Guidance Note On Implementing The Debt Limits Policy In Fund Supported Programs

The Debt Limits Policy (DLP) establishes the framework for using quantitative conditionality to address debt vulnerabilities in IMF-supported programs. In October 2020, the Executive Board approved reforms to the DLP which will enter into effect on June 30, 2021. The risk-based approach to setting debt conditionality informed by Debt Sustainability Analyses under the previous DLP approved in 2014 is maintained. The reforms aim to provide countries with more financing flexibility in practice while still adequately containing debt vulnerabilities through appropriate safeguards. This note provides operational and technical guidance related to the implementation of the DLP, including the operationalization of the approved reforms. In particular, it outlines the core principles underpinning the DLP, including when debt conditionality in IMF-supported programs is warranted and how to account for country-specific circumstances in the design of debt limits. The note also describes the process of setting and implementing debt conditionality, including: (i) identifying debt vulnerabilities to inform the focus of debt conditionality; (ii) designing debt conditionality; and (iii) implementing debt conditionality through the review cycle. The Guidance Note is intended for use by both IMF staff and country officials. In this regard, in addition to the guidance presented in the main body, the note also contains several annexes that cover definitional, technical, and operational issues arising in the determination and implementation of public debt limits.

Abstract

The Debt Limits Policy (DLP) establishes the framework for using quantitative conditionality to address debt vulnerabilities in IMF-supported programs. In October 2020, the Executive Board approved reforms to the DLP which will enter into effect on June 30, 2021. The risk-based approach to setting debt conditionality informed by Debt Sustainability Analyses under the previous DLP approved in 2014 is maintained. The reforms aim to provide countries with more financing flexibility in practice while still adequately containing debt vulnerabilities through appropriate safeguards. This note provides operational and technical guidance related to the implementation of the DLP, including the operationalization of the approved reforms. In particular, it outlines the core principles underpinning the DLP, including when debt conditionality in IMF-supported programs is warranted and how to account for country-specific circumstances in the design of debt limits. The note also describes the process of setting and implementing debt conditionality, including: (i) identifying debt vulnerabilities to inform the focus of debt conditionality; (ii) designing debt conditionality; and (iii) implementing debt conditionality through the review cycle. The Guidance Note is intended for use by both IMF staff and country officials. In this regard, in addition to the guidance presented in the main body, the note also contains several annexes that cover definitional, technical, and operational issues arising in the determination and implementation of public debt limits.

Section I. What Is The Debt Limits Policy?

1. The Debt Limits Policy (DLP) establishes the framework for using quantitative conditionality to address public debt vulnerabilities in IMF-supported programs. The policy dates back to the 1960s when the initial rationale for the inclusion of performance criteria (PCs) on external borrowing was to ensure that adjustment objectives, particularly the restraint of domestic demand, were not threatened by unforeseen foreign borrowing. Over time, the policy has evolved to help address debt vulnerabilities.1

2. Debt conditionality is one tool in the IMF’s toolkit to address debt vulnerabilities, intended to complement other conditionality and policies (Figure 1).

  • Fiscal conditionality is generally the most powerful tool to address public debt vulnerabilities, and with a sufficiently broad measure of the fiscal balance it can fully capture the net flow of new borrowing and thus debt accumulation. When country circumstances are such that fiscal conditionality alone cannot contain debt vulnerabilities—for instance due to narrow fiscal data coverage or because debt composition matters (e.g., excessive short-term or foreign currency debt)—debt conditionality may be used as a complement to provide additional safeguards.

  • The design and calibration of debt conditionality is informed by the identification of debt vulnerabilities in the Debt Sustainability Analysis (DSA), which requires adequate disclosure of debt data. The specification of debt conditionality also depends on the quality of debt management. It is thus appropriate that debt conditionality also strengthens debt transparency and encourages improvements in debt management.

  • Strengthening debt transparency and debt management in low-income countries (LICs) typically requires a sustained effort, addressed mainly by capacity development (CD) support. The Joint IMF-WB Multipronged Approach to Address Debt Vulnerabilities sets out a strategy aimed at supporting improvements in debt transparency and debt management over time.2

Figure 1.
Figure 1.

Debt Limits Policy and Other Policies and Tools

Citation: Policy Papers 2021, 037; 10.5089/9781513583617.007.A001

3. The DLP aligns in key areas with the World Bank’s Sustainable Development Finance Policy (SDFP) and it serves as an important reference framework for lending decisions by other creditors, including other Multilateral Development Banks (MDBs) and Export Credit Agencies.3

  • Both the DLP and the SDFP aim to help countries contain debt vulnerabilities while providing incentives for countries with access to concessional financing to seek it (see Section V for further details on linkages between the DLP and the SDFP). Although the two policies are closely linked, there are differences in how the two policies work. The DLP applies to all IMF members with Fund-supported programs, whereas the SDFP applies to all IDA-eligible countries.

  • The DLP (together with the SDFP) is a key anchor for implementing the recommendations of the OECD’s guidelines for lending to lower-income countries.4 Specifically, the decision by an OECD member country to provide financial support to a lower-income country’s public sector should be consistent with the prevailing limits set under the DLP. Official export creditors planning to provide credits to countries subject to limits on non-concessional borrowing (NCB) are called on to inform the IMF and the World Bank about such financing. The OECD also requires consistency with the IMF and the World Bank debt limits as a condition for treating a loan as a form of Official Development Assistance (ODA).5

  • Consistency between official lending practices and the debt limits set under the DLP is also an important component of the G20’s Operational Guidelines for Sustainable Financing.6 Commitment to observe the debt limits set under the DLP and the SDFP is also a condition for participation in the G20/Paris Club Debt Service Suspension Initiative (DSSI).7

4. The 2020 Review of the DLP introduced reforms that respond to changes in the credit landscape by providing countries more financing flexibility while still adequately containing debt vulnerabilities.8 The 2020 DLP Review maintains the same risk-based approach of the previous DLP, according to which conditionality is set according to debt vulnerabilities as identified through a Debt Sustainability Analysis. Annex I provides an overview of the reforms, which aim to: (i) better encourage adequate debt data disclosure to the IMF; (ii) allow for greater tailoring of debt conditionality for countries normally relying on concessional financing that have also been accessing international financial markets on a significant scale; (iii) encourage the broader use of debt conditionality in present value terms, which provides greater flexibility to countries on the mix of borrowing terms; (iv) facilitate the utilization of the existing policy for accommodating non-concessional borrowing (subject to safeguards); and (iv) clarify the definition and measurement of concessional debt.

5. This Guidance Note provides practical guidance on the DLP framework and explains how to design, implement, and monitor debt conditionality. Section II outlines the core principles underpinning the DLP. Subsequent sections take the user, step-by-step, though setting and implementing debt conditionality:

  • Section III describes the process of identifying debt vulnerabilities to inform debt conditionality, including: how to determine whether there is a comprehensive picture of debt; measures to improve debt information if needed; and utilizing DSAs to assess vulnerabilities.

  • Section IV explains the process of designing debt conditionality, including: debt conditionality requirements for countries that normally rely on concessional external financing, which are determined by external debt distress risk ratings in the LIC-DSF; conditionality design for countries that do not normally rely on concessional external financing, where the Market-Access Country (MAC DSA) informs tailored conditionality, until it is replaced by the Sovereign Risk and Debt Sustainability Framework (SRDSF);9 and how to design limits targeted at specific vulnerabilities, when necessary.

  • Section V focuses on the implementation of debt limits, including: ensuring and assessing compliance with debt conditionality (including how to assess the concessionality of loans); considerations for modifying debt conditionality when needed (including for exceptions to NCB limits); and coordination with the World Bank.

Section II. Principles Of Debt Conditionality

A. When are Public Debt Limits Appropriate?

6. The Fund’s Guidelines on Public Debt Conditionality in Fund Arrangements (the “Guidelines” set forth when public debt conditionality is warranted.10 In general, the use of debt conditionality is justified under either of the following conditions:

i. When the quality and coverage of national fiscal statistics favor the use of debt conditionality instead of, or as a complement to, “above-the-line” fiscal conditionality.

  • There could be merits in using a public debt limit as a complement to fiscal budgetary targets in cases where: (i) important public debt-creating activities are not adequately captured in the fiscal accounts (e.g., bank recapitalization, issuances of government guarantees, materialization of contingent liabilities from noncommercial SOEs and other off-budget entities); (ii) these activities pose a risk to the overall fiscal position; and (iii) addressing them is assessed to be of critical importance for achieving the goals of the member’s program or for monitoring the implementation of the program, in accordance with the Guidelines on Conditionality.11 12 Section IV.D discusses some of the operational considerations on the coverage of debt limits, including consideration of capacity constraints and the treatment of liabilities of the central bank or other public institutions, IMF financing, and commercially-viable SOEs.

  • There would be merits in setting quantitative conditionality on measures of 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. The circumstances justifying limits on debt accumulation as a “below-the-line” approach to fiscal conditionality are long-established and will not be discussed further in this Guidance Note.13

ii. When a country has significant debt vulnerabilities that cannot be tackled with fiscal conditionality alone. Since debt sustainability depends not only on the level and expected trajectory of debt, but also on certain features and terms of debt (e.g., maturity, interest rate/concessionality, currency composition, and collateralization), debt conditionality may be needed to ensure new debt has certain financing characteristics. For example, limiting borrowing on non-concessional terms can help limit the debt and debt service burden; limiting short-term debt can help ensure debt has longer maturities to reduce rollover risks). The complementarity between debt limits and fiscal conditionality is discussed further in Section III.C.

B. Key Differences Across Countries in Debt Conditionality Specification

7. The specification of debt conditionality needs to reflect country-specific circumstances appropriately. Relevant factors include the composition of public sector financing, the objectives of the Fund-supported program, the extent and type of debt vulnerabilities, the quality and timeliness of the financial information produced by the country’s public sector accounting system, and other macroeconomic circumstances of member countries. Depending on the circumstances, and in accordance with the Guidelines on Debt limits, debt limits could be set on total public debt or on public external debt; be specified in nominal or in present value (PV) terms; and cover debt stocks or new flows.

8. The country’s reliance on concessional external financing is a key factor driving different DLP conditionality requirements. Some external debt limits focus on encouraging more reliance on concessional borrowing, and it is important that such limits apply only when relevant and feasible for a country given the financing that it can access. As outlined in Section IV, there are different requirements for countries that normally rely on concessional external financing and those that do not. “Normally rely on” is taken to mean countries for which concessional financing has long been, and continues to be, a key source of public external financing, if in not every particular year. For operational purposes, all countries using the LIC-DSF are considered as countries normally reliant on concessional external financing.14 Countries that do not normally rely on concessional financing are defined as those that use the MAC DSA (or the SRDSF when effective). These countries often do not have access to concessional financing, and fiscal accounts tend to be more comprehensive, typically limiting the need to adopt debt conditionality to those countries with specific vulnerabilities such as in the composition of debt or contingent liabilities.

9. Among countries that normally rely on concessional external financing, DLP conditionality requirements also depend on whether they have access to international financial markets on a significant scale. As the credit landscape has evolved in recent years, some countries that still normally rely on concessional external finance have also started to access international financial markets on a significant scale. Typically, countries in this subgroup envisage continued tapping of these markets during IMF-supported programs, which may be inconsistent with the application of NCB limits. To address this tension, the 2020 Reform of the DLP introduced different conditionality requirements for countries that normally rely on concessional financing but have access to international financial markets (see Section IV.B). Yet, given the need for stability in conditionality and to appropriately manage risks, this market access should be significant rather than small and sporadic. Being considered as a country that has significant access to international financial markets requires meeting well-defined criteria, described in Box 1.

Identifying Countries with Significant Access to International Financial Markets

Determining whether a country that normally relies on concessional external financing has significant access to international financial markets depends on whether it has a significant amount of international financial market borrowing and a demonstrated capacity to manage significant levels of market borrowing, In particular, the criteria include:

1. Significant amount of international financial market borrowing. The country has had (a) significant access to international financial markets in recent years or (b) access to these markets is a key element of the Fund-supported program, as elaborated below:

  • a. Cumulative market borrowing of at least 50 percent of quota in at least two of the preceding three years (based on the latest available assessment of the Eligibility to Use the Fund’s Facilities for Concessional Financing);1 or

  • b. Access to these markets is an integral component of the country’s annual borrowing plan or of the baseline projections of the Medium-Term Debt Strategy (MTDS), provided that the Fund-supported program envisages financing on international financial markets in an amount equivalent to at least 100 percent of quota over the course of the program period.2 This threshold only applies in cases where the country does not meet Criterion 1a. Criterion 1b would ensure that even if a country had a temporary interruption in access to international financial markets (e.g., due to global developments), it could still be considered as having significant access, provided that there are credible expectations for such borrowing over the course of the program.

AND

2. Demonstrated capacity to manage significant levels of market borrowing. Capacity could be demonstrated by:

  • (i) the existence of a recent and credible Medium-Term Debt Strategy; and

  • (ii) an actively updated annual borrowing plan in line with international best practices.

1 Based on data used for the most recent annual assessment for Eligibility to Use the Fund’s Facilities for Concessional Financing. See for example Eligibility to Use the Fund’s Facilities for Concessional Financing, 2020. Note that, while similar, the DLP test differs from the PRGT-blending criteria in two key ways: it does not include an income criterion and it has a shorter horizon (three years) to reflect a country’s more recent financing circumstances.2 Given the assessment to meet this criterion would be based on projections, there would be some flexibility to also consider countries that are just under this threshold by a small margin.

C. Overview of Designing Debt Conditionality

10. Figure 2 outlines the process for designing debt conditionality, which is elaborated on in Sections III and IV:

  • The first step is to determine whether the available debt data provide a comprehensive and updated picture of debt in order to identify key vulnerabilities (Section III.A). In the event there is not a comprehensive or updated picture of debt (for instance due to capacity or institutional limitations), additional measures to improve debt information like technical assistance or conditionality (e.g. prior actions or structural benchmarks) may be implemented to improve debt information while following a risk-based approach (Section III.B).

  • A second step is to understand the coverage of fiscal conditionality and whether it can address all key debt vulnerabilities (Section III.C).

  • A third step is to identify the nature and extent of debt vulnerabilities (Sections III.D-IV). The LIC-DSF and MAC DSA (SRDSF when effective) serve as key tools in this regard as they can inform whether conditionality is required and the type of conditionality that may be employed (e.g. total public debt limits, PV or NCB limits, targeted debt limits, etc.).

  • Finally, calibrating quantitative debt limits involves an iterative process, which considers the authorities’ annual borrowing plan (and, by extension, its macroeconomic framework)15 and its MTDS (Section IV). This calibration must ensure that additional borrowing does not move a country into an unsustainable debt situation or lead to an unacceptable worsening of debt vulnerabilities. For countries classified as having a high risk of debt distress this would generally mean avoiding a rise in risk (as informed by indicators in the DSA). For countries with a low/moderate risk classification this would generally mean avoiding a move into or near a high risk of debt distress. A natural starting point for discussing debt conditionality with the authorities is the annual borrowing plan, staff’s macroeconomic framework, and the MTDS (if available). Ultimately, debt conditionality should be consistent with the financing assumptions in the macroeconomic framework.

Figure 2.
Figure 2.

Designing Debt Limits Flow Chart

Citation: Policy Papers 2021, 037; 10.5089/9781513583617.007.A001

Section III. Identifying Debt Vulnerabilities

A. Determining Whether There is a Comprehensive Picture of Debt

11. Establishing a comprehensive view of public debt vulnerabilities is an essential first step in designing debt conditionality, as explained in Section II. Ensuring that debt conditionality appropriately complements fiscal conditionality requires sufficiently detailed information about the magnitude, structure, and composition of debt (both disbursed and undisbursed). It requires sufficiently broad debt coverage without gaps, including in terms of the institutional coverage (i.e. which subsectors of the public sector are covered) and instrument coverage (loans, debt securities, other accounts payable).16 In addition, to identify and address specific debt vulnerabilities, data on key terms and conditions of the contracted debt must be available. Nevertheless, pursuing debt information can be costly and time consuming both for the country authorities and the IMF staff. Hence, rather than pursuing debt disclosure on all aspects of debt and for every entity of the public sector, staff should use a risk-based approach and focus their efforts on areas where information on debt is expected to sufficiently enhance program design. For example, this implies that detailed information on terms and conditions could be limited to those instruments that constitute the bulk of public debt and to those instruments whose characteristics pose a specific vulnerability (e.g. certain collateralized debt).

12. A risk-based approach to assessing whether the debt picture is comprehensive can be informed by a range of information, including a number of “red flags”. Determining the presence of omitted debt, or of vulnerability-enhancing terms and conditions, may not be straightforward. In these cases, some “red flags” can provide signals for when it is critical to pursue debt data disclosure, including:

  • Indications of significant omitted debt, such as:

    • (i) Extra-budgetary central government units, and state and/or local governments not already captured in debt numbers that have the capacity to borrow (and indications that they have used this capacity);

    • (ii) Large state-owned enterprises (SOEs) with capacity to borrow not already captured in the debt perimeter (and indications that they have used this capacity) and; Public-private partnerships (PPPs) or special purpose vehicles (SPVs) that are not adequately evaluated or monitored;

    • (iii) Large, unusual, or growing liabilities not included in debt statistics, such as accounts payable, which will ultimately have to be paid (e.g., unpaid wages and salaries, unpaid VAT refunds, unpaid taxes or social contributions from SOEs, overdue trade credits, etc.).

  • Signs of “hidden debt” that has not been disclosed due to confidentiality clauses in loan contacts. Staff may inquire whether the scope of confidentiality conditions could allow them to review the loan agreements and other documents (i.e. there may be clauses that are not subject to confidentiality provisions), as well as whether domestic law allows for the disclosure of such information. If there is any significant inconsistency between the amount and terms of loans reported by the authorities and those credibly reported by third parties (e.g., in academic or commercial debt databases, in the news media reports), country teams should request this information from the authorities, while ensuring that staff complies with the Fund’s policies on treatment of confidential information.

  • Missing information about the terms and conditions of a significant amount of debt or of significant contingent fiscal liabilities (such as PPPs or SPVs), combined with indications that vulnerability-inducing or vulnerability-enhancing terms and conditions could exist. Examples include the presence of unrelated collateral, margin call clauses in repurchase agreements,17 step-up rates, non-standard event-of-default clauses (including failure to deliver an agreed amount of a commodity as part of a tied purchase and sale agreement).

  • Debt sustainability residuals. One readily-available indicator that can help teams in assessing the potential for debt vulnerabilities stemming from omitted debt is the magnitude of the debt residual in both the LIC-DSF and the MAC DSA (SRDSF when effective).18 A large historical residual or unexpected change in debt is a red flag for the potential presence of omitted debt and significant risks from contingent liabilities (the LIC-DSF output contains a graph showing where each country is located in the distribution of unexpected changes in debt among LICs).

13. When there are indications of significant omitted debt, the DSA should inform the Board about the extent to which there are gaps in public debt coverage (Box 2). Information on debt coverage and contingent liabilities is already reported in the LIC-DSF, and the MAC DSA (SRDSF when effective) follows a similar approach. Red flags for seeking more information include: (i) when the DSA coverage falls short of recommended levels; and (ii) contingent liabilities (excluding the financial sector) entail significant debt-related risks, as revealed by the results of the contingent liability stress tests. Where contingent liabilities cannot be quantified, an alternative signal could be the existence of large unaudited SOEs or other relevant non-financial SOEs.19 Additional information sources that staff may consult for indications of potential missing components of debt include: Bloomberg (for data on SOEs and local government bond issuances) and the Perfect Information database of sovereign bond prospectuses; Dealogic or similar databases (for data on bonds or syndicate loans); information from the authorities, technical assistance reports, the World Bank Debt Transparency Heat Map,20 and other studies from the World Bank and other agencies (for data on PPPs and SPVs). Country teams may also seek assistance from SPR/LEG/STA in utilizing such sources.

How to Identify Debt Data Gaps in the DSA?

Full disclosure of the scope of public debt coverage is expected under DSAs. Both the LIC-DSF and MAC DSA (SRDSF when effective) require that the coverage of public debt be comprehensive, covering all parts of the public sector and contingent liabilities, in particular, government guarantees. Notwithstanding, there might be cases where headline debt indicators fall short of fully capturing debt vulnerabilities. Country teams are expected to discuss the debt coverage used for the analysis and to identify omitted components upfront. Debt outside of the perimeter does not imply that it is unknown to the country teams or that it cannot be quantified, and several country teams have enlarged the perimeter of debt over time as more information has become available. Debt vulnerabilities potentially arising outside of the perimeter of debt data should thus also be identified and quantified, both in terms of gaps in coverage of public sector debt and/or the omission of debt(-like) instruments (e.g., repos, swaps, PPPs). For those countries with annual Fiscal Risk Statements prepared by the authorities, the information contained can be adequate for the monitoring and evaluation of risks as long as their coverage includes those public sector entities (SPVs and PPPs) posing the greatest fiscal risks.

The magnitude of contingent liabilities should be assessed under a stress test. The Fund/Bank DSAs require customizing the contingent liability stress test to reflect country-specific debt vulnerabilities arising outside the baseline debt projections, including from SOEs (when not covered in the headline debt numbers) and PPPs. When the stress test flags a large deviation from the baseline debt projections and accurate information is not readily available, staff should seek fuller disclosure of information on omitted sectors.

The following illustrates cases where upfront disclosure or an improvement in debt data disclosure (in case of capacity limitations) is required:

Extra-budgetary central government units, and state and/or local governments that have large debt outstanding (if not included the debt data) with or without central government guarantee.

Large debt collateralized by assets or future revenues (e.g., by export receipts generated from natural resources) that are unrelated to the projects a loan is intended to finance.1

Large SOEs (not covered explicitly in the debt numbers) with significant debt liabilities including those guaranteed by the general government. Large stock of PPPs that is not properly regulated or monitored, which could potentially be contingent liabilities of the general government.

1 For more examples of unrelated collateral and its differences with the more common form of related collateral, see Collateralized Transactions : Key Considerations for Public Lenders and Borrowers. Examples of countries with Fund-supported programs in which conditionality has been adopted to help disclose unrelated collateralized or collateral-like debt, monitor its evolution and, in those cases where it has been assessed to be macro critical, limit further issuance are Angola (IMF Country Report No. 18/370), Republic of Congo (IMF Country Report No. 19/244), and Ecuador (IMF Country Report No. 19/79).

14. In the case of missing information on the terms and conditions of a significant amount of debt, or of significant contingent fiscal liabilities, the following should raise red flags for seeking more information:

  • PPPs: In addition to being a potential source of omitted debt,21 the background analysis for the 2020 DLP reform shows that lenders and borrowers may circumvent NCB limits by having concessional terms in loans but offsetting the grant element by overpricing in PPPs.22 Thus, a private creditor providing budget and project financing on highly concessional terms to a government without clear motives is a red flag, requiring staff to exercise due diligence regarding the overall terms and conditions of such financing.

  • Collateralized debt: full information on collateralized debt can improve program design and help reduce fiscal vulnerabilities when the nature and extent of collateralized debt is assessed to be macro-critical.23 It is most common for loans to be collateralized in the form of assets or revenues which are directly part of the specific project being financed, which has limited implications for the resources available to service other debts. In contrast, the risks will be notably higher when significant debt is collateralized on assets or future revenues (e.g., by export receipts generated from natural resources) that are unrelated to the projects a loan is intended to finance.24 An example would be collateralized lending to finance the budget. In this case, conditionality could take the form of a prior action on the full disclosure of the terms and conditions of collateralized debt when there is a presumption of the existence of such debt which is not publicly available (see next section). Red flags to seek further information include:

    • the debtor country has natural resources (e.g. oil, gas and other minerals) whose revenues can serve as collateral;

    • the debtor country has contracted or owes a large amount of debt to creditors that are traditionally known to request such collateral (teams may wish to contact SPR for updated information).

For countries that have bonds listed in exchanges, staff may consult the bond prospectuses with support from SPR and LEG if needed, which generally contain detailed information on debt characteristics that could be used to identify specific vulnerabilities.

B. Steps to Improve Debt Information if Needed

15. The expectation under the DLP is that critical debt data disclosure gaps should be addressed up front in Fund-supported programs. A critical debt data disclosure gap is defined as a case when the missing information could result in a vulnerability that would have a critical impact on the implementation or monitoring of the Fund-supported program. This would need to be assessed on a case-by-case basis, with due attention to the potential size and nature of the problem. When a critical data gap is identified, teams have some options to help pursue further information, where the appropriate approach will depend both on capacity constraints and the likelihood of debt vulnerabilities materializing in the near term:

  • Prior action. In cases where upfront disclosure of missing data is critical to the successful implementation of the program (e.g. unavailability of information could cast some doubts over fiscal/debt sustainability or financing assurances), prior actions could be used to ensure that data gaps are addressed prior to approval of an arrangement or completion of a review, provided that they are in line with the Guidelines on Conditionality, including that the IMF should take into account the strain that they can place on the country’s implementation capacity.

  • Structural benchmarks. Additional debt-related information could also be obtained over the course of the IMF-supported program, using structural benchmarks where warranted (i.e. where capacity limitations prevent full disclosure of debt data upfront, and staff assesses that delayed disclosure would not critically impact program design or near-term performance).

  • Capacity development. When data disclosure gaps are due to capacity constraints, the authorities’ efforts to enhance debt data disclosure and transparency would need to be supported by technical assistance where appropriate, including in collaboration with the World Bank.25 This assessment would need to be based on staff’s evaluation of the country’s capacity needs and the CD strategy.

16. Conditionality on debt data disclosure to IMF staff would only be expected if the information potentially revealed by such disclosure is deemed critical for achieving the goals of the Fund program or for monitoring its implementation. This would require a case-by-case assessment, with due attention to the potential size and nature of the problem.

C. Requirement for a Debt Holder Profile

17. The 2020 DLP reform requires that a table with information on the country’s debt holder profile must be included in staff reports for all program requests and reviews. The debt holder profile provides information on a country’s reliance on international financial markets, domestic and foreign banks, as well as bilateral, plurilateral, and multilateral creditors. Such information is needed to better assess the nature and extent of debt vulnerabilities, such as rollover risks, and would therefore help to enhance both the focus and calibration of debt conditionality. For example, debt conditionality could be used to better contain risks posed by the frequent use by some creditors of non-standard lending instruments and terms (these may be used by both official and private creditors, and could include advanced sales of commodities, financing tied to commodity purchase and sale agreements, lending with unrelated collateral, and overcollateralized repurchase agreements).26 Another example could be debt conditionality designed to encourage diversification of financing sources if this is critical for the IMF-supported program to address the risks from a concentrated composition of private sector credit. Such information can also be an important input into understanding the pattern of burden-sharing across additional financing sources within an IMF-supported program.

18. The debt holder profile table to be included in IMF-supported program documents should include granular information on public debt and debt service composition by creditors and instruments as described in the template in Table 1. Key components include:

  • Debt stock coverage: the debt holder profile table should include a breakdown of the public debt stock into external and domestic and by subcategories of creditors/instruments, in US$, as percent of GDP, and as percent of total debt as of the end of the period for the most recently completed year.27 The coverage of external and domestic debt should, in principle, be as broad as the DSA. Nonetheless, if a debt holder breakdown is not available for a specific sector within the DSA coverage, e.g., SOEs, a narrower coverage can be used. This excluded sector should be identified in a footnote to the table and explained in the DSA/staff report when significant.

  • Creditor composition: External debt should be decomposed by group and sub-group of creditors (Multilateral,28 Bilateral (Paris Club, non-Paris Club), Bonds, Commercial, and Other International) with a list of the two largest creditors for each sub-group of creditors.29 The information should be presented as reported by country authorities and using their debt holder classifications (including official and commercial), and indicated as such in a footnote. Creditors (other than the two largest ones) which account for 5 percent or more of total debt should also be reported individually to ensure macroeconomically significant debt is reported. Disclosure of information on individual creditors beyond these requirements is encouraged but not required.

  • Debt service: the table is expected to include a breakdown of the debt service by creditor sub-group for at least three years (i.e. the previous, the current and the next year). Disclosure of more granular information on debt service by individual creditors is also encouraged but not required.

  • Collateralized debt: since collateralized debt can involve specific vulnerabilities that may need to be addressed as part of a Fund-supported program, the table includes information on the stock of such debt as a memo item. It is common practice for lending to be collateralized by the assets or revenues of the projected being financed, but is important to identify that part of debt where the asset or future revenue forming the collateral is unrelated to the loan, such as lending to the budget collateralized by commodity revenues, as this earmarking makes other loans riskier, potentially resulting in a deterioration in financing terms if unrelated collateral is significant.

  • Domestic debt: information on domestic debt is expected to include the breakdown of public debt by instruments (bonds, T-bills, loans). There should also be a breakdown between resident and non-resident holdings, if available (or otherwise marked in the table as n/a). If data is available by resident institutions (banks, non-bank financial institutions etc.) reporting is encouraged but not required.

Table 1.

Template for Table on Debt Holder Profile*

article image

As reported by Country authorities according to their classification of creditors. Including by official and commercial Debt coverage is the same as the DSA, except for[(missing entities if applicable).]

Some public debt is net shown in the table due to [confidentially clauses/capacity constraints]. (Include for all creditor groups where applicable)

Individual creditors accounting for more than 5 percent of total debt.

Debt is collateralized when the creditor has rights over an asset| or revenue stream that would allow it, if borrower defaults on its payment obligations, to rely on the asset or revenue stream to secure repayment of the debt. Collateralization entails a borrower granting liens over specific existing assets or future recevables to a lender as security against repayment of the loan. Collateral is "unrelated" when it has no relationship to a project financed by the loan. An example would be borrowing to finance the budget deficit, collateralizaed by oli revenue receipts. See the joint IMF-World Bank note for the G20 "Collateralizaed Transactions; Key Considerations for Public Lenders and Borrowers" for a discussion of issues raised by collateral.

Includes other-one off guarantees not included in publicly guaranted debt (e.g. credit lines) and other explicit contingent liabilities not elsewhere classified (e.g. potential legal claims, payements resulting from PPP arrangements). See 2014 Government Finance Statistics Manual (7.252) for more information.

If unavailabel mark as "n/a".

*This table should be included in the staff report for all program requests and reviews. T0 is the current year. Items shaded in Yellow are expected to be completed. When an item in the yellow area is missing, it should be footnoted indicating whether this is due to confidentiality clauses or capacity constraints. Missing elements of the table are expected to be completed no later than the second program review.

19. The publication of information included in the debt holder profile table must be consistent with the Fund’s policies on confidential information and the Transparency Policy. Some debts may subject to confidentiality clauses which limit the authorities’ disclosure of certain terms (e.g., interest rates) or have broader limitations on disclosure (including the existence of the debt itself). To the extent information on debt can be provided to Fund staff on a confidential basis, it would be subject to the Fund’s policies on handling confidential information (Box 3). Staff should therefore clarify upfront whether the authorities consent to disclosure—both to the Executive Board and/or the public—of the information to be reported in the debt holder profile table. If consent to disclosure is not provided upfront, including due to confidentiality clauses in the debt contracts, staff should explain the consequences to the authorities of failure to consent to such disclosure (e.g., see Box 3, bullet 1) and encourage the authorities to work with the creditor as needed to permit disclosure.30 If, despite these requests, the authorities do not consent to disclosure of the information, staff may not report these data in the table in the staff report.31 When an element of public debt data is missing, the table should add a footnote explaining the reasons for the missing information, including whether it is due to confidentiality clauses.32 Where confidentiality clauses exist, staff are encouraged to include an estimate of the amount of debt subject to confidentiality clauses in the discussion of debt data in the DSA,33 or a qualitative assessment of the prevalence of such clauses if an estimate is not feasible. Separately, the authorities may choose to share certain debt information on a confidential basis with the Executive Board and not the public. In such cases, the information shall not be included in the staff report but may be shared on a confidential basis with Executive Directors.

20. Missing elements in the debt holder profile table are expected to be filled in by the time of the second program review.34 If an issue of capacity, technical assistance could be mobilized if needed to support the authorities’ efforts to enhance debt data disclosure. Staff reports should clearly document the reasons for any missing items in the table. If there are factors that delay the support needed, these should be explained in the staff report for the review. For example, during the COVID-19 pandemic, these could include supply and logistical constraints to providing in-country training and technical assistance missions, which cannot be effectively provided virtually. It may also be the case that although staff has encouraged the authorities to work with the creditor as needed to permit disclosure of such information, the creditor has not consented to such disclosure.

Handling Confidential Information Relating to Debt1

When reviewing debt documentation, IMF staff may receive information subject to confidentiality clauses or non-disclosure agreements (NDAs). Staff should always seek to clarify upfront the conditions under which such information is being provided.

• Such information would need to be shared with the Executive Board if it is considered necessary for the Board to make a decision with respect to the provision of IMF financing. If the authorities do not consent to disclosure to the Executive Board in such a case, staff and management would not be able to recommend approval of the arrangement or completion of the review.

• Any disclosure of information to the public requires the consent of the information provider. Staff shall not include in the staff report any information provided on the understanding that it will remain confidential.

Thus, there could be elements in the debt holder profile table that are not completed if either:

  • (i) a country cannot provide such information (e.g. due to capacity issues); or

  • (ii) a country can only provide such information to staff and/or the Executive Board on a confidential basis and the information provider does not consent to its disclosure more broadly.

In terms of the practical ways to ensure proper handling of confidential information, in previous program negotiations, country authorities have made available the detailed debt information (e.g., multi-party agreement, loan agreement, purchase and sale agreement, etc.) through the offices of their Executive Director offices, who have allowed Fund staff to review them under “data room” rules. Alternatively, some country authorities have made the information available through their legal advisors, some of which have representative offices in Washington DC or through video conferencing, who have also responded to Fund staff questions. LEG and SPR have assisted country teams in reviewing the detailed debt documentation and in the interactions with the authorities’ legal advisors.

1 For a general description of the treatment of confidential information, see Updated Guidance Note on the Fund’s Transparency Policy, April 2014, Appendix IX.

D. Understanding Complementarity with Fiscal Conditionality

20. In determining the appropriate deployment of debt limits, staff must have a clear understanding of the coverage of fiscal conditionality, particularly to determine whether it alone can address any identified macro-critical debt vulnerabilities. In line with the Guidelines on Conditionality,35 program conditionality should be parsimonious and there is a need to avoid duplication between fiscal and debt conditionality. Fiscal conditionality is generally included in all programs and is central to containing debt vulnerabilities (see Section II.A). Determining whether there is also a need for debt conditionality typically requires more judgement in countries that do not normally rely on concessional financing where the coverage of the fiscal conditionality may be more comprehensive. For example, when the coverage of fiscal conditionality is comprehensive, encompassing both domestic and external financing and all relevant entities that may generate debt vulnerabilities (e.g. general government, SOEs), a standard fiscal PC may suffice to safeguard debt sustainability. For countries that normally rely on concessional financing, fiscal coverage may be narrower than debt coverage, and some external financing may be excluded from the fiscal PC. This group will have certain requirements for debt limits determined by the LIC-DSF rating, although additional debt conditionality may also be deployed depending on country-specific risks.

21. Even when fiscal conditionality plays a central role in addressing debt vulnerabilities, there are several ways in which debt conditionality can serve as a complement:

  • First, debt can arise from outside the targeted fiscal sectors (e.g. guarantees and non-guaranteed SOE debt) and, accordingly, debt conditionality may perform a broader fiscal control function.

  • Second, certain forms of debt or liabilities create specific vulnerabilities that cannot be addressed by fiscal conditionality alone, such as a high share of unrelated collateralized debt.

  • Third, debt vulnerabilities could arise from the structure of debt, such as a preponderance of short-term claims or foreign-currency denominated debt. This complementary role of debt limits also depends on strengthening debt transparency to increase information about the extent of public sector liabilities and improve our understanding of debt vulnerabilities.

  • Finally, debt conditionality set on a contracting basis can capture vulnerabilities associated with the accumulation of debt at a different point in time than fiscal conditionality, which captures debt when it is disbursed (conditionality is typically established on debt contracting rather than debt disbursements when debt monitoring is weak).

E. Utilizing Debt Sustainability Analyses to Assess Debt Vulnerabilities

23. The main tools for assessing the extent and nature of debt vulnerabilities are the LIC-DSF for countries that normally rely on concessional financing and the MAC DSA (SRDSF when effective) for countries that do not normally rely on concessional financing.36

  • LICs: For countries in this group, the assessment of debt vulnerabilities is principally informed by the risk of external debt distress and, where relevant, the overall risk of debt distress. An external risk rating of moderate, high or in debt distress would signal the presence of significant external debt vulnerabilities. The extent of debt vulnerabilities related to domestic debt will be determined by the analysis of the public DSA and reflected in the overall risk of debt distress, as contained in the conclusions of the LIC-DSF.

  • MACs: For these countries, the MAC DSA is due to be replaced as the main tool for identifying the extent of debt vulnerabilities in 2020Q4–2021Q1. Once the SRDSF becomes effective, it will provide a bottom line assessment regarding the country’s risk of developing sovereign stress (distinguishing between “low”, “moderate”, and “high” risk) broadly like the LIC-DSF.37 The framework also provides horizon-specific signals (distinguishing between the short-, medium-, and long-run). As the distinction between domestic and external debt is typically less relevant to MACs, the SRDSF focuses on the total debt burden.

24. In addition to the extent of debt vulnerabilities, it is important for debt conditionality to be tailored to the nature of the debt vulnerabilities, which fall into three categories: i) risks related to the structure of debt; ii) risks stemming from high debt levels; and iii) risks emanating from high debt service obligations. In all cases, risks can be concentrated domestically, externally, or be present along both dimensions. Any resulting debt conditionality should be parsimonious and tailored to the source of the risk.

25. Table 2 provides an illustrative guide as to what types of debt conditionality can be designed to address specific debt vulnerabilities, which will be discussed in more detail in Section IV.38 The intensity of debt conditionality is expected to be proportional to the identified risks to debt sustainability, while reflecting a desire for parsimony along with the aim of limiting negative spillovers to other elements of the program and macroeconomic stability in general.

  • When vulnerabilities are due to a high level of domestic debt, the onus of safeguarding debt sustainability normally lies with fiscal conditionality, which limits domestic borrowing (among other things)—making it the most direct way of controlling domestic debt accumulation.

  • When vulnerabilities are stemming from a high level of external debt, debt conditionality can fulfil an important role—typically by limiting external borrowing. This lever may be less relevant for MACs (or LICs that are well-integrated in international capital markets) for whom the distinction between domestic and external borrowing is less material (due to significant participation by foreign investors in domestic debt markets); in such cases, the onus of safeguarding debt sustainability again lies with fiscal conditionality.

  • When vulnerabilities are related to high debt service obligations, conditionality could focus on limiting the amount of new borrowing along the vulnerable dimension (domestic or external) and on encouraging concessional borrowing which has lower debt service.

  • When dealing with risks originating from the structure of debt, tailored forms of debt conditionality can be invoked to address the specific vulnerability, as detailed in Table 2.

  • Additional conditionality may be in order, depending on whether risks are stemming from a high interest burden or from a heavy amortization profile (e.g., a bunching of maturities).

  • When vulnerabilities are stemming from a high interest burden, conditionality may be designed to encourage the substitution of domestic borrowing with external borrowing, provided that the latter is thought to give a cost advantage that is not offset by higher exchange rate risk. When designing a balanced financing mix between external and domestic borrowing, teams could incorporate the results from the latest External Sector Assessment (for example, countries with an external balance which is substantially weaker than warranted by fundamentals and desirable policies should in principle avoid increasing the share of foreign-currency denominated debt). A MTDS (if available), might also provide information on the cost and risk tradeoff of an alternative financing strategy and an anchor for the desired shift in creditor composition. For countries with significant access to international financial markets where domestic and external borrowing are close substitutes (typically the case for most MACs and some frontier LICs), fiscal conditionality may need to play a bigger role (ensuring that the elevated interest burden is matched with a stronger primary fiscal balance).

  • When vulnerabilities are stemming from a heavy amortization profile, conditionality could focus on lengthening maturities through new debt issues or restricting the issuance of debt that is set to mature in years which are already characterized by large repayment obligations. Such measures can be supported by conditionality calling for specific debt management operations if there is a need to reduce/smooth financing pressures in specific years.

Table 2.

Illustrative Forms of Debt Conditionality in Response to Different Vulnerabilities

article image

Section IV. Debt Conditionality Design

A. Tailoring Conditionality to Debt Vulnerabilities

26. Debt conditionality design should reflect the nature and extent of debt vulnerabilities identified in the LIC-DSF or SRDSF/MAC-DSA while taking account of country-specific circumstances (particularly financing), as discussed in Sections II and III.D. The DSA structure differs according to whether the country belongs to the LIC or MAC category, but both frameworks produce detailed information on core debt-related metrics to inform assessments of the precise nature and extent of debt vulnerabilities. In addition, the DSA can be used to determine a country’s space to borrow under the DLP—the general guiding principle being that exhaustion of a borrowing limit should not put a country near or at high risk of debt distress, or, for countries already at high risk or in debt distress, allow a further rise in risk by the end of the program (informed by a worsening of indicators in the DSA). The buffer that is needed between the high-risk thresholds and the projected debt burden indicators which are at risk of being breached will depend on the volatility of each economy. This calibration of limits typically requires an iterative process between the DSA and the underlying macroeconomic framework (including the financing assumptions), with the latter assuming borrowing takes place up to the ceiling of any preliminary debt limits.

27. Table 3 summarizes conditionality requirements under the DLP based on risks identified in the DSA and the country’s financing circumstances. Guidance on how the LIC-DSF (for countries that normally rely on concessional financing) and the MAC DSA/SRDSF (for countries that do not normally rely on concessional financing) can be deployed to determine the need for, and appropriate nature and calibration of debt conditionality, are provided in sections B and C respectively. Specific operational considerations for designing debt conditionality are discussed in section D.

Table 3.

Debt Limits Policy Conditionality Requirements

article image

Where the country team determines that there is weak capacity to monitor the incurring of all forms of debt, nominal NCB limits for moderate risk countries (or a memo item on CB for high risk countries) would apply, supported by a more focused capacity building effort.

Limits can be set on the basis of the currency of debt denomination if accurate high-frequency data on external borrowing is not available because of foreign investors moving in and out of domestic instruments, as well as between domestic-currency and foreign-currency bond issues.

The process for granting such exceptions is described in Section V.C.

B. Countries that Normally Rely on Concessional Financing

28. For countries that normally rely on concessional financing, the LIC-DSF provides information on the three forms of debt vulnerabilities laid out in Section III.D.39

  • Risks stemming from high debt levels are captured by comparing projections for the PV of debt (under the baseline and under the relevant stress scenarios) to the appropriate high-risk thresholds. Figure 1 of the LIC-DSF does so for external public debt (which is scaled by GDP as well as by exports; see Box 4 for example of figures), while Figure 2 does so for total public debt (which is scaled by GDP as well as by government revenues). Relevant stress scenarios could include, for example, shocks to commodity prices and their impact on export values, government revenues, and GDP.

  • Risks stemming from high debt service obligations can be assessed by comparing projections for debt service (under the baseline and under the relevant stress scenarios) to the appropriate high-risk thresholds. Figure 1 of the LIC-DSF does so for external public debt service obligations (which are scaled by exports as well as by government revenues), while Figure 2 analyzes total public debt service obligations (which are scaled by government revenues). When making the final risk of debt distress assessment, any liquid asset buffers held by the sovereign should be considered as a mitigating factor, while the country’s vulnerability to sudden stop-type events should also be taken into account (e.g., the institutional setup).

  • Risks stemming from debt structure are conveyed through the modules dealing with debt coverage and the currency composition of debt. Details on the maturity structure and possible collateralization arrangements are not part of the DSF-template, but the relevant information should be obtained from the authorities on a risk-based approach (potentially subject to confidentiality arrangements) as discussed in Section IIIA

Example Figures for LIC-DSF

Figures 1 and 2 in the LIC-DSF show the risks stemming from high debt levels as well as risks stemming from high debt service obligations.

LIC-DSF Figure 1.
LIC-DSF Figure 1.

Indicators of Public and Publicly Guaranteed External Debt under Alternative Scenarios

Citation: Policy Papers 2021, 037; 10.5089/9781513583617.007.A001

LIC-DSF Figure 2.
LIC-DSF Figure 2.

Indicators of Public Debt Under Alternative Scenarios

Citation: Policy Papers 2021, 037; 10.5089/9781513583617.007.A001

29. The external debt limits required for countries in this group depend on their external risk rating and whether they have significant access to international financial markets (see Section II.B) as described in sections (i) and (ii) below.

(i) Countries that Normally Rely on Concessional Financing without Significant Access to International Financial Markets

In practice, the configuration of the DLP for this group will depend on the country’s assessed risk of external debt distress, where a higher assessed risk typically leads to tighter debt conditionality.

30. Low risk of external debt distress.

  • For countries in this group, the use of debt conditionality is not required unless specific debt vulnerabilities warrant targeted conditionality. Provided that the fiscal conditionality has comprehensive coverage, there will typically be no need to offset the absence of a debt ceiling with new conditionality in other areas. Nonetheless, in cases where debt levels are projected to increase significantly, even while the risk of debt distress remains low, the staff report should discuss the underlying reasons for the debt build-up and how related risks are being addressed under the program.

  • A macroeconomic program that is projected to result in an increase in the risk of debt distress from low to moderate would not be precluded. However, the potential benefits of such borrowing should be carefully weighed against the cost of downgrading the risk of debt distress and increased debt vulnerabilities. Furthermore, the debt conditionality relevant for countries at moderate risk of external debt distress would apply in such instances (see below).

31. Moderate risk of external debt distress:

  • For countries in this group, in most cases a PC on the PV of external borrowing contracted would be required. Under a PV limit, the PV is calculated using the terms of individual loans and the unified discount rate, thus allowing more space for concessional debt. Thus, PV Limits can enable countries to access some non-concessional financing while providing incentives to optimize the terms of their borrowing (as they can borrow greater volumes if loans have higher concessionality), while eliminating distortive threshold effects (currently, all financing provided at a rate of concessionality even slightly below 35 percent is considered as non-concessional).40 The calibration of the PV would be informed by the authorities’ external borrowing plan and the DSA, to ensure new borrowing does not lead to a deteriorating debt situation.

  • In limited circumstances where the capacity of countries to monitor and observe conditionality on aggregate debt levels (including in PV terms) is not adequate, a PC setting a nominal NCB limit and a memo item on CB should be used instead.

    • This determination would be made by staff based on assessments of capacity for debt recording, monitoring, and reporting, rather than on broader debt management capacity. Sources of such information include: the World Bank’s publicly available Debt Management Performance Assessments (DEMPAs); IMF technical assistance reports; UNCTAD’s Debt Management and Financial Analysis System (DMFAS) reports or Commonwealth Secretariat’s (COMSEC) reports; IMF staff experience with the quality of debt monitoring and reporting in previous IMF-supported programs; and the Debt Reporting Heat Map.41 Teams may consult SPR/MCM to assist with the interpretation of these various information sources.

    • Applying these assessment tools will require a case-by-case approach, combining the tools’ results and staff judgement to determine whether a relevant capacity concern has been identified. Thus, the basis for the assessment should be clearly explained in the staff report to facilitate oversight by IMF management and the Executive Board.

    • As with a PV limit, the nominal NCB limit and the memo CB memo item would be informed by the authorities’ external borrowing plan and the DSA (to ensure new borrowing does not lead to a deteriorating debt situation). The CB memo item would aim to help country authorities in improving their monitoring and recording of debt on an aggregate basis, with a view to eventually moving to PV limits.

    • Whether in nominal or PV terms, the debt limit should be calibrated in such a way that the baseline projection does not generate a shift in the external debt distress rating from moderate to high (Box 3). In particular, higher scrutiny of borrowing plans for countries at moderate risk of debt distress with limited space to absorb shocks is required. Relatively small deviations from the baseline macro and financing assumptions could cause these countries to be reclassified into the high risk of debt distress category. Thus, the limit should be calibrated to have an adequate buffer to accommodate such shocks without tipping the country into high risk of debt distress. Such a buffer could take the form of flexibility in project execution to phase the disbursement of debt over time. Nevertheless, even with appropriate buffers, a shift into high risk could arise during the course of the program due to unanticipated exogenous shocks. In such cases, staff reports should provide a clear explanation of the factors driving the deterioration in the risk of external debt distress.

32. High risk of or in external debt distress:

  • For countries assessed at high risk of debt distress or in external debt distress,42 debt conditionality will typically take the form of a PC with a zero limit on NCB and an IT (or PC)43 on the PV of external borrowing. NCB is relatively less common in these countries (that normally rely on concessional borrowing without significant access to international financial markets) and tends to be project oriented. The zero NCB limit (with possible exceptions below) can usefully encourage concessional financing. The IT/PC on the PV of public external borrowing should be calibrated based on the principles outlined in Box 5, with the aim to prevent the country from moving further into the high-risk category by the end of the program. A memo item on CB should be used as an alternative in cases where capacity to set conditionality on aggregated debt levels was deemed inadequate (see ¶31).

  • NCB may be allowed only in exceptional circumstances. Such exceptions may be warranted where: (a) financing is needed for a project integral to the authorities’ development program for which concessional financing is not available; or (b) non-concessional borrowing is used for debt management operations that improve the overall public debt profile.44 These circumstances are discussed in more detail in Section V.C. including safeguards on repeated use of exceptions.

(ii) Countries that Normally Rely on Concessional Financing with Significant Access to International Financial Markets

33. For countries in this group at low risk of debt distress, debt conditionality is not required unless specific debt vulnerabilities warrant targeted conditionality. This would follow the same treatment as those countries that normally rely on concessional financing without significant access to international financial markets (¶30).

How to Determine the Level of Quantitative Debt Limits

Calibrating the appropriate level of a quantitative debt limit starts with a country’s borrowing plan, which is an integral part of the macroeconomic policy framework for an IMF-supported program. Regardless of a country’s financing mix, a borrowing plan specifying the projected financing from each of the main sources to cover a planned fiscal deficit is usually prepared. Its design is determined by several factors, including: the appropriateness of the underlying fiscal deficit from a demand management perspective; the compatibility of the fiscal deficit and borrowing plan with maintaining debt sustainability over the medium-term; and the feasibility of implementing envisaged investment projects given capacity constraints. Key features of borrowing plans and requirements for program documentation are further described in Annex III.

The amount of borrowing space available for the borrowing plan will depend on the nature and extent of vulnerabilities identified in the DSA. The size of the borrowing envelope should be based on country needs with an eye towards keeping (or bringing) the PV of debt below the ‘binding’ high-risk thresholds under the baseline, while delivering acceptable dynamics under the relevant stress scenarios. Investment plans, and associated borrowing, should take into account public capital gaps while fully recognizing inefficiencies in public investment spending.1 The need for fiscal buffers to accommodate the impact of an exogenous shock or to implement countercyclical fiscal policy to combat a possible future recession should also be considered. Ultimately, for moderate-risk cases, a borrowing limit should typically aim to avoid breaching any DSA thresholds under the baseline (as that would result in a high-risk rating if the breaches are not temporary). For high-risk cases, the IT on the PV of external borrowing should typically not move the country further into the high-risk category over the course of the program. In all cases, program targets should remain realistic, as informed by Figure 4 of the LIC-DSF.2

Determining the level of debt limits will require close coordination with the authorities. This process is linked to the broader work with the country authorities on a suitable fiscal program, including a realistic borrowing plan that does not pose undue risks to debt sustainability. An initial borrowing plan based on the planned financing for fiscal and project needs would be assessed using the DSA to analyze whether the envisaged levels and mix of borrowing pose significant risks to debt sustainability. As needed, the fiscal program and investment projects and the associated borrowing plan are adjusted in a manner that reduces these risks. Once a suitable borrowing plan is agreed it will determine the debt limits ensuring that these limits are consistent with both the fiscal framework and preserving debt sustainability.

In cases where Fund-supported programs entail a sovereign debt restructuring, debt limits should be consistent with the targeted debt reduction under the authorities’ debt restructuring strategy. Calibrating debt limits will generally involve a two-step process. At program request, debt limits consistent with the classification of debt distress risk would apply, such as high risk or in distress. This would typically be a zero NCB limit, with the possibility for exceptions for debt management operations that improve the debt profile, which may be particularly relevant in a restructuring context (Section V.C.). If, over the course of the program, following the implementation of the restructuring and other developments, the updated DSA places the country firmly back at moderate risk of debt distress earlier, there may be some space for a positive NCB limit as long as it preserves the buffer needed to prevent the country returning to high risk of debt distress as a result of exogenous shocks.

1 Since the marginal return to investment is likely to be higher in countries featuring larger gaps, more investment (even if financed through borrowing) may—up to a point—improve debt sustainability (Akitoby and Stratmann, 2008; IMF, 2018). However, in cases with a low efficiency in implementing infrastructure projects, additional debt is matched by modest effective improvements in infrastructure in practice, putting debt sustainability at risk. Capacity development and possibly structural conditionality could then be warranted to improve this efficiency.2 See Guidance Note on the Bank-Fund Debt Sustainability Framework for Low-Income Countries.

34. For countries in this group which are at moderate risk, high risk or in external debt distress, debt conditionality would need to account for these countries’ specific financing circumstances.45 Since access to market financing may be an important part of the borrowing plan, zero NCB limits for countries at high risk may not be appropriate. Thus, conditionality should take the following forms:

  • The default of a PC on the PV of external borrowing. As with PV limits for moderate risk countries, the level of these limits would need to be calibrated based on the authorities’ borrowing plan and the DSA (Box 5).

  • Alternative formulations when better tailored to address critical debt vulnerabilities. To use such a formulation, teams would need to make the case, using the DSA and applying judgement that such a formulation would be better tailored compared with a limit on the PV of external borrowing. An example could be a PC on the aggregate level of public debt (including domestic debt), specified in nominal terms, in cases where there is substantial foreign investor participation in the domestic bond market. With foreign investors moving in and out of domestic instruments and/or between domestic and foreign-currency bond issues, drawing a distinction between domestically and externally sourced financing may be difficult (see |44).46 Other examples could follow the conditionality design for countries that do not normally rely on concessional financing (see Section IV.C).

C. Countries that Do Not Normally Rely on Concessional Financing

35. If needed, borrowing limits should typically be of the nominal type, since these countries are generally not able to borrow on concessional terms. When the coverage of fiscal conditionality is comprehensive—encompassing both domestic and external financing (with the latter sometimes being excluded from fiscal PCs in LICs)—a standard fiscal PC may suffice to safeguard debt sustainability (see ¶21–22). In cases where risks are stemming from contingent liabilities (and broadening fiscal coverage is not deemed possible/effective), fiscal conditionality may be augmented with nominal limits on the issuance of new government guarantees, borrowing by SOEs, or on the contracting of PPPs.47

36. Along similar lines to countries that normally rely on concessional financing, the framework used to analyze debt sustainability the SRDSF can be used to determine the nature of debt vulnerabilities. The framework is currently being revised, with IMF (2021) laying out the envisioned structure of the new framework. Like the LIC-DSF, the SRDSF assesses debt vulnerabilities along the three dimensions discussed in Section III.D.48

  • Risks stemming from high debt levels are mostly captured by the fanchart-module (reported in SRDSF Figure A4; see Box 6 for examples of figures), which incorporates information from the (projected) level of debt under the baseline, the uncertainty regarding its future path (proxied by the fanchart’s width), alongside the probability of medium-term debt stabilization. In addition, the contribution of the debt burden to the logit stress probability (reported in Figure A3) can inform this risk assessment as well.

  • Risks stemming from high debt service obligations can be assessed via the Gross Financing Needs (GFN)-module (reported in SRDSF Figure A4). This module generates GFN projections under the baseline and a stress scenario and produces a “risk index” while taking the size of a country’s banking system into account, the latter’s initial exposure to the sovereign, as well as any liquid asset buffers the country may hold. Based upon the resulting index, the country receives a low/medium/high-risk rating, which should be a key input in assessing the need for GFN-focused conditionality. In addition to looking at the aggregate index (which averages over all projection years), teams should also analyze the year-by-year GFN trajectory—particularly looking for any spikes in the amortization profile (which may warrant debt management operations).

  • Risks stemming from debt structure are conveyed through the modules dealing with debt coverage (SRDSF Table A1), while Figure A1 produces detailed information on debt by currency, the debt holder profile, the nature of debt instruments, as well as the maturity of debt. For example, a large share of FX-denominated debt and high currency risks could warrant a debt limit on foreign currency borrowing. Details on possible collateralization arrangements are not part of the DSA template, so the relevant information should be obtained from the authorities using a risk-based approach (potentially subject to confidentiality arrangements) as discussed in Section III.A.

37. Similar to how the LIC-DSF can inform the calibration of debt limits (Box 5), teams may wish to use the fanchart- and GFN-modules of the SRDSF to inform the calibration of debt (and fiscal) conditionality under the program baseline. An objective could be to prevent or move a country from being assessed as “high risk.” Additionally, teams may wish to check that the program baseline would keep countries out of the high-risk zone of the sovereign stress logit when assessed at the end of the program. These objectives should be achievable without giving rise to severe realism concerns (as conveyed through Figure A2 of the SRDSF). If not, the program may need to be redesigned so that it gains credibility.

Example Figures for SRDSF (Formerly MAC DSA)

Table A1 and Figure A1 in the SRDSF/MAC DSA shows the risks stemming from debt structure, and Figure A4 shows risks stemming from high debt levels as well as risks stemming from high debt service obligations.

SRDSF/MAC DSA Table A1.

Debt Coverage

article image
SRDSF/MAC DSA Figure A1:
SRDSF/MAC DSA Figure A1:

Public Sector Debt Structure Figure

Citation: Policy Papers 2021, 037; 10.5089/9781513583617.007.A001

1/ Includes SDR allocations, holdings by regional financing arrangements, and Currency union Central banks, where applicable
1/ For more information, please see Review of The Debt Sustainability Framework For Market Access Countries.

D. Other Operational Considerations

IT versus PC

38. For debt conditionality outside the requirements discussed in Section IV.C, the decision as to whether debt limits should be a PC or an IT will follow 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 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 should take the form of ITs to assess progress in meeting program objectives. ITs can also be used in cases where PCs cannot be established due to substantial uncertainty about economic trends.

Coverage of Debt Limits

39. For the purpose of the debt limits policy, public sector debt coverage should be broad, thereby complementing fiscal conditionality, such that it:49

  • i. Should normally cover public and publicly guaranteed debt. However, institutional constraints and lack of data availability could justify the use of narrower sectoral coverage (e.g., central government), or focusing on specific public debt instruments (e.g., short-term external debt, issuance of central government guarantees).50 Coverage expectations are similar to those in the LIC-DSF, which focuses principally on external PPG debt for ratings purposes. However, in line with the Guidelines on Conditionality, when determining coverage for debt conditionality, careful attention is needed to ensure the authorities are able to feasibly monitor such debt in a program context, potentially resulting in narrower coverage.

  • ii. Would typically refer to debt of the nonfinancial public sector. This comprises debt of the consolidated central government, state governments, local governments, social security funds, as well as nonfinancial public enterprises and other nonfinancial official sector entities to the extent such coverage is feasible. In any case, the definition of public sector debt should be made clear in program documents (e.g., in the technical memorandum of understanding or TMU) with any exclusions explicitly documented.

40. The inclusion of liabilities of the central bank or other public financial institutions in public debt would depend on country circumstances. Central bank debt issued solely for monetary purposes would normally be excluded from the definition of public sector debt with the exception of liquidity papers,51 which may be included if they are included in public debt in the SRDSF/MAC DSA.52 The inclusion of other liabilities of the central bank or other public financial institutions could be justified when they pose significant fiscal risks and excluding them would undermine the effectiveness of debt limits in addressing debt vulnerabilities. Similarly, the external debt of the central bank should be included under public external debt limits if such debt contributes to external debt vulnerabilities and associated fiscal risks. In countries where the central bank also acts as the agent of the government and issues treasury bonds on behalf of the government, such issuance would constitute public debt. Similarly, government borrowings from the central bank should normally be included in public debt. In general, the decision to include any financial instrument (including swaps, deposits, or repo agreements) for the purposes of public debt conditionality should be based on whether they present a fiscal risk, and it should be spelled out clearly in the program documents (e.g., the TMU). For SRDSF/MAC DSA countries, the treatment of bilateral FX swap liabilities should be consistent with the DSA.53

41. “Commercially viable” SOEs may be explicitly excluded from the limits, when they do not pose fiscal risks. Given the increasing importance of SOE investment activities and the frequent materialization of contingent liability associated with SOEs, debt limits should cover those with substantial borrowing activities that could impose a fiscal burden on the general government (unless government debt already includes SOE debt guaranteed by the government). However, a case for selective exclusion can be made for commercially-viable SOEs that may borrow without a government guarantee and whose operations pose limited fiscal risk to the government. This treatment follows the treatment of SOEs in the LIC-DSF.54 In particular, the decision to exclude a particular SOE should be guided by whether it poses a limited fiscal risk, including an assessment of whether (i) it does not carry out uncompensated quasi-fiscal activities; and (ii) it has positive operating balances. Additional relevant indicators should be considered as well: SOE’s managerial independence; financial relations with the government (including offering collateral to government borrowing unrelated to the SOE’s business); the periodicity and independence of audits; publication of comprehensive annual reports and protection of shareholder rights; financial performance (e.g., return on assets (ROA), return on equity (ROE)) and sustainability; and other risk factors.55

Treatment of IMF and World Bank Financing

42. IMF financing should be excluded via language in the TMU to facilitate the use of IMF resources. Such exclusions would preserve borrowing space for the IMF financing when limits are set on a disbursement basis and leave open the possibility for additional IMF financing (e.g., RCF/RFI requests) that may not always meet concessionality definitions when limits are set on a contracting basis. A TMU exclusion may also be needed to accommodate domestic policies related to IMF borrowing. For example, when debt conditionality covers guaranteed debt and borrowings by the central bank require guarantees from the Ministry of Finance. In that regard, IMF financing generally does not fall within the definition of “debt” under domestic law.

43. Use of Adjustors for World Bank-Financed Projects in Debt Limits: In line with the objective of close coordination with the World Bank in IMF-supported programs and more broadly, the design of debt limits should, to the extent feasible, seek to avoid the crowding out of planned project lending by the World Bank. Teams are encouraged to make use of downward adjustors to cover World Bank lending projects that feature in the authorities’ borrowing plan and are expected to be contracted within the debt limit period. The size of these adjustors should be limited to the value of the identified projects and should be revisited at each program review to ensure that they incorporate updated information.56 If previously anticipated Bank financing is no longer expected to materialize or is delayed past the debt limit period, the overall debt limit should be revisited at the next review to consider the potential use of freed-up borrowing space. In instances where conditionality is specified in terms of a zero limit on NCB, an NCB exemption can be requested if the World Bank financing does not meet the 35 percent grant element threshold (as could be expected to occur in the case of IBRD financing for IDA-blend countries or project finance for IDA countries via the IDA Scale-up Window).57 More broadly, TMU exclusions for World Bank financing for budget support can also be used given their role in filling program financing gaps.

External Public Debt: Residency or Currency Based

44. The appropriate criterion to define external debt (whether residency or currency) for the purposes of debt limits will depend on country circumstances.58 For countries that do not have significant access to international financial markets, the choice of currency or residency criterion for measuring external debt is not a major issue, as the two approaches are likely equivalent. However, high nonresident participation in the primary and/or secondary domestic debt markets (in both local- and foreign currency) can challenge this equivalence. This opens the possibility that external debt defined on a residency basis may not be satisfactorily monitorable under the program, or sufficiently under the control of the authorities.

45. For countries with relatively closed financial accounts or very limited financial integration with the rest of the world, the use of the residency criterion would remain appropriate. Nonresident acquisition of domestically-issued debt in the secondary market is expected to be very limited in this case, and it may be reasonable to exclude such debt from the definition of external debt for the purpose of the debt limits policy. In such cases, vulnerabilities associated with the excluded debt instruments would need to be addressed in the program, including, through the following safeguards (tailored to country circumstances as appropriate):

  • i. The program relies on an appropriately broad concept for the fiscal balance performance criterion, to close any definitional loopholes;

  • ii. All new borrowing on the domestic market is normally in local currency;

  • iii. Relevant changes are made in the program’s design (e.g., higher NIR targets), if needed to mitigate vulnerabilities associated with significant nonresident holdings of domestically-issued debt;

  • iv. Any significant accumulation of domestically-issued debt by nonresidents is reflected, to the extent possible, in the LIC-DSF with an explicit assessment of the potential vulnerabilities (e.g., higher rollover risk in the case of short-term borrowing, a potential threat to the exchange rate and/or reserves in the event of sudden withdrawals);

  • v. The authorities report to the Fund the terms of new domestically-issued debt, including the currency composition, and take steps over time to improve their monitoring of secondary market transactions; and

  • vi. The authorities are not signatories in transactions that involve the immediate repackaging of domestically-issued debt instruments for the sole purpose of reselling the repackaged instruments to nonresidents.

46. For countries with an intermediate degree of financial integration and where the use of residency criterion becomes problematic, a currency of denomination criterion could be used. These are typically countries that have seen increasing nonresident ownership of their domestically-issued debt, but still fall into the category of countries that normally rely on concessional external financing. For these countries, a definition of external debt based on the residency criterion encompassing all debt held by nonresidents could be difficult for the authorities to monitor and control. In such cases, the definition of external public debt and concessionality requirements (where needed) could be applied to foreign-currency denominated public debt. However, risks from local currency debt held by nonresidents should be actively tracked and discussed in the DSA and in the External Sector Assessment (ESA). When collecting debt data and analyzing debt vulnerabilities, country teams should also be mindful of the possibility that domestically-issued debt could be foreign-currency indexed or that the repayment currency may not be the same as the currency of issuance.

Contracting versus Disbursement

47. When invoked, debt limits may be set either on a contracting or a disbursement basis:

  • In cases where much of the external financing takes the form of project loans that are disbursed over an extended period, it would typically be more appropriate to specify debt conditionality in terms of the contracting of new debt which is typically under the control of the authorities, rather than on the disbursement of new debt, which may be driven by the uncertain pace of project implementation over time. When limits are set on a contracting basis, country teams would need to make some assumptions about the phasing of the disbursement to evaluate the impact on the DSA. However, the full contracted amount of the debt would be taken upfront in the debt limit, upon signing of the loan agreement. Given that some contracted debt may never end up being disbursed or new donor financing could be catalyzed over the course of the program, these disbursement assumptions should be updated to ensure a consistent DSA. This could also warrant revisions to contracted debt limits at subsequent program reviews.

  • Since country practices differ as to the procedures to be followed in regard to the contracting of public external debt, the TMU should include a clear indication of the precise stage in these processes at which, for program purposes, the debt is viewed as having been contracted (see Annex II).

Section V. Implementing Debt Limits

A. Monitoring Debt Limits

48. The primary responsibility for ensuring that debt contracted or disbursed observes the limits agreed in the program lies with the authorities. This is consistent with the national ownership principle in the Guidelines on Conditionality. Failure to accurately report compliance with the debt limits established as PCs is subject to misreporting, as the authorities should made aware.

49. Nevertheless, country teams play a role in supporting countries to ensure their compliance. Two important tools in this regard are the PV-tool for setting and monitoring debt limits as well as the grant element calculator.59 The PV tool is an Excel file that allows users to calculate the present value and grant element for multiple loans at the same time. It provides summary statistics for the entire debt portfolio, including total present value, weighted average grant element, and other useful statistics such as interest rate range and variable interest loan exposure. The web-based grant element calculator facilitates the calculation of the grant element of an individual debt instrument, accounting for commissions, fees, and alternative standard repayment profiles. Country teams should also seek consultation and information sharing requirements in the TMU to ensure countries inform them sufficiently early before the contracting of any new borrowing. When there are concerns about the financing terms or possible collateralization, a best practice would be to request the authorities to provide information on the financing agreement or the agreement itself. LEG and SPR can help in determining how best to classify it.

B. Assessing Concessionality

50. Debt is considered concessional if it has a grant element of 35 percent. This definition is well anchored in global practices, including those employed by the World Bank and OECD. A standardized definition can also help countries attract concessional financing. The grant element of a debt is the difference between the PV debt and its nominal value, expressed as a percentage of the nominal value of the debt, and further details on its calculation are discussed in Annex IV. For operational purposes, a grant element calculator is also available on the IMF’s DLP website.

51. A concessionality definition with a grant element higher than 35 percent is only permitted in cases where it is deemed an integral part of restoring debt sustainability. These would include countries undergoing debt restructuring or those receiving debt relief including under the Highly-Indebted Poor Countries (HPIC) initiative. Two key considerations for setting a concessionality threshold with a higher grant element are: (i) the extent to which it restores debt sustainability through improvements in indicators in the DSA; and (ii) the availability of highly concessional financing, which could be verified through consultations with the authorities and key donor partners.

52. Some non-standard types of financing are automatically treated as non-concessional, with a zero-grant element. These include blended financing arrangements that include the provision of a financially significant amount of grants in kind (e.g., grants provided in the form of equipment or machinery for a project) and financing involving unrelated collateral (e.g., general budget borrowing collateralized by earmarking of commodity receipts). Annex IV describes these types of financing in more detail.

C. Exceptions to NCB Limits

Project-Related NCB Exceptions and the Signal-Based Approach

53. Exceptions to zero NCB limits to allow for certain project financing may be warranted if both: (i) the project is integral to the authorities’ development program; and (ii) concessional financing is not available.60 The project should credibly be projected to yield a high economic and social return, and importantly, it should be aligned with the authorities’ development program to underpin ownership within the Fund-supported program.61 Given that a key objective of the DLP is to encourage creditors to provide and debtors to seek concessional financing, it is also important to establish that concessional financing for such a project is not available. Nonetheless, repeated use of such exceptions could warrant closer attention as discussed at the end of this section.

54. Determining when a project meets these two criteria should follow a signal-based approach. Although it is important to credibly establish that a project has high expected economic and social returns, individual project assessment is beyond the mandate and expertise of the IMF. Staff should instead seek out signals to determine whether NCB exceptions are justified. Under the signal-based approach, at least one strong signal would be required indicating that the project is integral to the authorities’ development program and one strong signal indicating concessional financing is not available. Table 4 establishes an indicative list of core options for signals. This list may be added to over time by country teams (for example, if another development priority like the Disaster Resilience Strategy or Covid-19 response emerged). Staff are encouraged to check internally as to whether new signals have been added (by consulting the Debt Policy division in SPR).

Table 4.

Signals for NCB Project Exceptions—Core Options1

article image

These signals are a set of core options and additional signals could be added over time, provided that they meet the information quality guidelines outlined in ¶53. When there are concerns about the quality of a given signal, an additional signal would be warranted.

55. The quality of information provided by signals needs to be carefully assessed to ensure their credibility. This should be carefully assessed on a case-by-case basis by the country team and during the internal review process. In cases where a signal is considered meaningful, yet there are some concerns about the quality of that signal, an additional distinct signal to meet the criteria would be required. Assessing the quality of a signal would involve the following considerations:

  • For signals to determine whether a project is integral to the authorities’ development plan, a strong signal would generally meet at least one of the following criteria: (i) the project is supported by a credible assessment of its economic and social returns; (ii) the project is explicitly reflected in a strategic development document of the authorities, which is supported by sound assessment of project benefits; or (iii) there is evidence of the authorities’ capacity to select and assess projects. Country teams should also carefully weigh the risks of using any signals and the timeframe (if assessments are involved) to avoid undue delays in program negotiations. Additional considerations for specific core options outlined in Table 4 are discussed in Box 7.

  • For signals to establish whether there is no alternative concessional financing for a project, it would be important to establish that the authorities made recent suitable efforts to seek financing from development partners for the project in question. It is also important to understand why these efforts may have been unsuccessful to ensure that the inability to raise project financing does not reflect development partners’ doubts regarding its economic and social returns.

Ensuring Quality in Signals to Determine When Projects Integral to Authorities’ Development Program—Specific Considerations

Project is commercial in nature. In a project that is commercial in nature (e.g. a joint venture with a foreign partner), commercial net revenues would likely exceed debt service costs, ultimately limiting the fiscal liability associated with any potential NCB. The analysis confirming this commercial assessment would need to be carefully examined to ensure it is credible (e.g., the assessment carried out by a reputable and internationally-recognized firm).

Project is part of an existing Public Sector Investment Program (PSIP) or National Development Plan (NDP). The inclusion of a project in a country’s PSIP or NDP could be an important indication from the authorities of a project being integral to national development. However, PSIPs and NDP can vary both in terms of the rigor by which identified projects are assessed and selected, and the extent to which the projects identified are national priorities. Thus, a key consideration should be the extent to which projects included the plan have been subject to a rigorous assessment of the economic and social returns. Known weaknesses in project appraisal and selection (e.g., as identified through public investment related TA) may be reason to be concerned about this signal’s quality.

Disaster Resilience Strategy (DRS). The inclusion of a project in a country’s DRS can be an important signal of a project being integral to national development, given that building resilience to natural disasters is macro-critical for some countries. A key consideration for use of this signal would be the extent to which the DRS follows IMF recommendations,1 particularly the extent to which the DRS is grounded in a comprehensive forward-looking diagnostic that includes risk assessment, cost-benefit analysis, and implications for the macroeconomic and debt trajectory.

COVID-19 response or mitigation strategy. While the DLP does not entail special transitional previsions or carve outs due to the COVID-19 shock, given the paramount importance of emergency response measures , the inclusion of a project as part of the authorities’ response or mitigation strategy can be a good indication of it being integral to national development. However, an important consideration to support the credibility of this signal, and the potential for the project to yield high economic and social returns, would be whether such spending followed best practices for fiscal transparency and accountability.2

The project is endorsed/undertaken by an MDB. Given shared objectives and track record of close coordination in a program context, an endorsement (or conducting) of a project by donor/partner MDBs could signal its high economic and social returns, considering their mandate and expertise.3 An assessment letter from the MDB would provide a clear signal. Projects endorsed or undertaken by other regional or plurilateral financing institutions could also be considered but may require some additional examination of how the institution has assessed the project.

Public Investment Management Assessment (PIMA). A PIMA of being an adequate performer would indicate that a country has sound frameworks and practices to select and manage projects with high economic and social returns. Average institutional and effectiveness scores for LICs (currently 1.7) may serve as a guideline benchmark. However, teams should consult the Fiscal Affairs Department to further interpret results and appropriately factor in potential limitations. For example, PIMA scores might not reflect a country’s current practices or may be influenced by the composition of previous public investments, including the share of externally financed projects. In cases where a recent PIMA is not available, FAD may also be able to provide useful information on current Public Investment Management (PIM) practices.

Credible independent assessment. While conducting individual project assessments on social and economic returns is beyond the expertise and core mandate of the IMF, a country may be able to obtain this elsewhere from an MDB or an independent assessor. Use of this signal should carefully assess the: (i) extent to which the assessment is credible (e.g., whether the independent assessor is internationally recognized); and (ii) whether the assessment is truly independent in that there are no conflicts of interest in terms of potential gains from the project (or appropriate oversight mechanisms to avoid potential conflicts).

1Building Resilience in Countries Vulnerable to Large Natural Disasters (IMF, 2019).2 See IMF staff note Keeping the Receipts: Transparency, Accountability, and Legitimacy in Emergency Responses. (April 20, 2020).3 These would include MDBs that lend alongside the Fund to provide budget support for countries facing macroeconomic vulnerabilities, including the African Development Bank (AfDB), Asian Development Bank (ADB), Inter-American Development Bank (IDB), and the World Bank (World Bank).

56. The case for any exceptions to NCB limits should be clearly set out in program documents to facilitate review by Management and the Executive Board. This should include: (i) clear identification of the signal(s) for each of the two requirements (integral to development program and lack of available concessional resources); (ii) a brief explanation as to why the selected signals are assessed as providing quality information; and (iii) in cases where there were concerns about the quality of a given signal, a similar discussion of the additional signals used would be required.

Debt Management Related NCB Exceptions

57. NCB exceptions for debt management operations may be warranted if the operation results in an improvement in the overall debt profile.62 Whether such an operation constitutes an improvement, is determined via the DSA. A debt management operation would need to show either an improvement in the key liquidity and/or solvency debt burden indicators without adversely affecting the risk rating. Staff should carefully assess the realism of any assumptions associated with the impact of the operation (including the realism of the expected financing conditions of such an operations and the extent to which all costs associated with the operation have been fully reflected). An additional consideration could be how the debt management operation fits into the broader debt management strategy, including the MTDS (if available). As with NCB exceptions for projects, the case for the use of such an exception should be documented in the staff report, including a brief discussion of the assessed impact on key DSA indicators.

Repeated Use of NCB Exceptions

58. In cases where NCB project exceptions are sought repeatedly for multiple projects, further assessment is required to guard against potential misuse. The primary consideration would be the track record of the use of NCB exceptions to ensure the policy is not being used to circumvent the debt limits. For example, an ex-post assessment that the economic and social returns for a project where an NCB exception was granted were low would suggest against further NCB exceptions. In addition, it would be important to establish whether the authorities were making sufficient efforts to seek concessional financing and why those efforts were not successful. For instance, a lack of available of concessional financing due to a relatively high income might be a valid reason to allow for an NCB exception.

59. Further assessment is also warranted in cases where NCB exceptions are repeatedly sought for debt management purposes to ensure the debt limit in place remains the correct fit. In cases where a country repeatedly requests debt management exceptions, it would be important to assess whether this could be a sign that the country should be recast as accessing international capital markets on a significant scale (with a default non-zero NCB limits and a possibility to tailor further). Past and prospective market financing should be examined (See Section II.B).

D. Coordination with the World Bank

60. The DLP is closely aligned with the World Bank’s Sustainable Development Finance Policy (SDFP). The DLP applies to all IMF members with a Fund-supported program and the SDFP applies to all IDA-eligible countries. Both policies aim to help countries contain debt vulnerabilities while providing incentives for countries with access to concessional financing to seek it63

61. Teams should coordinate with World Bank counterparts to ensure that close alignment between the debt policy commitments is maintained. When initiating a new Fund-supported program, program design should take account of borrowing limits set under the SDFP together with an updated assessment of debt vulnerabilities. The SDFP maintains the approach of aligning, in principle, borrowing limits with the DLP when there is an IMF-supported program. When an IMF-supported program expires, the SDFP would take account of borrowing limits set under the DLP together with an updated assessment of debt vulnerabilities. The policies on NCB exceptions for both institutions are also broadly aligned as both the DLP and SDFP would only allow for NCB exceptions related to critical projects where concessional financing is not available, and/or related to debt management operations. The calibration of limits for countries normally relying on concessional financing under both policies would be underpinned by the joint LIC-DSF.64 Hence, the development of the LIC-DSF is also an opportunity to discuss debt limits under the new Fund-supported program.

62. As with the LIC-DSF, in instances of diverging views between the IMF and World Bank staff, country teams should seek to first resolve any disagreement at the working level. During a Fund-supported program, the SDFP will in principle align debt limits with those of the program. A main potential area of difference could be when a PV-limit is used in the DLP, while the SDFP may opt to maintain a nominal NCB limit. In these instances, country teams should aim to calibrate a PV-limit assuming NCB borrowing that also adheres to the SDFP. Other potential areas of divergence could include: the coverage of debt limits; and instances of tailored debt limits either to address specific debt vulnerabilities or for those country cases that may use tailored options under the DLP.65 If agreement cannot be reached at the working level, the IMF mission chief and the MTI director should, after consultation with their respective review departments (SPR in the Fund, Global Macro and Debt Analytics Unit in the Bank), seek a resolution.

Annex I. Overview of 2020 Debt Limits Policy Reforms

article image

Annex II. Program Documentation

A. Staff Report Documentation Requirements

The following provides guidance on information to be included in staff reports.

article image

B. Suggested TMU Language

This section provides guidance on the TMU language for setting performance criteria and indicative targets on debt in Fund arrangements.1 Country teams should omit sections that are not relevant to their country.

Definition of Debt

1. For program purposes, the definition of debt is set out in paragraph 8(a) of the Guidelines on Public Debt Conditionality in Fund Arrangements attached to Executive Board Decision No. 15688-(14/107), adopted December 5, 2014.2

(a) For the purpose of this guideline, the term “debt” will be understood to mean a current, i.e., not contingent, liability, created under a contractual arrangement through the provision of value in the form of assets (including currency) or services, and which requires the obligor to make one or more payments in the form of assets (including currency) or services, at some future point(s) in time; these payments will discharge the principal and/or interest liabilities incurred under the contract. Debts can take a number of forms, the primary ones being as follows:

  • (i) loans, i.e., advances of money to the obligor by the lender made on the basis of an undertaking that the obligor will repay the funds in the future (including deposits, bonds, debentures, commercial loans and buyers’ credits) and temporary exchanges of assets that are equivalent to fully collateralized loans under which the obligor is required to repay the funds, and usually pay interest, by repurchasing the collateral from the buyer in the future (such as repurchase agreements and official swap arrangements);

  • (ii) suppliers’ credits, i.e., contracts where the supplier permits the obligor to defer payments until some time after the date on which the goods are delivered or services are provided; and

  • (iii) leases, i.e., arrangements under which property is provided which the lessee has the right to use for one or more specified period(s) of time that are usually shorter than the total expected service life of the property, while the lessor retains the title to the property. For the purpose of the guideline, the debt is the present value (at the inception of the lease) of all lease payments expected to be made during the period of the agreement excluding those payments that cover the operation, repair or maintenance of the property.

(b) Under the definition of debt set out in this paragraph, arrears, penalties, and judicially awarded damages arising from the failure to make payment under a contractual obligation that constitutes debt are debt. Failure to make payment on an obligation that is not considered debt under this definition (e.g., payment on delivery) will not give rise to debt.

Coverage of Debt

2. For the purposes of this debt limit ceiling, public sector debt covers public and publicly guaranteed debt.

3. The public sector comprises the central government, state governments, local governments, social security funds, [the central bank], nonfinancial public enterprises (enterprises and agencies in which the government controls the entity,3 but which are not consolidated in the budget), and [other official sector entities].4 The government’s control of an entity will be assessed according to the methodology defined in GFSM 2014, Chapter 2. The following entities however are excluded from the coverage: [Please list any specific nonfinancial public enterprises or other public entities that are excluded from the definition of public sector].

4. Text to be inserted where relevant to the program design, if central bank is not included in the coverage of public sector debt [Notwithstanding the terms of paragraph [2] above, repurchase agreements and swaps of the central bank or other public financial institutions, shall be excluded from the definition of debt.] [Central bank debt issued solely for monetary purposes is excluded from the definition of public sector debt.] [the definition of debt also excludes the following central bank debt: …….]

5. Text to be inserted where relevant to the program design, if central bank is included in the coverage of public sector debt [Central bank treasury bonds issued in the name and on behalf of the government, as well as those issued in the name of the central bank but on behalf of the government, are included in the coverage of public debt.]5

6. For program purposes, a ‘guaranteed debt’ is an explicit promise by the public sector to pay or service a third party obligation (involving payments in cash or in kind).

Contracting of Debt and Treatment of Credit Lines

7. For program purposes, a debt is considered to be contracted when all conditions for its entry into effect have been met, including approval by [specify relevant decision making unit or parliament].6 Contracting of credit lines (which can be drawn at any time and entered into effect) with no predetermined disbursement schedules or with multiple disbursements will be also considered as contracting of debt.7

External Debt

8. For the purposes of the ceiling on the contracting or guaranteeing of new [non-concessional] [concessional] external debt, external debt is any debt contracted or guaranteed by the public sector on [non-concessional] [concessional] terms [with non-residents] [denominated in foreign currency, i.e., currency other than [country’s] currency]].

Domestic Debt

9. For the purposes of the ceiling on the domestic debt, domestic debt is any debt contracted or guaranteed by the public sector [with residents] [denominated in [country’s] currency].

Short-Term Debt

10. For the purposes of the ceiling on the contracting and guaranteeing of debt, short-term debt is any debt contracted or guaranteed by the public sector with original maturities of up to and including one year.

Concessionality

11. For program purposes, a debt is concessional if it includes a grant element of at least 35 percent, calculated as follows: the grant element of a debt is the difference between the present value (PV) of debt and its nominal value, expressed as a percentage of the nominal value of the debt. The PV of debt at the time of its contracting is calculated by discounting the future stream of payments of debt service due on this debt.8 For debts with a grant element equal or below zero, the PV will be set equal to the nominal value of the debt. The discount rate used for this purpose is the unified discount rate of 5 percent set forth in Executive Board Decision No. 15248-(13/97).

12. For debts carrying a variable interest rate in the form of a benchmark interest rate plus a fixed spread, the PV of the debt would be calculated using a program reference rate plus the fixed spread (in basis points) specified in the debt contract. The program reference rate for the six-month USD LIBOR is [3.34] percent and will remain fixed for the duration of the program. The spread of six-6[For background (not to be included in TMUs): if no specific definition is included in the program documents, a debt will be deemed contracted in accordance with the terms of the contract and as provided for by the law of the member that determines the requirements for the effectiveness of such contracts.] month Euro LIBOR over six-month USD LIBOR is [-250] basis points. The spread of six-month JPY LIBOR over six-month USD LIBOR is [-300] basis points. The spread of six-month GBP LIBOR over six-month USD LIBOR is [-100] basis points. For interest rates on currencies other than Euro, JPY, and GBP, the spread over six-month USD LIBOR is [-200] basis points.9 Where the variable rate is linked to a benchmark interest rate other than the six-month USD LIBOR, a spread reflecting the difference between the benchmark rate and the six-month USD LIBOR (rounded to the nearest 50 bps) will be added. Given the anticipated global transition away from LIBOR, this TMU can be updated to reflect the relevant benchmark replacements (U.S. Secured Overnight Financing Rate (SOFR); U.K. Sterling Overnight Index Average (SONIA); EURIBOR; and Tokyo Overnight Average Rate (TONAR)) prior to the complete phase out, once operationally feasible.

[Debt Conditionality Applicable to Countries that Normally Rely on Official Concessional External Financing]

[High Risk of or in Debt Distress and No Significant Access to International Financial Markets]

13. A performance criterion (ceiling) applies to the nominal value of new non-concessional external debt, [an indicative target] (ceiling) applies to the PV of new external debt, [[a performance criterion] [an indicative target] (ceiling) applies to the nominal value of domestic debt], contracted or guaranteed by the public sector with original maturities of one year or more. The ceiling applies to debt contracted or guaranteed for which value has not yet been received, including private debt for which official guarantees have been extended.

[High or Moderate Risk of Debt Distress and Significant Access to International Financial Markets]

14. A performance criterion (ceiling) applies to the PV of new external debt (i.e., understood as debt in foreign currency,10 [[a performance criterion] [an indicative target] (ceiling) applies to the nominal value of domestic debt], contracted or guaranteed by the public sector with original maturities of one year or more. The ceiling applies to debt contracted or guaranteed for which value has not yet been received, including private debt for which official guarantees have been extended.

15. [An adjustor of up to 5 percent of the external debt ceiling set in PV terms applies to this ceiling, in case deviations from the performance criterion on the PV of new external debt are prompted by a change in the financing terms (interest, maturity, grace period, payment schedule, upfront commissions, management fees) of a debt or debts. The adjustor cannot be applied when deviations are prompted by an increase in the nominal amount of total debt contracted or guaranteed.]

[Moderate Risk of Debt Distress and No Significant Access to International Financial Markets]

16. A performance criterion (ceiling) applies to the PV of new external debt, [[a performance criterion] [an indicative target] (ceiling) applies to the nominal value of domestic debt], contracted or guaranteed by the public sector with original maturities of one year or more. The ceiling applies to debt contracted or guaranteed for which value has not yet been received, including private debt for which official guarantees have been extended. The ceiling is subject to an adjustor defined below.

17. [An adjustor of up to 5 percent of the external debt ceiling set in PV terms applies to this ceiling, in case deviations from the performance criterion on the PV of new external debt are prompted by a change in the financing terms (interest, maturity, grace period, payment schedule, upfront commissions, management fees) of a debt or debts. The adjustor cannot be applied when deviations are prompted by an increase in the nominal amount of total debt contracted or guaranteed.]

[Debt Conditionality Applicable to Countries that Do Not Normally Rely on Official Concessional External Financing]11

18. [[A performance criterion] [An indicative target] (ceiling) applies to the nominal value of total debt (external and domestic) of the public sector.]

19. [[A performance criterion] [an indicative target] (ceiling) applies to the nominal value of short-term debt (external and domestic) of the public sector.]

20. [[A performance criterion] [an indicative target] (ceiling) applies to the nominal value of external debt, [contracted or] guaranteed by the public sector with original maturities of [one year or more]. The ceiling applies to debt [contracted or] guaranteed for which value has not yet been received, including private debt for which official guarantees have been extended.]

Annex III. Borrowing Plan

A. What is a Borrowing Plan?

1. A borrowing (or financing) plan is an integral component of a country’s fiscal program, and hence of the planned macroeconomic policy framework. Regardless of a country’s financing mix, a borrowing plan underlying the financing of a planned fiscal deficit is usually prepared. Its design is determined by several factors, including: the appropriateness of the underlying fiscal deficit from a demand management perspective; the compatibility of the borrowing plan with maintaining debt sustainability over the medium-term; the consistency of the projected trajectory of public investment and savings; and the feasibility of implementing envisaged investment programs given capacity constraints. A description of the key features of a country’s borrowing plan is expected to be included in program documentation (See Annex II).

2. The key features of a borrowing plan depend on a country’s financing circumstances, including the degree of financial market integration and the extent to which the borrowing plan covers typical budgetary financing or financing of public investments deemed critical for growth. Key features include the breakdown of the sources of new borrowings across different characteristics: financing terms (including concessionality mix); maturities and currency composition; and uses of financing (relevant where a significant share of financing is destined to investment projects). Information related to the design of debt conditionality, e.g., indicators of PV of debt or indicators of currency composition (where debt limits are used to address vulnerabilities from exposure to currency risks) would also be relevant in assessing the implementation of the borrowing plan.

3. The borrowing plan to be disclosed in program documentation would need to preserve the confidentiality nature of some information. Adequate flexibility and discretion should be provided to avoid unduly limiting the country’s ability to secure the most favorable terms on current and future debt negotiations. While disclosure of specific financing terms is not required, some sensitive information could be derived in cases where disclosure of concessionality requirements (e.g., the PV of debt) is a key element in the borrowing plan and where the borrowing plan only includes a small number of loans. In this case, the level of disaggregation of the information should be carefully tailored to avoid unintentional disclosure of sensitive information.

B. Role in Setting Debt Limits

4. As discussed in Section IV, a borrowing plan is a key component in deriving quantitative debt limits. It plays an important role in the process of assessing and deriving the appropriate level of borrowing underlying a debt limit in any country.

5. Borrowing plans would also have a role in the assessment of implementation of debt conditionality in program reviews. A nonobservance of debt conditionality would require an assessment of the circumstances leading to it (e.g., whether there was a change in the projected financing mix or the level of new borrowing accommodated under the debt limit). To this end, depending on the specific circumstances, the assessment of the implementation of the components of the borrowing plan would help in determining the cause of the nonobservance and point to modifications needed to the program.

C. Presentation of Borrowing Plans in Program Documents: Illustrative Examples

6. The examples below provide an illustration of the most general cases that can be identified depending on the financing characteristics: countries with limited or no access to concessional financing and countries relying primarily on concessional financing. These are not meant to be exhaustive.

Countries that Do Not Normally Rely on Official Concessional Financing

7. Example A illustrates the borrowing plan of a country in which (i) there is no project financing and all borrowing is untied for budgetary purposes only; (ii) has no access to concessional financing; and (iii) public debt vulnerabilities are primarily related to large rollover needs. Hence, the debt conditionality is designed to address rollover risks by constraining the issuance of short-term debt. Consistent with the characteristics of the country the borrowing plan and debt limits are presented in nominal terms.

Example A. Summary Table on Public Borrowing Program

article image

8. Based on this borrowing plan staff can assess whether the realized pattern of debt accumulation was consistent with program expectations regarding constraining rollover risks and ascertain the implications for the program if there were any deviations. Similarly, where public debt vulnerabilities are related, e.g., to foreign currency borrowing, a breakdown between FX and local currency issuance would be shown in the summary table.

Countries that Normally Rely on Concessional Official External Financing

9. Example B below provides an illustrative borrowing plan of a country: i) relying primarily on concessional external financing with some non-concessional loans foreseen, and where ii) the bulk of financing is used for financing infrastructure projects with only a small portion going directly to budget financing (in this case representing a sovereign bond issuance); and iii) no domestic financing is envisioned. In this case, the external borrowing plan of the country gives a fairly detailed breakdown of the sources of external financing and its main uses, key features for the design and assessment of debt conditionality in this country. As discussed in paragraph 3, when the borrowing plan entails only one loan under one specific category, sensitive information about financing terms could be unintentionally disclosed. To avoid this risk, the presentation of the borrowing plan in program documents could remove the creditor by creditor details of the PV of debt and present only the aggregate PV values.

Example B. Summary Table on External Borrowing Program

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.

Annex IV. Technical Considerations in Setting and Monitoring Debt Limits

A. Issues Relevant for Countries that Normally Rely on Official External Concessional Financing

Calculating the Present Value (PV) and Grant Element (GE) of External Debt1

1. For the purpose of setting and monitoring debt limits, the following will apply:

  • The calculation of the PV and GE will be based on the Fund’s concessionality calculator.

  • A single discount rate is used and set at 5 percent2. The level of the discount rate was kept at 5 percent in the 2017 LIC-DSF review. 3

  • The calculation of the PV and GE will be based on the loan amount contracted in a given year. Specifically, it will be assumed that all new loans contracted are fully disbursed at the time when they are contracted. A loan, therefore, will contribute to an annual PV target on debt accumulation only in the year during which it is contracted. The relevant face value for the PV and GE calculations should reflect the net financing provided to the sovereign for project costs in the transaction, which cannot include amounts associated with financial costs such as capitalization of interest or of other fees and charges. This would unduly inflate the face value of the loan and therefore generate a higher GE.

  • For loans with a grant element equal to or below zero, the PV will be set equal to the face value of the loan.

  • For loans carrying a variable interest rate, a “program reference rate” will be specified in the program documents (e.g., the TMU). This rate would be based on staff’s “average projected rate” for the six-month USD LIBOR over the following 10 years.4 The average projected rate will be updated semi-annually on the Fund’s concessionality calculator, based on the spring and fall vintages of the World Economic Outlook (WEO). The program reference rate would be set equal to the most recently available average projected rate and will be fixed until the subsequent program review is conducted to avoid a missing of the debt performance criterion purely due to an unforeseen change in the variable interest rate. . The PV of the loan would be calculated using the program reference rate, plus the fixed spread (in basis points) specified in the loan contract. Where the variable rate is linked to a benchmark interest rate other than the six-month USD LIBOR, a spread reflecting the difference between the benchmark rate and the six-month USD LIBOR (rounded to the nearest 50 bps) will be added.

Assessing the Concessionality of Financing Packages

2. When a financing package that is composed of more than one financing instrument (typically non-concessional loans combined with monetary grants) is proposed, Fund staff may assess on a case-specific basis whether an envisaged combination of financing instruments can be treated as a package for purposes of meeting conditionality under Fund-supported programs.5 A number of elements are taken into consideration to determine whether a financing package in any particular case can be regarded as an integrated incurrence of debt for the purposes of the debt ceilings from the legal and economic points of view. The list below illustrates several of the most important taken into consideration to support the determination:

  • Identical intended use or purposes for the financing (i.e., two financing instruments are not supposed to finance separate projects);

  • Disbursement and repayment schedule in line with the expected concessionality (e.g., Interrelated schedules for disbursement);

  • Cross-conditions for:

    • entry into legal effect (for example, whether the contracting or guaranteeing of the debt is conditional upon the provision of the grant);

    • availability of funds (for example, whether the availability of loan disbursements is contingent on release of scheduled grant disbursements); and

    • default (for example, whether default in one of the contracts is considered default on other contracts); and

    • Identical parties (i.e., donors and recipients) to the financing.

3. Once the components have been determined to constitute an integrated package of debt, the overall concessionality of the package is calculated using the weighted average of the grant elements of its various components. These weights would be based on the nominal value of the loan amounts. If, however, the financing package is not found to be integrated, the concessionality of each element of the package will be assessed individually.

4. Financing packages that include the provision of a grants in kind would be treated as non-concessional (with zero grant element).6 Without international competitive bidding in which multiple competitors participate, or neutral and fair evaluations of grants in kind conducted by third-party technical consultants, it is very difficult to assess the fair value of grants in kind (e.g., prices of equipment, machines, or engineering service). Note that these guidelines relate only to cases where grants in kind form part of blended financing packages. Staff is also not in a position to verify whether the concessional element is undermined by over-invoicing in other dimensions of a broader deal.

Assessing Financing Involving Collateral Unrelated to the Transaction

5. The key distinction for classifying collateralized debt as concessional or non-concessional is whether the collateral is related or unrelated.7 A debt instrument is collateralized when the creditor has rights over an asset or revenue stream that would secure repayment of the debt if the borrower defaults. The use of collateral can presents an evaluation challenge: the standard method of estimating the concessionality of a loan relies only on the nominal interest rate charged, and does not take account of the value of the collateral pledged, which may reduce risks to the lender and thereby lower the nominal lending rate. To determine whether a collateralized loan can be treated as concessional for the purpose of the DLP, the key distinction is whether the loan is collateralized on an asset or future revenue flow that is related to the original transaction (“related collateral”) or if it is collateralized on an unrelated asset or future revenue flow (“unrelated collateral”).

6. Related collateral transactions are standardized and straightforward to compare and evaluate, and can thus be treated as concessional. In related collateral transactions, the asset financed or the future flow of revenue it is expected to generate serve as the collateral. The two most common types of related collateral transactions are trade and project finance. In both cases, collateralization is standard practice, and the comparison with noncollateralized loans is straightforward: the difference is the reduced risks to the lender that collateral provides, which should in principle be reflected in a lower nominal lending rate. Furthermore, no evident debt vulnerability is created, since the loan is directly tied to a transaction to finance the purchase of a new asset or to generate a new revenue stream that is used to ensure repayment of the loan.

7. Unrelated collateral transactions are complex, non-standardized, difficult to compare and evaluate, and should thus be treated as non-concessional (with zero grant element). An example of an unrelated collateral transaction is a loan used to finance current expenditures which is collateralized on government revenues from the export of a commodity. Unrelated collateralized transactions are complex and have several particular features which makes them very difficult to compare, including differences in enforceability, which depend on the nature of the collateral (e.g., whether it is a physical or a financial asset, whether it is in the form of the balance in an escrow account or a security assignment deed, whether the collateral is priced in a hard currency, and how the price of the collateral varies in response to various shocks), the contract conditions (e.g., whether it requires keeping the value of the collateral constant and thus margin calls are involved), and other transaction costs. The multiple dimensions of unrelated collateralized transactions imply they need to be assessed on a case-by-case basis, and that there is no simple way to compare them with noncollateralized transactions. This is the main reason for treating them as non-concessional (with zero grant element). In addition, unrelated collateral increases debt vulnerabilities, as it can increase credit risks for new loans that are not collateralized, tending to raise the cost of such lending.

8. To assess whether a collateralized loan is related or unrelated, conditionality may need to established as a prior action before arrangement approval to ensure proper disclosure of loan agreements prior to the start of the program. The cases of Angola’s 2018–20 EFF and Ecuador’s 2019–21 EFF illustrate how a combination of prior actions and structural benchmarks can be established to ensure adequate disclosure before arrangement approval of loan agreements which would allow assessing the nature of collateralized and collateralized-like transactions, and determining whether collateralization is related or unrelated.

Treatment of Grants in Setting Debt Conditionality

9. The inclusion of grants in a PV target would not affect the PV limits given that the present value of grants is zero. In theory, the change of an envisaged concessional loan into grants could create some space to contract more non-concessional loans, but grants tend not to be substitutes for concessional loans in terms of size and targeting sector.

Addressing Uncertainties in the Projection of Present Value of Debt

10. Uncertainty about financing terms of loans could introduce uncertainty on the projected present value of new debt to be contracted. In many instances, the precise financing terms of loans assumed in a borrowing program may still be under negotiation at the time program conditionality is being set. This creates uncertainty about the actual PV of a particular loan that, even without any change in the total nominal loan amount, could generate deviations from the originally-envisaged PV debt target. Given this uncertainty, the use of an adjustor on debt conditionality set in PV terms may be warranted. Recognizing that deviations may come from one large loan or a few number of small/medium loans, program conditionality could accommodate moderate changes to the debt limit set on PV terms and use an adjustor of up to 5 percent of the new borrowing in PV terms (i.e., the envisioned debt limit) only when deviations are prompted by an unexpected change in the financing terms of a loan or loans.

B. Issues Relevant for All Countries

Effectiveness of a Loan

11. For the purposes of setting and monitoring debt conditionality, the assessment of when a loan becomes effective will be based on country-specific definitions, reflecting the application of the national decision-making processes and when all required contractual steps to the loans effectiveness have been completed. For example, in some countries, the relevant date is when a loan is approved by the highest relevant decision-making unit in the government, while in other it is when a parliament approves the loan. Since country practices differ on the procedures to be followed in regard to the contracting of public external debt, program documentation (e.g., TMUs) should make every effort to include a specific definition as to the precise stage in these processes at which, for program purposes, the debt is viewed as having been contracted. If no specific definition is included in the program documents, a debt will be deemed contracted in accordance with the terms of the contract and as determined by the law applicable to such contracts.

Treatment of Credit Lines

12. Credit lines with uncertain disbursement schedules or allowing for multiple disbursements should be included in debt conditionality on a contracting basis.8 Credit lines usually make available funds that can be disbursed at any time upon the borrower’s demands. As such, the disbursement schedule is not determinable at the time debt conditionality is set. In such cases, staff is expected to discuss with the authorities the expenditures to be financed by the credit lines up to its expiration date. When disbursements can take place over multiple years, annual sub-limits on maximum disbursements under the credit line could be established based on the most likely disbursement schedule.

Treatment of Trade Credits

13. Short-term trade credits are frequently excluded from debt limits on the grounds of facilitating trade. For example, the TMU language on debt conditionality could exclude suppliers’ credits less than 90 days. Nevertheless, when considering an exclusion teams should carefully assess the potential for fiscal risks that may arise from such forms of financing and ensure that it would not open the possibly for exclusion of debt limits.

Treatment of Pre-Financing Arrangements

14. Teams may want to introduce targeted conditionality on pre-financing arrangements where domestic suppliers take out a bank loan that would subsequently be transferred to the government after disbursement. Although not a widespread practice, pre-financing could be used to circumvent other debt limits e.g. NCB limits). Thus, a continuous zero limit on such arrangements could help mitigate risks when they are a concern. When such conditionality is deployed, TMU language should detail the specific type of arrangement that is being limited and the parties involved (including government sub-sectors).

Treatment of Standby Letters of Credit

15. Teams should consider the interaction between standby letters of credit and any debt conditionality (e.g., NCB limits) carefully. Standby letters of credit that are undrawn are contingent liabilities9 and are thus not “debt” as usually defined in a TMU.10 However, any amounts drawn under standby letters of credit will constitute debt, and should be included in the DSA, thus informing the calibration and/or design of debt limits at subsequent program reviews. Amounts drawn under standby letters of credit, while constituting debt, may be excluded from certain debt limits (e.g., NCB limits) through language in the TMU on a case-by-case basis.11 This could include, for example, standby letters of credit that are issued for the purposes of guaranteeing payment obligations for the purchase of power/electricity, under Power Purchase Agreements (PPAs) in the context of a PPP agreement. Teams should consult SPR and LEG regarding the treatment of such arrangements.

1

For background on the evolution of the DLP, see Review of the Policy on Debt Limits in Fund-Supported Programs.

3

The IMF and World Bank provide a regular update of debt limits on their external websites to support creditors that use it as a reference framework. The table on the IMF website is available here.

9

As part of the reform of the IMF’s MAC DSA framework, it has been renamed to the Sovereign Risk Debt Sustainability Framework (SRDSF). See Review of the Debt Sustainability Framework for Market Access Countries (2020). The SRDSF is expected to become effective in Q4 2021 or Q1 2022.

12

These activities could include public debt-creating activities recorded below-the-line (such as government lending programs or equity injections); cases where government takes on substantial contingent liabilities (issuance of government guarantees, SOE debt), or where the fiscal accounts do not record operations such as debt assumptions or guarantee calls above the line

13

The approach to conditionality adopted by country teams should reflect an assessment as to the key fiscal variables that need to be monitored, coupled with a pragmatic assessment as to the quality, timeliness, and the adequacy of coverage of fiscal operations (see GFSM 2014). In cases where country teams propose to change the specification of a performance criterion or indicative target from that used previously, the program documentation should explain the factors justifying this change.

14

See Guidance Note on the Bank-Fund debt Sustainability Framework for Low Income Countries, ¶7–9 for details on the criteria for countries to use the LIC-DSF.

15

See Annex III for further details on the requirements for a Borrowing Plan.

16

For a full discussion on the concepts of debt and contingent liabilities, please see 2014 Government Finance Statistics Manual (particularly Section 7.236–7.271 and Appendix III) and the External Debt Statistics: Guide for Compilers and Users.

17

In contrast with the standard overnight or short-term repos used for liquidity management purposes with AAA assets posted as collateral, the repos referred to are vulnerability-enhancing. They have longer maturities and the collateral in the debtors’ own bonds is a multiple (e.g., double or triple) of the loan value. Such repos can accentuate the procyclicality of financing needs owing to the collateralization in debtor’s own bonds and margin call requirements.

18

The magnitude of the debt residual in the historical debt changes decomposition is an indicator of changes in debt not accounted for by identified debt-creating flows (i.e. the primary fiscal balance, the interest rate/growth differential, exchange rate depreciation, etc.). This residual is presented as a share of GDP and can be compared with the debt-to-GDP ratio to obtain a sense of its relative magnitude. An alternative source for assessing the magnitude of the fiscal cost of contingent liabilities covering a longer period (1990–2014) is the database developed in Bova et al. (2016) “The Fiscal Costs of Contingent Liabilities: A New Dataset,” IMF Working Paper WP/16/14, January 2016.

19

Commercially viable SOEs may be excluded from DSAs and debt conditionality (see ¶41).

21

PPP agreements may represent omitted debt if they impose a non-contingent liability on the government. This would be the case, for example, of a minimum revenue or minimum profit guarantee that is set so high relative to the expected revenue or profit that it ends up being a certain payment rather than a contingent one based on whether certain conditions are met.

23

Even in cases where collateralized debt is not a large share of total debt, collateralized debt can still present a significant debt vulnerability due to the nature of the contractual agreement. For example, some repos can amplify risks even if they are small in relation to total debt because they are overcollateralized using the government’s own bonds and may have margin calls, making them highly procyclical financing instruments.

24

A significant share of debt with such priority increases credit risks on other lending, which raises the potential for a loss of access to uncollateralized financing that also makes it difficult to service existing debt, as discussed in the G20 note on Collateralized Transactions : Key Considerations for Public Lenders and Borrowers.

25

The World Bank is often better placed to address some of the underlying legal and institutional issues and other capacity shortfalls that give rise to inadequate transparency, particularly where there are multiple competing priorities for structural conditionality in an IMF-supported program.

27

While the debt holder profile table requirements focus on the existing debt stock, country teams should also have a good understanding of the size and terms of amounts of debt that are contracted, but not yet disbursed to inform both the DSA and the Borrowing Plan (Annex II).

28

The IMF does not have a list of institutions considered as multilaterals. In practice, the Fund’s judgment on whether an institution is a multilateral creditor is based on a number of factors, including: (i) global, rather than regional, membership of the institution; (ii) Paris Club’s treatment of the claims of the institution, and the institution’s participation in the Paris Club; and (iii) the treatment of the institution under the HIPC Initiative. Plurilateral institutions are defined as those creditors composed of two or more official creditors which do not meet the multilateral criteria. If the Fund adopts a new definition of multilaterals as part future policy reviews, the classification of institutions would be adapted accordingly if needed.

29

If there is no debt in any of these major categories it would be useful to report a 0 figure rather than omit the category. If a country has external debt that is outside these categories, a suitable new category should added.

30

Staff may not encourage the authorities to breach terms in their contracts.

31

Under the Transparency Policy, the Managing Director will generally not recommend that the Executive Board approve a request for use of Fund resources unless that member explicitly consents to the publication of the associated staff report. The Transparency Policy also strictly limits the conditions (such as market sensitivity) under which deletions of non-public information may be made prior to publication.

32

As part of efforts to promote debt transparency, the Executive Board is seeking clarity on the extent of confidentiality clauses in sovereign debt contracts.

33

For the purposes of the estimate, such clauses would include those relevant for completing the debt holder profile table. Such clauses could include, for example, those limiting or preventing disclosure of the existence of debt, the amount of the debt, creditors, or key terms regarding the debt profile (e.g., currency). Confidentiality clauses relating to commercially sensitive information for the debtor and creditor, but that do not affect reporting the debt stock in the debt holder profile table (e.g., interest rate), would not generally be expected to be discussed.

34

This expectation assumes a semi-annual review cycle.

36

While the DSA is the main tool for assessing debt vulnerabilities and the one that determines conditionality requirements, teams may consult other complementary tools to deepen their understanding of debt vulnerabilities. For example, structural models like the DIGNAR-19 maintained by RES/ICD, can enable a quantitative assessment of macroeconomic shocks, with an emphasis on growth and debt.

37

While the 2013 MAC DSA does not have a specific risk of debt distress rating, the judgment on the extent of debt vulnerabilities is informed by a set of tools provided within the MAC DSA framework; specifically, it is based on a set of benchmarks related to the debt level, the gross public financing needs and a number of other indicators related to the debt profile.

38

See “Overall Trends in DLP Implementation” and “Experience with Implementation of the DLP” in Reform of the Policy on Public Debt Limits in IMF Supported Programs (IMF, 2020) for debt limits that were used in PRGT and GRA-Supported programs between July 2015-April 2020, including those aimed at specific vulnerabilities.

39

The LIC-DSF is described in IMF and World Bank (2018).

40

See Reform of the Policy of Public Debt Limits (2020) for further discussion as well as a list of previous IMF-supported programs that used PV-limits from July 2015- April 2020.

41

See World Bank Debt Reporting Heat Map. Under the World Bank Debt Sustainability Enhancement Program (DSEP): (i) all IDA countries will be screened annually for debt-related vulnerabilities in order to identify those countries for which concrete performance and policy actions (PPAs) will need to be defined and implemented. For IDA countries identifying through the screening process, PPAs will be established and updated each year to help identify and prioritize areas critical in addressing debt vulnerabilities, such as strengthening debt transparency, fiscal sustainability, and debt management. The design of PPAs will be informed by indicators in the Debt Reporting Heat Map.

42

See also Box 3 if the “in debt distress” rating is due to ongoing or impending debt restructuring negotiations.

43

The choice between an IT or PC should follow the principles outlined in Section 4.D.

44

See “Overall Trends in DLP Implementation” and “Experience with Implementation of the DLP” in Reform of the Policy on Public Debt Limits in IMF Supported Programs (IMF, 2020) for discussion of past applications of the NCB exception policy.

45

For a in debt distress country, market financing would not normally be a part of the borrowing plan. See also Box 5 if the “in debt distress” rating is due to ongoing or impending debt restructuring negotiations.

46

Both the significance and stability of foreign investor participation would need to be considered in this case. For example, if private investors largely exit from the domestic bond market in bad times, such inflows would not be viewed, at that juncture, as a significant source of stable funding.

47

For guidance on assessing the fiscal risks posed by PPPs, see Irwin et al. (2018), “How to Control the Fiscal Costs of Public-Private Partnerships,” International Monetary Fund, How-To Note 18/04, October 16, 2018.

48

In the 2013 MAC-DSA, risks stemming from debt structure and debt service obligations can be gauged from the relevant cells in the heatmap. Risks stemming from high debt levels are also mostly captured by the fanchart.

49

See also Annex IV.B for specific coverage of debt limits issues including treatment of IMF financing, credit lines, trade credits, and pre-financing arrangements. Treatment of World Bank financing is also discussed in Section V.D.

50

A comprehensive range of public debt instruments is provided in Government Finance Statistics Manual 2014 (GFSM 2014) and the Public Sector Debt Statistics: Guide for Users and Compilers (IMF 2011).

51

These would include central bank liabilities such as issuances of treasury securities.

52

See Review of the Debt Sustainability Framework for Market Access Countries—Annexes. Liquidity papers that are issued solely for monetary policy purposes would normally be excluded from the debt definition used for the DSA, provided (i) no financing to the government can be provided through their issuance; (ii) the government is not de facto responsible for paying debt service thereon; and (iii) the securities do not represent a material fiscal risk (as indicated, for example, by a track record of central bank independence and monetary stability). Where one or more of these conditions is not met, liquidity papers would be included in public debt and GFNs for DSA purposes unless their outstanding stock can be deemed de minimis.

55

For more detail, see Appendix III of Guidance Note on the Bank-Fund DSF for LICs

56

The authorities can propose adjustments to their borrowing plan at program reviews, providing them with the flexibility to modify their investment plans if, for whatever reason, proposed projects are dropped or substantially delayed.

57

Using the signal approach there would likely be a strong basis for an NCB exception. A project being undertaken by an MDB is a signal that a project is critical for national development. The World Bank would also be able to provide that alternative concessional financial is not available. Excluding such financing from the debt limit via specific language in the TMU would be permitted, but the same criteria must be met.

58

This may differ from the residency definition used in the Balance of Payments and International Investment Position Manual 6th Edition and other statistical manuals. The definition of external debt in the context of debt conditionality may also differ from the LIC-DSF, given the importance of ensuring the authorities capacity to monitor and control such debt. For instance, although there may be sufficient information to define debt on a residency basis in the LIC-DSF, if the authorities are unable to closely track nonresident holdings of domestic debt, a currency-based definition may be preferred in order to avoid misreporting concerns.

59

Both tools can be accessed via the IMF Debt Limits Policy page. A more detailed version of the PV tool is also available to country teams internally.

60

For nominal non-zero NCB limits and PV-limits, exceptions would not apply. Critical projects should be part of the authorities borrowing plan, which would inform the quantitative calibration of such limits (see Section IV. B).

61

The use of non-standard financing packages (described in Section IV,B and Annex IV would not necessarily preclude the use of an NCB exception.

62

Debt management operations refer to asset and liability operations typically performed by debt management offices with the objective of improving the debt profile (e.g. reducing interest costs by refinancing existing debt, reducing rollover risks by prefunding maturing obligations, or reducing refinancing and interest rate risks by improving the maturity profile of the debt stock).

63

See “Alignment between the IMF’s DLP and World Bank’s SDFP” in Reform of the Policy on Public Debt Limits in IMF Supported Programs Supplementary Information (IMF, 2020) for more details.

64

For countries that do not normally rely on concessional financing, the calibration of limits, if any, would be supported by the IMF’s MAC DSA for the purposes of the DLP, and the World Bank’s analysis of debt vulnerabilities for the purposes of the SDFP.

65

Countries that do not normally rely on concessional finance or countries that normally rely on concessional finance with significant access to international financial markets.

1

It also provides guidance on the TMU language for setting assessment criteria and other targets on debt in PSIs and PCIs

3

[For background (not to be included in TMUs): According to GFSM 2014, government control of an entity can be established based on any of the following criteria: (i) ownership of the majority of the voting interest; (ii) control of the board or other governing body; (iii) control of the appointment and removal of key personnel; (iv) control of key committees of the entity; (v) golden shares and options; (vi) regulation and control; (vii) control by a dominant public sector customer or group of public sector customers; or (viii) control attached to borrowing from the government (see GFSM 2014, Chapter 2, Box 2.2 for further details).]

4

This definition may differ from that used in other sections of the TMU, including regarding fiscal targets.

5

[For background (not to be included in TMUs): The decision to include any financial instrument (including swaps or repo agreements) and/or other liabilities of the central bank or other public financial institutions for the purposes of public debt conditionality should be based on whether they present a fiscal risk, and excluding them would undermine the effectiveness of debt limits in addressing debt vulnerabilities. Central bank debt issued solely for monetary purposes is excluded from the definition of public sector debt. Similarly, the external debt of the central bank should be included under public external debt limits if such debt contributes to external debt vulnerabilities and associated fiscal risks. If the central bank also acts as the agent of the government and issues treasury bonds on behalf of the government, such issuance would constitute public debt. Similarly, government borrowings from the central bank should normally be included in public debt.]

7

[For background (not to be included in TMUs): this definition applies where debt limits are set on a contracting basis. For debt limits that are set on a disbursement or stock basis, such definition is not needed in the TMU. Instead it should be stated clearly in the TMU that the performance criterion or indicative target (ceiling) applies to the disbursements or stock on the appropriate type of debt (e.g., external public and publicly guaranteed debt).]

8

The calculation of concessionality takes into account all aspects of the debt agreement, including maturity, grace period, payment schedule, upfront commissions, and management fees.

9

The program reference rate and spreads are based on the “average projected rate” for the six-month USD LIBOR over the following 10 years from the most recent Fall World Economic Outlook (WEO).

10

Reference could be replaced with alternative conditionality if deemed better targeted toward vulnerabilities.

11

[For background (not to be included in TMUs): examples of debt limits that could be used where relevant to the program design. For countries that do not normally rely on official concessional financing, debt limits, if warranted, would be set in nominal terms and could take the form of either limits on total public debt or targeted debt limits, guided by the type of debt vulnerabilities identified.]

1

The present value (PV) of public external debt is equal to the sum of all future debt service payments (principal and interest), discounted to the present using a set discount rate. The grant element (GE) measures the concessionality of a loan, and is defined as the difference between the nominal and the present value, expressed as a percentage of the nominal value.

3

IMF (2017), “Review of the Debt Sustainability Framework for Low Income Countries: Proposed Reforms”.

4

Given the LIBOR is in the process of being phased out, TMU language should include a reference indicating that such rates will be updated to reflect the relevant benchmark replacements (U.S. Secured Overnight Financing Rate (SOFR); U.K. Sterling Overnight Index Average (SONIA); EURIBOR; and Tokyo Overnight Average Rate (TONAR)) ahead of the phase out and once operationally feasible.

5

This is assessed jointly by LEG and SPR based on the underlying financial contracts. A country team encountering potential financing packages is encouraged to consult with LEG and SPR at an early stage of such loan negotiation.

6

Only the loan part, ignoring grants in kind, would be counted toward non-zero or PV limits in line with the grant element of the loan.

7

See Collateralized Transactions: Key Considerations for Public Lenders and Borrowers for a discussion of issues raised by collateral and more examples of related and unrelated collateralized transactions.

8

Credit lines would not include standby letters of credit. In general, standby letters of credit are intended to serve as a guarantee and are thus contingent in nature and cannot be drawn on at a borrower’s discretion. Standby letters of credit should be distinguished from other types of letters of credit, that are not intended to serve as a guarantee. Teams should consult SPR and LEG for assistance in evaluating such documents.

9

See Footnote 5 in “Limits on External Debt or Borrowing in Fund Arrangements – Proposed Change in Coverage of Debt Limits”, EBS/00/128, June 30, 2000.

10

See “Definition of Debt” under Annex B above.

11

Country teams could consider seeking such exclusion on a preemptive basis (i.e., before such standby letters of credit are drawn). For example, this could be considered in cases where a program includes a zero limit on NCB; any drawn amounts under a standby letter of credit will constitute debt. As discussed, such exclusion through language in the TMU will be considered on a case-by-case basis.

Guidance Note On Implementing The Debt Limits Policy In Fund Supported Programs
Author: International Monetary Fund. Strategy, Policy, & Review Department