fees.

Figure AII5.
Figure AII5.

Debt Level in Early HIPCs, 2007 vs. 2013 1/

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

Sources: WEO.◆ Red diamonds are countries experiencing deteriorating risk of debt distress during 2007-14 (Cameroon, Mozambique).1/Early HIPC refers to HIPCs reached completion point before January 1, 2007. Mauritania was excluded as an outlier (public debt-to-GDP ratio were 97% and 88% in 2007 and 2013, respectively).
Figure AII6.
Figure AII6.

Debt Level in Non-HIPCs, 2007 vs. 2013 1/

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

Sources: WEO.◆ Red diamonds are countries experiencing deteriorating risk of debt distress during 2007-14 (Cape Verde, Dominica, Maldives).1/Two HIPC-eligible countries, Chad and Sudan, are included in this chart as they have not (fully) benefited from the debt relief.

9. It is informative to look more closely at those countries where debt levels have risen significantly—defined here as those countries that experienced a) an increase in the public debt-GDP ratio during 2007–13 of at least 5 percentage points, and b) an increase in the public external debt-GDP ratio over the same period of at least 5 percentage points (Appendix Table 2). Thirteen of the countries reviewed met this threshold; raising the threshold levels to require an increase of at least 10 percent points in the public debt-GDP ratio would eliminate three early HIPCs from the list. Early HIPC cases that stand out include Ghana and Malawi (where the public debt burden has risen by close to 30 percentage points of GDP, mainly on the back of increased domestically-issued debt) and Senegal (an increase in the public debt-GDP ratio of 22 percentage points, the bulk of it financed externally). Honduras and Tanzania have also recorded increases in the external debt-GDP ratio of at least 10 percentage points of GDP.

Appendix Table 2.

Changes of Total Public Debt and Public External Debt Level in Early HIPCs and Non-HIPCs

(List of countries with both changes > 5 percent of GDP)

article image
Source: WEO; staff reports; and compilation of country teams responses.

Early HIPC refers to HIPCs reached completion point before January 1, 2007.

Annex III. Assessing the Quality of Debt Monitoring

This annex provides an overview of existing indicators that could be used by country teams to identify weakness in debt monitoring and reporting in countries that rely significantly on official external financing on concessional terms. These indicators are part of wider frameworks developed and regularly applied by the World Bank to assess debt management practices. Preliminary analysis suggests that in about 40 percent of PRGT-eligible countries debt records are either incomplete or produced with significant delays. In cases where potential weaknesses in debt monitoring, and in particular in tracking of concessional debt, are confirmed, the specification of debt limits should make appropriate allowance for these capacity limitations.

A. Background

1. As discussed in Section II.B, the specification of debt limits needs to make appropriate allowance for a country’s debt management and monitoring capacity constraints. The proposed design of debt limits for countries with significant access to concessional resources would require monitoring of a larger number of loans. Moving from a system where no real-time monitoring of concessional debt is required to a performance criterion covering both concessional and non concessional debt on a continuous basis may lead to more frequent occurrence of misreporting in countries with weaker administrative capacity.

2. A few indicators could help to identify weakness in debt monitoring and reporting. Together with countries’ track record and relevant TA assessments, these indicators could be used to identify countries that would not be able to move to a performance criterion encompassing both concessional and non-concessional debt right away. In contrast with the existing capacity assessment exercise, which covers various aspects of the management of public resources, the diagnostics described below focuses on the assessment of the quality of debt monitoring, i.e., the timeliness and completeness of debt records and reports at least at the central government level.1

B. Available Indicators for Assessing the Quality of Debt Recording and Reporting

3. The quality of debt monitoring could be assessed using selected components of the three relevant frameworks applied by the World Bank and its partner institutions: the Country Policy and Institutional Assessment (CPIA), the Public Expenditure and Financial Assessment (PEFA), and the Debt Management Performance Assessment (DeMPA). These frameworks cover a wide range of aspects related to management of public debt and resources (see Box AIII1) including the quality of debt data recording and reporting. Table AII1 describes and compares the sub-indicators (PEFA’s PI-17(i), DeMPA’s DPI-14(1, 2) and DPI-15 (1, 2, 3), and CPIA’s A3) which assess debt monitoring practices.

Existing Frameworks to Assess Management of Public Debt and Resources

The Debt Management Performance Assessment (DeMPA)’s main purpose is to assess strengths and weaknesses in public debt management (PDM) operations and ensure that debt management weaknesses would not endanger the gains from debt relief. DeMPAs assessment covers six core functions of PDM: (1) governance and strategy development; (2) coordination with macroeconomic policies; (3) borrowing and related financing activities; (4) cash flow forecasting and cash balance management; (5) operational risk management; and (6) debt records and reporting. DeMPA assessments are undertaken and published at the authorities’ request.

The Public Expenditure and Financial Assessment (PEFA) framework measures performance of a country’s Public Financial Management (PFM). It offers a comprehensive analysis of the PFM, using 28 indicators grouped in three areas: credibility of the budget (4 indicators); comprehensiveness and transparency (6 indicators) and budget cycle (18 indicators), which include elements of both ex-ante and ex-post scrutiny.

Country Policy and Institutional Assessment (CPIA) index consists of 16 indicators grouped into four categories: (1) economic management; (2) structural policies; (3) policies for social inclusion and equity; and (4) public sector management and institutions. Countries are rated on their current status in each of these performance criteria, with scores from 1 (lowest) to 6 (highest).

4. Indicators differ by relevance and frequency of updates (Table AIII1). In general, the DeMPA framework allows for a deeper assessment by producing separate scores for different aspects of debt recording and debt reporting. One of the sub-indicators on debt reporting (DPI-15) also extends its coverage beyond the central government. However, DeMPA reports usually remain confidential. PEFAs do not have the same level of detail: its PI-17(i) rating for the quality of the central government debt data recording and reporting corresponds to four dimensions in the DeMPA framework. However, PEFA assessments are updated more frequently (on average every three years) and, because they are published, are available for a larger sample of countries. The CPIA’s score for the A3 component on debt policy and management is produced annually, but is even broader: in addition to assessing the quality of public debt data, the score also reflects the risk of debt distress and the assessment of debt management practices in general.

5. Available data suggests that there is a good correlation between assessments made under different frameworks, and various indicators could be used to complement each other in identifying potential weaknesses in debt monitoring practices. Table AIII2 summarizes data on publicly available PEFA/DeMPA ratings and CPIA Debt policy scores for PRGT-eligible countries. Countries are grouped by PEFA ratings, and sorted by CPIA Debt Policy score within each group. DeMPA scores are included, where available. PEFA ratings and DeMPA scores C and above indicate that minimum requirement for a selected dimension has been met. For the CPIA, adequate debt recording systems correspond to scores of above 3.2 While some differences could be explained by time lags between the assessments and the publication of results, two main observations that stand out based on the data in Table AIII2.

  • First, in cases where PEFA and DeMPA reports were issued around the same year, the assessments generally produce similar results. Therefore, where recent DeMPA (PEFA) scores are not available, the assessment could rely on the assessment produced by the alternative framework.

  • Second, low CPIA scores may also signal potential weaknesses in the monitoring systems. Given the differences in the scoring systems, the direct comparison of PEFA ratings and CPIA scores is not possible. However, as evident from Figure AIII1, the average CPIA Debt Policy score for countries with PEFA rating set at D (incomplete and inaccurate debt records) is significantly below averages for other three groups. With only a one exemption (Nigeria), D rating under the PEFA framework corresponds to CPIA scores below 3.

Figure AIII1.
Figure AIII1.

CPIA Debt Policy Scores Grouped by PEFA PI-17

(i) Ratings

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

C. Proposed Methodology to Indentify Cases With Potential Weaknesses in the Quality of Debt Monitoring

6. More weight should be given to lower ratings when looking for weaknesses in debt recording systems. Based on this principle, criteria to identify cases where the quality of debt monitoring may be insufficient would be the following:

  1. A country scoring of ‘D’ in any of the most critical dimensions including DPI-14(1), DPI-15(2) (where relevant), and PI-17(i) over the last three years; or

  2. A CPIA Debt Policy score equal to 3 or below.

Figure AIII2 describes step-by step process that could be applied to identify deficiencies in data monitoring.

Figure AIII2.
Figure AIII2.

Using DEMPA, PEFA and CPIA Scores to Diagnose Weaknesses in the Quality of Debt Monitoring

Citation: Policy Papers 2014, 083; 10.5089/9781498342568.007.A001

7. Score-based diagnostics suggests that in about 40 percent of PRGT-eligible countries debt records may be either incomplete or are produced with significant delays (Table AIII3). In about half of the cases the choice was triggered by low DeMPA and/or PEFA ratings, including several countries where CPIA scores were low as well. Overall, for majority of selected countries CPIA Debt Policy scores are equal or below 3. All of these countries are classified as lower capacity under the current capacity assessment framework.

8. The final assessment of a country’s capacity to adequately monitor its debt should take into account other evidence, such as recent track record, fragile state status, relevant TA reports, etc. The assessments would be produced by country teams at the time of a program request and updated in subsequent reviews. In cases, where the quality of debt monitoring, and in particular in tracking of concessional debt, is low, these capacity limitations should be appropriately treated in the specification of debt limits (where relevant) and addressed in structural conditionality. The program documents will be expected to provide justification for any deviation of the proposed design of debt limits from the one covering both concessional and non-concessional debt. As debt management and monitoring capacity is strengthened, the specification of these limits could be gradually enhanced over time to converge to a present value limit on all new debt.

9. A list of the 26 countries deemed to have weaknesses in the quality of debt monitoring using the above methodology is presented in Table AIII3.

Table AIII1.

Comparison of PEFA, DeMPA and CPIA Indicators on the Quality of Debt Data Recording and Reporting

article image

The overall score for the indicator is the average of the scores of the two components. In addition to criteria mentioned in the table, the assessment also takes into account risks related to contingent liabilities, legal framework for the public borrowing, availability of the medium-term debt management strategy.

Table AIII2.

Available Indicators on the Quality of Debt Data Recording and Reporting

article image
Sources: PEFA Secretariat; World Bank; and Fund staff calculations.

The CPIA rates countries against a set of 16 criteria grouped in four clusters: (a) economic management; (b) structural policies; (c) policies for social inclusion and equity; and (d) public sector management and institutions. The score used in this table reflects the average CPIA score of the overall score from 2011-13.

PEFA data contains the most recent status of a national assessment as of April 16, 2013. The data is updated on a six-monthly basis in which PEFA Partners and other agencies that lead PEFA assessments are contacted about the status of their assessments. The PEFA Secretariat collects and verifies this information before updating the assessment portal. For futher information on methodology, please visit http://www.pefa.org/en/content/pefa-framework-material-1

Each indicator seeks to measure performance against a four point ordinal scale from A to D. The highest score is warranted for an individual indicator if the core PFM element meets the relevant objective in a complete, orderly, accurate, timely and coordinated way.

Debt Management Performance Assessment Tool (DeMPA) is a methodology for assessing performance through a comprehensive set of performance indicators spanning the full range of government debt management (DeM) functions. The scoring methodology will assess each dimension and assign a score of either A, B, or C, based on the criteria listed. In the cases where a dimension cannot be assessed, an N/R score is assigned.

Program, Near Program, and World Bank only assessed LICs; as of June 26, 2013.

Table AIII3.

Cases with Significant Weaknesses in Debt Monitoring

article image
Sources: PEFA Secretariat; World Bank; and Fund staff calculations.

PEFA data contains the most recent status of a national assessment as of April 16, 2013. The data is updated on a six-monthly basis in which PEFA Partners and other agencies that lead PEFA assessments are contacted about the status of their assessments. The PEFA Secretariat collects and verifies this information before updating the assessment portal. For futher information on methodology, please visit http://www.pefa.org/en/content/pefa-framework-material-1.

Each indicator seeks to measure performance against a four point ordinal scale from A to D. The highest score is warranted for an individual indicator if the core PFM element meets the relevant objective in a complete, orderly, accurate, timely and coordinated way.

The CPIA rates countries against a set of 16 criteria grouped in four clusters: (a) economic management; (b) structural policies; (c) policies for social inclusion and equity; and (d) public sector management and institutions. The score used in this table reflects the average CPIA score of the overall score from 2011-13.

Debt Management Performance Assessment Tool (DeMPA) is a methodology for assessing performance through a comprehensive set of performance indicators spanning the full range of government debt management (DeM) functions. The scoring methodology will assess each dimension and assign a score of either A, B, or C, based on the criteria listed. In the cases where a dimension cannot be assessed, an N/R score is assigned.

Annex IV. The World Bank’s IDA Lending and Non-Concessional Borrowing Policy

1. The financing terms of World Bank IDA lending to IDA-only countries are determined by country ratings of the risk of external debt distress.1 These risk ratings stem from the World Bank-IMF debt sustainability framework (LIC DSF).

2. The World Bank’s IDA terms are designed to help slow the reaccumulation of unsustainable debt. Specifically, the credit-grant mix provided by IDA reflects the risk of external debt distress as assessed in the LIC DSF (the “traffic light system”). Countries at low risk of debt distress receive 100 percent of their IDA financing in the form of credits (currently 38 years’ maturity, 6-year grace period, interest rate of 0.75 percent); countries at moderate risk of debt distress are eligible for a 50–50 mix of IDA credit and grant financing;2 countries at high risk of debt distress are eligible for 100 percent IDA grant financing. To help ensure IDA’s long term financial sustainability, IDA contributors have agreed to reimburse IDA for any shortfall in repayments as a result of countries receiving grants.

3. The Bank’s Non-Concessional Borrowing Policy (NCBP), like the traffic light system, is also designed to help slow the reaccumulation of unsustainable debt. While not precluding non-concessional borrowing, the Bank’s NCBP allows the Bank to ensure that scarce IDA grant finance is used effectively to pursue debt sustainability and at the same time provides incentives for countries with moderate and high risk of debt distress, or those that have received debt relief under MDRI, to seek concessional financing. Should a country take on non-concessional borrowing that may undermine debt sustainability, the Bank may modify its IDA financing framework outlined above on a case-by-case basis. These modifications can take a number of forms: (i) converting allocations on grant terms to credit terms, (ii) a reduction in the volume of annual IDA allocations, (iii) a hardening of financing terms (applying a higher interest rate), and (iv) a combination of the above. Waivers are granted should there be country- and loan-specific justification for the non-concessional borrowing, inter alia based on strong economic returns of the project financed and limited impact on the DSA. Since 2006, roughly 30 cases of breaches of the limits set under the NCBP have taken place: the Bank has implemented adjustments to the financing terms or volume of IDA funding in three cases, with waivers having been granted in the remaining cases.

4. For countries implementing Fund-supported programs (including the Policy Support Instrument), the ceilings on non-concessional borrowing allowed under the NCBP have been based on the ceilings on NCB set in the Fund-supported program. Typically, waivers approved for breaches of NCB ceilings under Fund programs would translate into waivers under the Bank’s NCBP. This said, every request for a waiver to the IMF Board related to non-concessional borrowing is also assessed by the Bank’s Non-Concessional Borrowing Policy Committee. In countries not implementing Fund-supported programs, the Bank may, at the request of the country authorities, agree on NCB ceilings with its clients: these ceilings may be linked to the execution of specific projects or be untied.

5. In designing reforms to the Fund’s DLP, careful consideration has been given to ensuring consistency with IDA policies and the NCBP. There has been ongoing consultation between Fund and Bank staffs as the Fund’s reform proposals have advanced. In technical terms, the proposed reform of the Fund’s DLP will continue to generate agreed target levels for NCB, as illustrated in Table 1 of the main text, enabling the Bank, should it wish, to take these agreed targets as the relevant limits for implementation of the NCBP.3 More fundamentally, the Fund’s DLP will continue to focus on providing countries with the flexibility to manage their investment financing needs within a fiscal framework consistent with the need to maintain a sustainable debt position over the medium-term—aligned with the NCBP objective of containing the reaccumulation of unsustainable debt.

2

For the purpose of this paper, the term “programs” refers to programs supported by Fund arrangements in the General Resources Accounts (GRA), other than FCL arrangements, under the Poverty Reduction and Growth Trust (PRGT), or under the PSI. Unless otherwise specified, ‘debt limits’ and ‘debt conditionality’ refer to performance criteria or indicative targets establishing some form of limit or sub-limit on the accumulation of public and publicly-guaranteed debt.

4

“Concessional borrowing” is used throughout this paper to refer to loans with a grant-element of at least 35 percent, consistent with current Bank-Fund usage; “non-concessional borrowing” refers to loans not assessed as concessional. The term “semi-concessional” refers to loans assessed to have a positive grant element.

5

These included large threshold effects associated with marginal changes in loan terms, difficulties in obtaining the independent project assessments that were a corner-stone of the policy, and (prior to the unification of discount rates in October 2013) shifts in the classification of loans on which terms had already been determined following the regular adjustment of discount rates, contributing to unintended misreporting (see “Review of the Policy on Debt Limits in Fund-Supported Programs,” March 1, 2013).

6

Ibid, provides a detailed account of implementation of debt conditionality across the Fund membership.

7

Annex I illustrates the use of fiscal and debt conditionality for a sample of recent programs in the General Resources Account (GRA).

8

As of end-September 2014, all PRGT-supported programs included a performance criterion on the contracting of new non-concessional debt.

9

For the purposes of this paper, the terms “debt,” “borrowing,” and “loan” are used interchangeably and refer to the concept of debt set out in Executive Board Decision No. 6230–(79/140), points 3 and 9, as revised on August 31, 2009 (Decision No. 14416–(09/91)). Changes to these definitions of debt are not proposed here: anomalies associated with the application of these definitions (such as the handling of central bank repo transactions) can be handled on a case-by-case basis in a program’s technical memorandum of understanding.

10

Limits on debt accumulation are not always essential to the effective design of fiscal conditionality. In countries where debt accumulation reflects only budgetary operations, limits on new borrowing and limits on the fiscal balance would be substitutes for one another, with the relative merits of using one or the other depending on technical features such as data timeliness and ease of monitoring.

11

Throughout the paper, the term “structure of debt” is used as a short-hand to refer to the currency composition, maturity structure, financial terms, and investor base for government debt.

12

This could be the case if the financing data is drawn from sources other than the budgetary accounting system (e.g., from banking system balance sheets or debt stock data).

13

The framework for public DSA in market access countries (MAC DSA) is currently in use for non-PRGT-eligible countries (see IMF, “Staff Guidance Note for Public Debt Sustainability Analysis in Market-Access Countries,” May 9, 2013): while some of these countries have access to concessional resources, these resources typically do not account for a significant share of the country’s external financing (e.g., India). The LIC DSF is currently used for all PRGT-eligible countries, although the guidelines allow for use of the MAC DSA when the country has durable and substantial access to market financing and is not “IDA-only” in terms of access to World Bank resources.

14

There are four possible risk ratings under the LIC DSF: low, moderate, high or in debt distress.

15

Similarly, the levels of borrowing that could be accommodated in the fiscal program would typically differ significantly between countries at high risk of debt distress and countries at moderate risk of debt distress—and, within the “moderate risk” group, between countries with higher debt burdens and countries with lower debt burdens.

16

In situations where the debt build-up stemmed from overly large budget deficits, the policy correction would be to set appropriately tight fiscal targets. In situations where the debt build-up stemmed from borrowing by public sector entities not adequately captured in the fiscal accounts (e.g., state-owned enterprises), the imposition of conditionality on the borrowing levels of these entities could be warranted.

17

In such cases, there may still be a need to distinguish between foreign- and domestic-currency denominated debt for risk management purposes.

18

The contracting of project loans is under the direct control of the national authorities; the disbursements of these loans depend on the (uncertain) pace of project implementation.

19

In countries where debt management weaknesses are a significant cause for concern, programs will often include a capacity building component, supported by targeted technical assistance; where warranted, steps to strengthen debt management could be a focus for structural conditionality.

20

The borrowing plan—used here as a shorthand expression for the government’s plans to meet its projected financing needs—is an integral component of a country’s fiscal program in any circumstances (i.e., whether or not debt conditionality is being deployed).

21

Where plans reflect a decision to provide fiscal stimulus in the face of weak demand conditions, the strategy needs to be assessed in light of financing constraints, projected debt dynamics, and output-inflation trade-offs.

22

Analyzing the growth pay-off from investment is an integral part of the existing LIC DSA (see IMF, “Staff Guidance Note on the Application of the Joint Bank-Fund Debt Sustainability Framework for Low-Income Countries,” November 5, 2013). For examples of country applications of model-based approaches to assess the growth payoff of debt-financed investment in LICs, see IMF, WP/12/127, WP/13/237, and WP/14/44.

23

As used here, the term “program documentation” includes the authorities’ Letter of Intent (LOI) and Memorandum of Economic and Financial Policies (MEFP) and the accompanying staff report.

24

For countries that have very limited, or no, access to concessional financing, the case for focusing on present value would disappear, as the present value and nominal levels would be broadly similar.

25

Specification of conditionality (PC/ IT) on the level of new concessional borrowing would not be justifiable in cases where the authorities do not have adequate capacity to monitor this variable in a timely manner. Inclusion of the programmed level of concessional borrowing as a memorandum item in the conditionality table will ensure that the level of such borrowing is systematically monitored by staff over time and reported on in program documents.

26

Should the assessment be made, prior to the commencement of a new program, that capacity-building efforts had failed to achieve this objective, an explanation of the factors impairing the capacity building effort, and a plan for overcoming these obstacles, would be needed in making the case for an ensuing Fund-supported program.

27

There may be cases where the authorities wish to move to a unified debt limit specified in present value terms during the program period; this preference could be accommodated in the context of a program review, once the requisite improvement in debt monitoring capacity has been achieved.

28

Low income countries are also more likely to face large infrastructure gaps, implying potentially high returns to public investment—another factor to be taken account of in the fiscal assessment.

29

Where external financing primarily takes the form of direct funding to the budget, the sectoral breakdown of the use of these funds would have little information value, and need not be provided.

30

Country authorities may wish to avoid publishing details on the specific terms obtained from individual creditors, out of concern that this could establish precedents for future loan discussions, again constraining scope to negotiate more favorable terms.

31

The adjustor would be capped in a manner that fully accommodates moderate ex post deviations from the assumed grant element of a loan but only partly accommodate large deviations from program assumptions.

32

The loan(s) involved would need to be large in scale for uncertainty regarding loan terms to translate into significant uncertainty regarding the projected present value of all new external debt to be contracted.

33

Public sector debt would typically refer to non-financial public sector debt.

34

The risk ratings discussed below are generated using the LIC DSF. There may be cases of countries that rely significantly on concessional financing where, for operational reasons; staff have considered that the MAC DSA is a more appropriate tool for analyzing debt vulnerabilities. In determining whether such countries face significant debt vulnerabilities, staff would seek to apply a similar standard as is used in the case of countries for which the LIC DSA is deployed; judgment would be called for in making this determination, given that the outputs of the DSAs differ significantly.

35

As noted in paragraph 50, the use of debt conditionality could still be warranted on grounds of data quality and/or the adequacy of coverage of fiscal data.

36

The assessment of debt monitoring capacity would be driven, but not mechanically determined by, the methodology outlined in Annex III; a final determination would be based on the judgment of Fund staff, formed in consultation with World Bank staff.

1

Loan concessionality information is available only on contracting basis and is not available on actual disbursement basis. This difference (contracting versus disbursement) should not cause any bias in the results.

2

Maldives did not have a risk rating in 2007, but its risk of external debt distress has deteriorated since its first rating in 2009.

3

In the case of Mozambique, large changes in the underlying balance of payments linked to commercial investment in the natural gas sector also contributed to the revision of the risk of debt distress.

4

For the purpose of this annex, public debt covers general government debt.

1

Please refer to the “IMF, Staff Guidance Note on Debt Limits in Fund-Supported Programs,” December 18, 2009.

2

As discussed above CPIA debt score takes into account not only the assessment of debt management practices, but also the extent of countries’ debt vulnerabilities. In some cases, the later component may actually overweigh higher score assigned the quality of debt monitoring. Therefore average scores falling slightly below 3 may still be consistent with adequate debt record systems.

1

Blend and gap countries receive only IDA credit and are not eligible for grants. Blend countries are countries that are IDA-eligible but also creditworthy for some IBRD borrowing. Gap countries are IDA countries with Gross National Income per capita above the operational cutoff for more than two consecutive years.

2

The volume of the latter is reduced by 20 percent to address moral hazard issues and to help finance foregone charge income on grants.

3

What would be different from the current situation is that, in many cases, these agreed targets will not be quantitative performance criteria in the Fund program.

Same Series

Other IMF Content

Other Publishers