## REVIEW OF BANK-FUND DEBT SUSTAINABILITY FRAMEWORK FOR LOW-INCOME COUNTRIES: BACKGROUND NOTE

## Source details

**Canonical URL:** [REVIEW OF BANK-FUND DEBT SUSTAINABILITY FRAMEWORK FOR LOW-INCOME COUNTRIES: BACKGROUND NOTE](https://www.imf.org/-/media/files/about/partners/cso/2026-lic-dsf-review-background-note-for-public-consultation.pdf)

## Other formats

- [Markdown version](/-/media/files/about/partners/cso/2026-lic-dsf-review-background-note-for-public-consultation.pdf.md)
- [Structured JSON version](/-/media/files/about/partners/cso/2026-lic-dsf-review-background-note-for-public-consultation.pdf.json)

---

### Background
- The joint IMF-World Bank Debt Sustainability Framework for Low-Income Countries (LIC-DSF) is described as the cornerstone of the assessment of the risk of public debt stress and sustainability in countries eligible for IMF and/or WB concessional financing (IMF and World Bank).
- Uses and linkages:
  - Used for macroeconomic analyses, policy advice, application of internal policies.
  - Linked to lending policies of other multilateral development banks, including the African Development Bank.
  - DSAs provide early warning signals on build-up of risks of public debt stress and have been used to determine restructuring envelopes, notably under the G20 Common Framework.
- Evolution:
  - Launched in 2005 and subject to periodic reviews to adapt to evolving debt risk landscape.
  - The 2017 review:
    - Reduced the number of thresholds and standardized stress tests.
    - Expanded debt-carrying capacity to include macroeconomic fundamentals.
    - Strengthened assessment methodology for overall public debt risks.
    - Introduced realism tools, tailored stress tests, a market-financing module, and enhanced guidance on structured judgment.
  - 2017 framework: classifies countries into three debt-carrying capacity groups: Weak, Medium, and Strong; mechanical risk assessment compares debt-burden indicators to thresholds calibrated from past predictive ability; structured judgment and judgment-based sustainability assessments remain central.

### Context for the current review
- Drivers:
  - Over the last 15 years, overall public debt increased markedly as a share of GDP across all LICs.
  - Since the COVID-19 pandemic, LICs faced relatively diminished access to external financing, prompting greater reliance on domestic financing (IMF and WB, 2025).
  - Frontier and relatively wealthier LICs have increasingly borrowed on non-concessional terms, narrowing the gap between present and face value of overall public debt.
- Emerging challenges identified:
  - Absence of a fully calibrated model of overall debt stress and insufficient consideration of domestic debt vulnerabilities risk under-capturing public debt stress.
  - Lack of a clear mechanical signal for debt sustainability could impede consensus during restructurings.
  - Increased heterogeneity across LICs strains DCC-specific thresholds’ ability to capture repayment capacity differences.
  - Single time horizon for stress assessments reduces information content and complicates cross-country comparisons.
  - Realism tools’ effectiveness may be strained as forecast uncertainty rises with more frequent and larger shocks.
  - Differences in debt coverage raise uniformity-of-treatment concerns.

### Identifying key reform areas (analysis and performance)
- Predictive performance observations:
  - LIC-DSF has flagged debt distress episodes ahead of time; all cases of new “In Debt Distress” ratings since rollout in 2018 were preceded by a “High-Risk” rating in a recent DSA.
  - With over one-third of LIC-DSF users rated at high risk in a given year, the up-to-two-years-ahead probability of debt distress conditional on a “High-Risk” rating has been around 5 percent.
  - Of the latest LIC-DSAs with “High-Risk” external debt distress ratings, 15 out of 16 countries with near-to-medium-term (i.e., 1-to-5 years ahead) debt vulnerabilities are implementing reforms supported by an IMF program.
  - Large undifferentiated “High Risk” group, sometimes driven by long-term risks, complicates interpretation and may overestimate near-term risks; mechanical reliance on LIC-DSF risk ratings could influence lenders and access to financing.
- Specific pressure points:
  - Performance and robustness: out-of-sample performance aligns with in-sample calibration, but re-estimation with updated data suggests resilience needs strengthening.
  - Reflecting heterogeneity: DCC grouping remains useful; “cliff effects” from DCC changes have not significantly impacted final ratings but may cause discontinuities in fiscal advice; Frontier LICs’ DCC generally aligns with market credit signals, but heterogeneity suggests further analysis of DCC variables/modalities.
  - Domestic debt vulnerabilities: need to strengthen overall public debt distress assessment given rising domestic debt importance.
  - Judgment-based assessments: improved outcomes for non-representative LICs, but contributed to higher false positives by reflecting long-term considerations without time differentiation of risk ratings; sustainability assessments would benefit from model-based signals.
  - Forecast accuracy: short- and medium-term projections fairly accurate though COVID-19 complicates conclusions; long-horizon forecasts show some optimism bias in exports and revenue forecasts.
  - Debt coverage: cross-country differences in debt data completeness, including SOEs, remain a challenge.
  - Climate risks and policies: coverage evolved but remains uneven; additional tools needed.
  - Realism tools and stress tests: can be further strengthened and retooled.
  - Continuity at graduation: two separate frameworks (LIC-DSF and MAC-SRDSF) risk discontinuity for graduating countries; further alignment would mitigate this risk.

### Potential reforms and proposed design changes
- Overarching objective: strengthen and future-proof the framework while keeping it tractable and easy to use.
- Modeling architecture enhancements:
  - Re-estimation and recalibration of the external debt stress model.
  - Development of a new model and mechanical risk signal for overall debt stress, with new solvency and liquidity debt burden indicators.
  - Updating the measurement of the Composite Indicator of Debt-Carrying Capacity (CI-DCC), calibrated to better reflect LICs’ heterogeneity.
  - Introducing a new module for analysis of domestic debt vulnerabilities.
- Mechanical signal and auxiliary indicators:
  - New model-based mechanical signal for overall public debt sustainability.
  - Judgment in sustainability assessment supported by auxiliary indicators of near-to-medium-term fiscal effort: distance between the primary balance and the debt-stabilizing primary balance, and magnitude of persistent breaches of debt-burden indicators in stress models.
- Long-term and data-confidence changes:
  - Introduce a new long-term module to model impacts of development challenges, including climate, and policies to address them.
  - Introduce a confidence flag in debt data to incentivize improvements in transparency, reliability, and coverage.
  - Introduce granularity of the “High-Risk” ratings by time horizon and type of risk.
  - Update and enhance realism tools and stress tests.
- Simplifications:
  - Propose a reduction in the number of debt-burden indicators used in the external public debt risk assessment.
  - Enhance stress tests.

### Conceptual framework refinements
- Terminology changes:
  - Adopt the term “risk of public debt stress” to refine the previous shortcut “risk of debt distress”.
  - Definitions proposed:
    - Risk of public debt stress: materialization of acute fiscal and/or market pressures related to public debt, with no presumption about modes of resolution (policy adjustments, external financing support, and/or debt restructuring).
    - Risk of unsustainable debt/public debt sustainability: for PRGT and IDA-eligible countries, ability to meet current and/or future financial obligations while preserving growth and making adequate progress toward development goals (IMF and WB, 2017). Public debt is unsustainable when no politically and economically feasible policy set reasonably prospects to resolve public debt stress without debt restructuring (IMF, 2024).
  - Terms “risk of unsustainable debt” and “public debt sustainability” remain unchanged in meaning relative to current LIC-DSF.
- Emphases:
  - Public debt stress can have solvency and liquidity dimensions.
  - Assessment remains forward-looking and informed by judgment.
  - Clarifying terminology aims to reduce misinterpretation of signals as implying unsustainability when they reflect broader stress-risk continuums.

### Framework architecture: models, CI-DCC, and DBIs
- Three interconnected models:
  - (i) risk of external public debt stress (re-estimated),
  - (ii) risk of overall public debt stress (new),
  - (iii) risk of public debt unsustainability (new).
- All three models use PROBITs with one solvency and two liquidity debt-burden indicators (one per PROBIT), and a common Composite Indicator of Debt-Carrying Capacity (CI-DCC).
- Dependent variables:
  - Criteria identifying onset of public debt stress refined and expanded.
  - Unsustainable debt events identified as a subset of severe public debt stress events that cannot be resolved without debt restructuring.
  - Overall public debt stress and unsustainable debt criteria aligned more closely with MAC-SRDSF.
- Debt-burden indicators (DBIs) changes:
  - Proposal to drop PV of external debt-to-exports from external DBIs.
  - Overall DBIs better aligned with MAC-SRDSF:
    - Overall public debt taken at face rather than present value in ratio to GDP.
    - Gross financing needs-to-GDP (GFN-to-GDP) replacing overall public debt service-to-revenues.
    - Interest payments as ratio to revenues excluding grants to replace public debt service-to-revenues.
  - DBIs for public debt unsustainability model will be a subset of indicators used in external and overall stress models.
- CI-DCC refinements:
  - CI-DCC remains a weighted average of macroeconomic indicators capturing debt-carrying capacity.
  - World growth removed from CI-DCC due to loss of statistical significance.
  - CI-DCC to include:
    - Institutional quality proxy: WB Country Policy and Institutional Assessment (CPIA) index.
    - Domestic economic fundamentals: real GDP growth.
    - External buffers: reserve coverage and remittances as ratio to GDP.
    - Measures of exposure to global macro-financial conditions and exogenous shocks: output volatility; interaction of interest rate on newly contracted external loans and share of external in overall public debt.
  - Higher cost of financing and greater exposure to global macro-financial conditions require concurrent strengthening of institutional frameworks, macro-fundamentals, or external buffers to maintain the same CI-DCC score.
- DCC classification:
  - Countries assigned to one of three DCC groups: Weak, Medium, and Strong.
  - CI-DCC calculated using centered averages of DCC variables based on latest five years of historical data and the first five years of projections.

### Mechanical risk signals and calibration
- Signal generation:
  - Mechanical signals generated by breaches of DCC-specific thresholds for DBIs in the baseline and stress scenarios over the first ten years of the forecast horizon.
- Risk categories:
  - "Low Risk" — no sustained breaches of DCC-specific thresholds in either the baseline or stress scenarios.
  - "Moderate Risk" — sustained breaches of DCC-specific thresholds occurring only in stress scenarios.
  - "High Risk" — sustained breaches of DCC-specific thresholds in the baseline.
- Calibration and use:
  - External thresholds re-calibrated and new thresholds for overall DBIs selected to achieve predictive power broadly comparable to that at time of 2017 review.
  - Sustained breaches defined as occurring more than once over the forecast horizon.
  - Final risk rating generally informed by the more conservative of mechanical signals from external and overall public debt stress models:
    - External model captures continued importance of external debt for LICs.
    - Overall model captures domestic debt vulnerabilities where relevant.

### Public debt sustainability and the unsustainability model
- New unsustainability model:
  - Formal model for risk of unsustainable debt, improving reliance on judgement.
  - Similar structure to stress models; probability of unsustainable debt determined by DBIs and CI-DCC.
  - Risk flags derived by comparing predicted probabilities to selected probability cut-offs.
  - Calibration targets a lower false positive rate (FPR) than in the stress model to avoid triggering self-fulfilling loss of financing.
  - Final sustainability assessment informed by mechanical signal and judgement, supported by indicators of fiscal effort required to stabilize overall public debt dynamics and severity of projected persistent threshold breaches in stress models in near- to medium-term.
- Mechanical signal for sustainability:
  - Only generated when final rating of public debt stress is "High Risk" or "In Distress".
  - Signal categories:
    - "Unsustainable" — when country-specific predicted probabilities from any of the three regression specifications equals or exceeds applicable probability cutoff.
    - "Sustainable" — otherwise.

### Auxiliary modules: Domestic Debt and Long-Term Modules
- Domestic Debt Risk Module (new):
  - Purpose: track domestic public debt risks and support differentiated analysis of domestic debt vulnerabilities.
  - Components:
    - Core diagnostics: recent developments and projections of domestic public debt-to-GDP and domestic public debt service-to-revenue (excluding grants) ratios; comparisons to medians of projected values across LIC-DSF users and to LICs that experienced domestic debt stress over past decade.
    - Classification: countries categorized as low-, closer-, or heightened-scrutiny depending on whether forecasts are below or above medians across LICs and whether they exceed cross-country medians observed in past stress events.
    - Realism tools: benchmark net domestic public debt issuance and financial sector exposures to sovereign against distributions across all LICs; consider market pressures in domestic government bond markets (e.g., changes in demand across tenors).
  - Use in judgement:
    - Domestic debt vulnerabilities serve as qualitative input into judgement for overall public debt risk and sustainability assessments.
    - Concerns about domestic public debt trajectories or borrowing plan consistency with macro-financial stability inform final ratings and sustainability assessments.
  - Notes on benchmarking and data:
    - Historical realizations and projections for 2018-32 presented for domestic public debt-to-GDP and domestic public debt service-to-revenue.
    - Median pre-stress historical pattern for 2015–24 and medians of average near- to medium-term projections across LIC-DSF users based on latest LIC-DSAs as of December 2025 are used for benchmarking.
    - Borrowing assumptions and shares in new domestic debt issuance shown for medium- and long-term and short-term debt.
- Long-Term Module (new):
  - Purpose: assess public debt stress and sustainability implications of policy and investment decisions associated with development and long-term challenges (including climate adaptation).
  - Approach:
    - Extend projections beyond the 10-year baseline up to a maximum of 30 years and overlay model-generated impulse responses capturing macroeconomic impacts of specific long-term challenges.
    - Allow country teams to model alternative policy scenarios in the first ten years; model long-term challenges and policies beyond first ten years to assess whether projected public debt-to-GDP remains consistent with maintaining public debt sustainability in the long run.
  - Outputs and limits:
    - Produces insights that do not generate mechanical signals but inform policy choices and may be incorporated into final risk and sustainability assessments via judgement.
    - Exercise of judgement must account for large uncertainty and wide confidence bands in long-term projections.
    - Extended projections (up to 30 years) build on 10-year baselines by assuming the economy remains on the implied steady-state path.
  - Guidance:
    - Projections of GFNs-to-GDP ratio facilitate investigation of a sustainable financing mix.
    - New Guidance Note will provide guidance on evaluating dynamic stability of public debt-to-GDP and supporting GFNs-to-GDP ratios.

### Incorporating climate risks and policies in long-term analysis
- Long-term climate module:
  - Two complementary modelling approaches available as standard options; country teams select the most relevant approach.
  - Starting point: climate risk-adjusted projection of public debt-to-GDP ratio.
  - Confidence intervals capture parameter and warming sensitivity in different climate models and different adaptation speeds.
  - Calibrated warming scenarios range from +1.8°C to +4.4°C by 2100.
  - Policy scenarios illustrate trade-offs between higher adaptation spending (increasing debt-to-GDP) and reduced GDP losses from natural disasters relative to a “no-extra-adaptation” scenario.
  - Illustrative outputs include:
    - Public Debt-to-GDP Ratio (Percent) projections.
    - Avoided Income Losses as Result of Additional Adaptation Investment (Percent of climate risk-adjusted GDP).
  - Modelling inputs and methods:
    - Use granular, spatially disaggregated geophysical and sectoral data into sectoral biophysical models, aggregated through a macro-structural model.
    - Pair reduced-form estimates of slow-moving temperature increases with dynamic general equilibrium modelling of recurrent natural disasters and the climate investment-growth-debt nexus.
  - Tailored natural disasters stress test updated to reflect latest country-specific information on effect on debt of extreme climate-related events.

### Development module: modelling long-term development policies
- Purpose: interface between LIC-DSF and quantitative tools assessing impact of policies addressing long-term development challenges (human and physical capital, infrastructure, structural reforms).
- Mechanisms:
  - Development policies can enhance potential growth via productivity improvements, labor force development, and improved resource allocation.
  - Increased public investment financed solely by debt generally worsens debt burden indicators initially; growth dividends can strengthen revenues and partially improve debt dynamics in second-round effects.
- Illustrative scenarios:
  - Human capital: increase in spending on human capital by 1 percent of GDP per year for ten years — initial larger deficits and borrowing needs; gradual improvements in potential output, revenues, and moderation of debt accumulation as human capital improves.
  - R&D: permanent additional government spending of 0.3 percent of GDP that increases total factor productivity by 0.09 percentage points.
  - Female labor force participation: assumed 2.5 percentage points increase through elimination of gender gap in secondary and tertiary education via increased government spending by 0.17 percent of GDP per year until it reaches 1.2 percentage points above the baseline where it is kept.
- Financing modalities:
  - Simulations use a standard mix of concessional and non-concessional financing, and purely market-based non-concessional financing.
  - Non-concessional financing scenarios show a surge in public debt-to-GDP and GFN-to-GDP, underscoring relevance of adequate funding schemes.
- Modelling tools referenced: Sustainable Development Goals Financing Tool; combination of MFMod-GJ and MFMod.

### Auxiliary indicators, data confidence, and debt perimeter
- Proposed auxiliary indicators to inform final assessments:
  - Distance between the primary deficit and the debt stabilizing primary deficit (DSPD) — reflects fiscal effort required to stabilize public debt-to-GDP at recent historical levels.
  - Magnitude of persistent breaches of DBIs used in stress assessments — forecasts persistence of public debt stress beyond five years.
- Confidence flag on underlying debt data:
  - Summarizes debt coverage used in the DSA (whether full public sector perimeter captured).
  - Composite signal based on two indicators of data reliability:
    - Historical debt data revisions between the most recent DSA and earlier vintages.
    - Discrepancies in public and publicly guaranteed external debt data reported in the DSA and in World Bank International Debt Statistics (IDS).
  - Summary of World Bank’s Debt Reporting Heatmap to reflect debt data transparency.
- Debt perimeter objective:
  - Near-complete coverage of external and domestic public and publicly guaranteed (PPG) debt, including central, state, and local governments, social security and extra-budgetary funds, central bank, state-owned non-financial public enterprises (SOEs), and private debt guaranteed by public sector.
  - Principle: include all SOE debt unless SOE poses limited fiscal risks; criteria from 2017 LIC-DSF Guidance Note to be clarified and streamlined.

### Enhanced realism tools and diagnostics
- Retain and update three core realism tools:
  - Decomposition of debt dynamics across DSA vintages to flag implausible shifts in debt-creating flows.
  - Benchmarking planned fiscal adjustment against cross-country and country-specific historical distributions.
  - Consistency checks between fiscal adjustment and growth based on plausible fiscal multipliers.
- Empirical bases refreshed and benchmarking expanded beyond Fund-supported programs.
- New checks to cover projections for growth, exports, revenues (excluding grants), and financing assumptions:
  - Real GDP Growth — compare baseline projections to 10-year historical averages to flag unusually high or low projected growth.
  - Changes in exports-to-GDP and revenues (excl. grants)-to-GDP — flag large projected five-year changes relative to country history and LICs.
  - More systematic scrutiny of financing assumptions, including diagnostics on maturity structure and cost of new domestic borrowing to identify understated gross financing needs.
- Empirical basis: LIC DSAs database and staffs’ calculations; histograms based on all LIC-DSAs over 2015-2024 with bottom and top 2.5% trimmed to exclude outliers. Projected 5-year changes are relative to latest data outturns; latest 5-year changes capture difference between variable outturns in periods t and t - 5.

### Stress testing framework: standardized, tailored, and customized tests
- Structure:
  - Retains standardized, tailored, and customized stress tests; mechanical signals remain for “Low” and “Moderate” risk ratings.
  - Standardized stress tests activated for all LIC-DSF users; tailored stress tests triggered only for relevant countries.
  - Stress tests applied to both external and overall public DBIs; new triggered banking crisis stress test applies only to overall DBIs (modelled assuming new debt created is financed domestically).
  - Recalibration and updates using revised and extended dataset compiled in the Review.
- Tailored stress test enhancements:
  - Natural disasters:
    - country-specific triggers revised; calibration improved to make growth impacts country- and source-specific, capturing storms, floods, droughts, heatwaves, cold snaps, and earthquakes.
    - Country teams can adjust scenario parameters for country-specific circumstances.
  - Banking crisis stress test:
    - Replace uniform-exposure assumption with triggered test activated when macro-financial indicators signal heightened risks (similar to MAC-SRDSF).
    - Default calibration includes a direct, first-round effect on the primary deficit of 10 percent of GDP (benchmarked by median fiscal cost of systemic banking crises in LICs in Laeven and Valencia (2020)) and a second-round effect on real GDP growth, inflation, and interest rates; users can tailor calibration further.
  - Domestic financing stress test:
    - New tailored test for LICs with non-zero domestic debt to assess impact of deterioration in availability of external financing resulting in a shift towards shorter-term domestic debt; calibration tailored to country circumstances.
- Additional updates:
  - Commodity price shock and market financing shock updated and recalibrated.
  - Contingent liability stress test unchanged.
  - Customized alternative scenarios continue to allow assessment of other country-specific risks.

### Structured and transparent use of judgment; final ratings and granularity
- Role of judgment:
  - Structured and transparent use of judgment remains integral to final assessments of risk and sustainability.
  - Use of judgment expected to be highlighted and justified in the DSA write-up.
  - Under staff proposal, judgment better grounded on signals from new auxiliary modules; new Guidance Note will elaborate on structured use.
  - Assessments expected to account for main risks including data quality, macroeconomic uncertainty, policy implementation risks, and global conditions.
- Final risk ratings:
  - Continue to combine mechanical risk signals and structured, transparent judgment.
  - Mechanical signals provide a first pass; judgement informed by auxiliary modules used where appropriate.
  - Final rating of risk of public debt stress expected to be at least as conservative as rating of external public debt stress.
  - Ratings scale unchanged: Low Risk, Moderate Risk, High Risk, or In Debt Distress.
  - For High Risk or In Debt Distress, public debt sustainability assessed as either Sustainable or Unsustainable.
- Granularity improvements:
  - “Moderate Risk”:
    - Anchored on overall public DBIs.
    - Fiscal space derived from minimum distance (from below) between DBIs and their DCC-specific thresholds.
    - Classification:
      - “Limited fiscal space to absorb shocks” — if a median-size observed shock would lead to threshold breaches (at least one DBI lacks sufficient distance).
      - “Substantial fiscal space to absorb shocks” — if threshold breaches occur only under shocks in top quartile of observed distribution (all DBIs have sufficient distance to withstand all but top-quartile shock).
    - Analysis primarily focuses on overall public debt-to-GDP and GFN-to-GDP ratios.
    - Granularity helps inform decisions about scaling-up development spending.
  - “High Risk”:
    - Complemented with snapshot of intensity and persistence of underlying debt vulnerabilities by time horizon and type of risk.
    - Candidate methodology uses average size and frequency of threshold breaches to capture intensity and persistence.
    - For each DBI, threshold breaches averaged over three time horizons: near term (1-2 years ahead), medium term (3-to-5 years ahead), and longer term (beyond 5 years).
    - Values for solvency indicator and maximum of two liquidity indicators compared with interquartile range of breaches across LIC-DSAs with “High Risk, Sustainable” ratings over 2007-25; disk sizes reflect number of projected breaches as percentage of maximum possible per horizon (scale: 0, 33, 67, 100) and legend percentiles: Above 75th percentile; Below 75th percentile and above median; Below median and above 25th percentile; Below 25th percentile.
- Guidance:
  - Further guidance on use of different ratings to be provided in new Guidance Note.

### Next steps and transition
- Once approved, a transition period of approximately 9-12 months will be needed before full implementation.
  - Purpose: finalize new template, user manual, and Guidance Note, and begin rolling out training for staffs and country authorities.

*Review of Bank-Fund Debt Sustainability Framework for Low-Income Countries: Background Note (IMF and World Bank).*

### References _______________________________________________________________________________________ 34

### REVIEW OF BANK-FUND DEBT SUSTAINABILITY FRAMEWORK FOR LOW-INCOME COUNTRIES: BACKGROUND NOTE

### Background
- The joint IMF-World Bank Debt Sustainability Framework for Low-Income Countries (LIC-DSF) is described as the cornerstone of the assessment of the risk of public debt stress and sustainability in countries eligible for IMF and/or WB concessional financing.
- The LIC-DSF is used for macroeconomic analyses, policy advice, application of internal policies, and has been linked to lending policies of other multilateral development banks, including the African Development Bank.
- Debt sustainability analyses (DSAs) provide early warning signals on the build-up of risks of public debt stress and unsustainability and have been used to determine restructuring envelopes, notably under the G20 Common Framework.
- Since launch in 2005, the LIC-DSF has undergone periodic reviews to adapt to the evolving debt risk landscape.
- The 2017 review:
  - Simplified the framework by reducing the number of thresholds and standardized stress tests.
  - Expanded the measure of debt-carrying capacity to include macroeconomic fundamentals.
  - Strengthened the methodology for assessing overall public debt risks.
  - Introduced realism tools, tailored stress tests, and a market-financing module.
  - Enhanced guidance on the application of judgment to achieve a more evenhanded and structured application.
- The 2017 framework provides tools for assessing risks of external and overall public debt distress and for assessing debt sustainability, capturing country heterogeneity by classifying countries into three debt-carrying capacity groups: Weak, Medium, and Strong.
- The mechanical risk assessment compares solvency and liquidity debt-burden indicators to threshold values calibrated from past predictive ability; near-, medium-, and long-term mechanical signals derive from the same panel regression model and are conflated into single ratings for external and overall public debt distress; structured judgment is applied to reflect country specifics; sustainability assessment is judgment-based, informed by the size and persistence of mechanical signals.

### Context for the current review
- The review occurs amid heightened debt vulnerabilities and increased heterogeneity in LICs’ borrowing behavior:
  - Over the last 15 years, overall public debt has increased markedly as a share of GDP across all LICs, mirroring emerging markets.
  - Since the COVID-19 pandemic, LICs have faced relatively diminished access to external financing, adding impetus to a shift toward greater reliance on domestic financing (IMF and WB, 2025).
  - Frontier and relatively wealthier LICs have increasingly borrowed on non-concessional terms, resembling emerging markets; this has narrowed the gap between present and face value of overall public debt.
- Emerging challenges identified:
  - Absence of a fully calibrated model of overall debt stress and insufficient consideration of domestic debt vulnerabilities risk under-capturing public debt stress.
  - Lack of a clear mechanical signal for debt sustainability could impede consensus building during restructurings.
  - Increased heterogeneity across LICs strains the ability of current debt-carrying capacity-specific thresholds to capture repayment capacity differences.
  - A single time horizon for stress assessments reduces information content and complicates cross-country comparisons.
  - Effectiveness of realism tools may be strained as forecast uncertainty rises with more frequent and larger shocks.
  - Differences in debt coverage raise uniformity-of-treatment concerns.

### Identifying key reform areas (analysis and performance)
- The predictive performance of the LIC-DSF depends on the discriminatory power of the econometric model and robustness of framework design to the changing debt landscape.
- Performance observations:
  - The LIC-DSF has flagged debt distress episodes ahead of time; all cases of new “In Debt Distress” ratings since rollout in 2018 were preceded by a “High-Risk” rating in a recent DSA.
  - However, with over one-third of LIC-DSF users rated at high risk in a given year, the up-to-two-years-ahead probability of debt distress conditional on a “High-Risk” rating has been around 5 percent.
  - Of the latest LIC-DSAs with “High-Risk” external debt distress ratings, 15 out of 16 countries with near-to-medium-term (i.e., 1-to-5 years ahead) debt vulnerabilities are implementing reforms supported by an IMF program.
  - The large undifferentiated “High Risk” group, sometimes driven by long-term risks, complicates interpretation and may overestimate near-term risks; mechanical reliance on LIC-DSF risk ratings could influence lenders and access to financing.
- Specific pressure points for review:
  - Performance and robustness: Out-of-sample performance aligns with in-sample calibration, but re-estimation with updated data suggests resilience needs strengthening.
  - Reflecting heterogeneity: DCC grouping remains useful; “cliff effects” from DCC changes have not significantly impacted final ratings, but may cause discontinuities in fiscal advice if not internalized; Frontier LICs’ DCC generally aligns with market credit signals, but heterogeneity among Frontier and other LICs suggests further analysis of DCC variables/modalities.
  - Domestic debt vulnerabilities: Need to strengthen overall public debt distress assessment given rising domestic debt importance.
  - Judgment-based assessments: Use of judgment has improved outcomes for non-representative LICs but has contributed to higher false positives by reflecting long-term considerations without time differentiation of risk ratings; sustainability assessments would benefit from model-based signals.
  - Forecast accuracy: Short- and medium-term projections appear fairly accurate, though COVID-19 complicates conclusions; long-horizon forecasts show some optimism bias in exports and revenue forecasts.
  - Debt coverage: Comprehensive coverage is key as borrowing outside the budget and SOE-related risks have increased; cross-country differences in debt data completeness, including SOEs, remain a challenge.
  - Climate risks and policies: Coverage has evolved but remains uneven; additional tools are needed to support assessment.
  - Realism tools and stress tests: These can be further strengthened and retooled.
  - Continuity at graduation: Two separate frameworks (LIC-DSF and MAC-SRDSF) risk discontinuity for graduating countries; further alignment would mitigate this risk.

### Potential reforms and proposed design changes
- The note outlines a package of reforms aimed at strengthening and future-proofing the framework while keeping it tractable and easy to use:
  - Modeling architecture enhancements:
    - Re-estimation and recalibration of the external debt stress model.
    - Development of a new model of and mechanical risk signal for overall debt stress, with new solvency and liquidity debt burden indicators.
    - Updating the measurement of the Composite Indicator of Debt-Carrying Capacity (CI-DCC), calibrated to better reflect LICs’ heterogeneity.
    - Introducing a new module for analysis of domestic debt vulnerabilities.
  - A new model-based mechanical signal for overall public debt sustainability.
    - Judgment in sustainability assessment will be supported by auxiliary indicators of near-to-medium-term fiscal effort: distance between the primary balance and the debt-stabilizing primary balance, and magnitude of persistent breaches of debt-burden indicators in stress models.
  - Introduction of a new long-term module to model impacts of development challenges, including climate, and policies to address them.
  - Additional framework updates:
    - Introduce a confidence flag in debt data to incentivize improvements in transparency, reliability, and coverage.
    - Introduce granularity of the “High-Risk” ratings by time horizon and type of risk.
    - Update and enhance realism tools and stress tests.
  - Staffs also anticipate simplifying other aspects:
    - Propose a reduction in the number of debt-burden indicators used in the external public debt risk assessment.
    - Enhance stress tests.

### Conceptual framework refinements
- Terminology refinement:
  - The note proposes adopting the term “risk of public debt stress” to refine the previous shortcut term “risk of debt distress”; this clarifies distinctions among:
    - Risk of public debt stress: materialization of acute fiscal and/or market pressures related to public debt, with no presumption about modes of resolution (policy adjustments, external financing support, and/or debt restructuring).
    - Risk of unsustainable debt/public debt sustainability: in the context of PRGT and IDA-eligible countries, debt sustainability refers to the sovereign’s ability to meet current and/or future financial obligations while preserving growth at a satisfactory level and making adequate progress toward development goals (IMF and WB, 2017). Public debt is unsustainable when no politically and economically feasible policy set reasonably prospects to resolve public debt stress without debt restructuring (IMF, 2024).
  - The terms “risk of unsustainable debt” and “public debt sustainability” would remain unchanged in meaning relative to the current LIC-DSF.
- The conceptual refinement underscores:
  - Public debt stress can have solvency and liquidity dimensions.
  - Assessment is forward-looking and ultimately depends on judgment.
  - Clarifying terminology aims to reduce misinterpretation of signals as implying unsustainability when they reflect broader stress-risk continuums.

*Review of Bank-Fund Debt Sustainability Framework for Low-Income Countries: Background Note (IMF and World Bank).*

### 14.      A continued separate assessment of the risk of external public debt stress is warranted

### 14. A continued separate assessment of the risk of external public debt stress is warranted

### Importance of a separate external public debt stress assessment
- Most LICs exhibit twin current account and fiscal deficits, underlining the importance of external public borrowing as the most important source of financing across the majority of LICs.
- External public borrowing remains central to maintaining macroeconomic stability and is a distinct source of risk relative to emerging markets, where private sector flows and greater substitutability between domestic and external financing reduce the need for a separate assessment.
- In principle, external debt should continue to be defined based on the residency of the creditor.
- The shift toward greater reliance on domestic financing (IMF-WB 2025) underscores the importance of better incorporating domestic debt vulnerabilities in the overall debt risk assessment.

### Underlying regression models and the Composite DCC Indicator (CI-DCC)
- Framework architecture:
  - Three interconnected models form the core: (i) risk of external public debt stress (re-estimated), (ii) risk of overall public debt stress (new), and (iii) risk of public debt unsustainability (new).
  - All three models comprise a set of PROBITs that include one solvency and two liquidity debt-burden indicators (one per PROBIT), and a common composite indicator of debt-carrying capacity (CI-DCC).
- Dependent variables:
  - Criteria identifying onset of public debt stress for external and overall public debt stress have been refined and expanded.
  - Unsustainable debt events are identified as a subset of severe public debt stress events that cannot be resolved without debt restructuring.
  - Overall public debt stress and unsustainable debt criteria aligned more closely with MAC-SRDSF.
- Debt-burden indicators (DBIs):
  - Proposal to streamline external DBIs by dropping the PV of external debt-to-exports.
  - Overall DBIs to be better aligned with MAC-SRDSF: overall public debt taken at face rather than present value in the ratio to GDP; gross financing needs-to-GDP (GFN-to-GDP) replacing overall public debt service-to-revenues; interest payments as a ratio to revenues excluding grants to replace public debt service-to-revenues.
  - DBIs for public debt unsustainability model will be a subset of indicators used in external and overall stress models.
- CI-DCC refinements and composition:
  - CI-DCC continues as a weighted average of macroeconomic indicators capturing debt-carrying capacity.
  - World growth removed from CI-DCC construction due to loss of statistical significance.
  - CI-DCC to include:
    - Institutional quality proxy: WB Country Policy and Institutional Assessment (CPIA) index.
    - Domestic economic fundamentals: real GDP growth.
    - External buffers: reserve coverage and remittances as a ratio to GDP.
    - Measures of exposure to global macro-financial conditions and exogenous shocks:
      - Output volatility to differentiate diversified economies (e.g., many Frontier LICs) from Fragile and Conflict-Affected States (FCS).
      - Interaction of the interest rate on newly contracted external loans and the share of external in overall public debt to capture exposure to global macro-financial conditions.
  - Under the revised methodology, higher cost of financing and greater exposure to global macro-financial conditions require concurrent strengthening of institutional frameworks, macro-fundamentals, or external buffers to maintain the same CI-DCC score.
- DCC classification:
  - Countries assigned to one of three DCC groups: Weak, Medium, and Strong.
  - CI-DCC calculated using centered averages of DCC variables based on the latest five years of historical data and the first five years of projections.

### Mechanical risk signals for External and Overall Public Debt Stress
- Signal generation:
  - Mechanical signals generated by breaches of DCC-specific thresholds for debt-burden indicators in the baseline and stress scenarios over the first ten years of the forecast horizon.
- Risk categories defined:
  - "Low Risk" ⎯ when there are no sustained breaches of DCC-specific thresholds in either the baseline or stress scenarios.
  - "Moderate Risk" ⎯ when there are sustained breaches of DCC-specific thresholds occurring only in stress scenarios.
  - "High Risk" ⎯ when there are sustained breaches of DCC-specific thresholds in the baseline.
- Calibration and use:
  - External thresholds re-calibrated and new thresholds for overall DBIs selected to achieve predictive power broadly comparable to that at the time of the 2017 review.
  - Sustained breaches defined as occurring more than once over the forecast horizon.
  - Final risk rating generally informed by the more conservative of mechanical signals from external and overall public debt stress models:
    - External model captures continued importance of external debt for LICs.
    - Overall model captures domestic debt vulnerabilities where relevant.

### Public debt sustainability and the unsustainability model
- New unsustainability model:
  - Introduces a formal model for risk of unsustainable debt, improving on the current LIC-DSF reliance on judgement.
  - Model similar to stress models; probability of unsustainable debt determined by DBIs and CI-DCC.
  - Risk flags derived from comparing predicted probabilities to selected probability cut-offs.
  - Calibration targets a lower false positive rate (FPR) than in the stress model to avoid triggering self-fulfilling loss of financing.
  - Final sustainability assessment informed by mechanical signal and judgement, supported by indicators of fiscal effort required to stabilize overall public debt dynamics and severity of projected persistent threshold breaches in stress models in near- to medium-term.
- Mechanical signal for sustainability:
  - Only generated in cases of a final rating of public debt stress as "High Risk" or "In Distress".
  - Signal categories:
    - "Unsustainable" ⎯ when country-specific predicted probabilities from any of the three regression specifications equals or exceeds the applicable probability cutoff.
    - "Sustainable" ⎯ otherwise.

### Auxiliary modules
- Domestic Debt Risk Module (new):
  - Purpose: track domestic public debt risks to support differentiated analysis of domestic debt vulnerabilities in DSAs.
  - Components:
    - Core diagnostics: recent developments and projections of domestic public debt-to-GDP and domestic public debt service-to-revenue (excluding grants) ratios; comparisons to:
      - Medians of average projected values over the near-to-medium term for LIC-DSF users with non-zero domestic public debt and domestic public debt service.
      - Analysis of LICs that experienced domestic debt stress over the past decade.
    - Classification: countries categorized as low-, closer-, or heightened-scrutiny cases depending on whether forecasts of solvency and liquidity indicators are below or above their medians across all LICs and whether they exceed cross-country medians observed in past stress events.
    - Realism tools: benchmark net domestic public debt issuance and financial sector exposures to the sovereign against distributions across all LICs to flag upper-percentile cases; consider market pressures in domestic government bond markets (e.g., changes in demand across tenors).
  - Use in judgement:
    - Domestic debt vulnerabilities analysis serves as qualitative input into judgement for overall public debt risk and sustainability assessments.
    - Concerns about domestic public debt trajectories or borrowing plan consistency with macro-financial stability inform final ratings and sustainability assessments.
  - Notes on charts and data (as described):
    - Historical realizations and projections for 2018-32 presented for domestic public debt-to-GDP and domestic public debt service-to-revenue.
    - Median pre-stress historical pattern for 2015–24 and medians of average near- to medium-term projections across LIC-DSF users based on latest LIC-DSAs as of December 2025 referenced for benchmarking.
    - Borrowing assumptions and shares in new domestic debt issuance shown for medium- and long-term and short-term debt.
- Long-Term Module (new):
  - Purpose: complementary tool for assessing public debt stress and sustainability implications of policy and investment decisions associated with development and long-term challenges (including climate adaptation).
  - Approach:
    - Extend projections beyond the 10-year baseline (up to a maximum of 30 years) and overlay model-generated impulse responses capturing how macroeconomic factors would be impacted by specific long-term challenges.
    - Model planned policy changes over the medium-to-long term by allowing country teams to use the long-term module to model alternative policy scenarios in the first ten years of the forecast period.
    - Model long-term challenges and policies beyond the first ten forecast years to assess whether projected public debt-to-GDP remains consistent with maintaining public debt sustainability in the long run.
  - Outputs and limits:
    - The module produces insights that do not generate mechanical signals but inform policy choices and may be incorporated into final risk and sustainability assessments via judgement.
    - Exercise of judgement must account for large uncertainty and wide confidence bands in long-term projections.
    - Extended projections (up to 30 years) build on 10-year baseline forecasts by assuming the economy remains on the implied steady-state path.
  - Guidance:
    - Projections of GFNs-to-GDP ratio facilitate investigation of a sustainable financing mix.
    - Guidance to teams on evaluating dynamic stability of public debt-to-GDP and supporting GFNs-to-GDP ratios will be provided in the new Guidance Note.

### Incorporating climate risks and policies in the long-term analysis
- General approach:
  - Long-term climate module assesses debt-related impacts of:
    - (i) physical climate risks (gradual long-term shifts and extreme weather events increasing in frequency and intensity),
    - (ii) benefits and costs of policies and investments to increase resilience and support the energy transition.
- Scenario construction and outputs:
  - Assess potential impact on trajectories of overall public debt and GFN as ratios to GDP under different warming scenarios.
  - Extended projections of macro indicators overlaid with impact of physical climate risks under a moderate warming scenario that incorporates recurrent natural disasters, calibrated to the specific country.
  - Produce a "climate risk adjusted path" of projections augmented with a confidence band to reflect uncertainty around warming scenarios.
  - The climate risk-adjusted trajectory can be considered a “no-extra-adaptation” scenario (i.e., not adding additional adaptation investments beyond the baseline).
  - Policy scenarios model fiscal costs and payoffs of adaptation and domestic mitigation policies and investments; alternative scenarios allow benchmarking of different assumptions about the government's share in additional economy-wide climate-related spending consistent with maintaining an unchanged debt risk outlook.

*Background note for public consultation: Review of Bank-Fund Debt Sustainability Framework for Low Income Countries.*

### 32.      While the long-term module would allow the integration of climate impacts in DSAs

### 2026-lic-dsf-review-background-note-for-public-consultation - 32.      While the long-term module would allow the integration of climate impacts in DSAs

### Integration of climate impacts in DSAs (long-term climate module)
- Two complementary modelling approaches will be available as standard options at launch; country teams may select the approach most relevant to country specificities.
- Starting point for both approaches: the climate risk-adjusted projection of the public debt-to-GDP ratio.
- Confidence intervals capture:
  - parameter and warming sensitivity in different climate models, and
  - different adaptation speeds.
- Calibrated warming scenarios range from +1.8°C to +4.4°C by 2100.
- Policy scenarios illustrate trade-offs between higher spending on adaptation investment (increasing the debt-to-GDP ratio) and reduced GDP losses from natural disasters relative to a “no-extra-adaptation” scenario.
- Illustrative outputs included:
  - Public Debt-to-GDP Ratio (Percent) projections (Box Figures 1.1 and 1.3).
  - Avoided Income Losses as Result of Additional Adaptation Investment (Percent of climate risk-adjusted GDP) (Box Figures 1.2 and 1.4).
- Modelling inputs and methods referenced:
  - granular, spatially disaggregated geophysical and sectoral data into sectoral biophysical models, aggregated through a macro-structural model; and
  - pairing reduced-form estimates of slow-moving temperature increases with dynamic general equilibrium modelling of recurrent natural disasters and the climate investment-growth-debt nexus.
- The tailored natural disasters stress test will be updated to reflect the latest available, country-specific information on the effect on debt of extreme climate-related events.

### Development module: integrating development challenges and policies
- Purpose: provide an interface between the LIC-DSF and quantitative tools assessing the impact of policies addressing long-term development challenges (public investments in human and physical capital, infrastructure, structural reforms, sectoral interventions).
- Key mechanisms:
  - Development policies can enhance potential growth via productivity improvements, labor force development, and improved resource allocation.
  - Increased public investment financed solely by debt generally worsens debt burden indicators initially; growth dividends can strengthen fiscal revenues and partially improve debt dynamics in second-round effects.
- Two modelling approaches will be available; country teams choose based on country specificities.
- Illustrative scenarios (Box 2):
  - Increase in spending on human capital by 1 percent of GDP per year for ten years: initial larger deficits and borrowing needs; gradual improvements in potential output, revenues, and moderation of debt accumulation as human capital improves (Box Figures 2.1 and 2.2).
  - R&D scenario: permanent additional government spending of 0.3 percent of GDP that increases total factor productivity by 0.09 percentage points (Box Figures 2.3 and 2.4).
  - Female labor force participation scenario: assumed 2.5 percentage points increase through elimination of the gender gap in secondary and tertiary education via increased government spending by 0.17 percent of GDP per year, until it reaches 1.2 percentage points above the baseline where it is kept.
- Financing modalities crucial:
  - Simulations use a standard mix of concessional and non-concessional financing, and purely market-based non-concessional financing.
  - Non-concessional financing scenarios show a surge in public debt-to-GDP and GFN-to-GDP, underscoring the relevance of adequate funding schemes for development policies.
- Modelling tools referenced: Sustainable Development Goals Financing Tool; combination of MFMod-GJ and MFMod.

### Auxiliary sustainability indicators and realism tools
- Proposed auxiliary indicators to inform final debt sustainability assessments:
  - Distance between the primary deficit and the debt stabilizing primary deficit (DSPD) — reflects fiscal effort required to stabilize public debt-to-GDP at recent historical levels.
  - Magnitude of persistent breaches of DBIs used in stress assessments — forecasts of persistence of public debt stress beyond five years to judge feasibility of policies preventing unsustainability.
- Modalities and reference values for auxiliary indicators will be detailed in the new Guidance Note.
- Confidence flag on underlying debt data to inform judgment and improve cross-country comparability; it will summarize:
  - a summary measure of debt coverage used in the DSA (whether the full public sector perimeter is captured);
  - a composite signal based on two indicators of data reliability:
    - historical debt data revisions between the most recent DSA and earlier vintages (may signal reporting weaknesses or data improvements); and
    - discrepancies in public and publicly guaranteed external debt data reported in the DSA and in the World Bank International Debt Statistics (IDS) database;
  - a summary of the World Bank’s Debt Reporting Heatmap to reflect debt data transparency based on observed government reporting practices.
- Debt perimeter objective: near-complete coverage of external and domestic public and publicly guaranteed (PPG) debt, including central, state, and local governments, social security and extra-budgetary funds, the central bank, and state-owned non-financial public enterprises (SOEs), as well as private debt guaranteed by the public sector.
- Principle: include all SOE debt unless the SOE poses limited fiscal risks; criteria from the 2017 LIC-DSF Guidance Note to be clarified and streamlined.

### Enhanced realism tools (proposed)
- Retain and update three core realism tools:
  - decomposition of debt dynamics across DSA vintages to flag implausible shifts in debt-creating flows;
  - benchmarking of planned fiscal adjustment against cross-country and country-specific historical distributions;
  - consistency checks between fiscal adjustment and growth based on plausible fiscal multipliers.
- Empirical bases refreshed and benchmarking expanded beyond Fund-supported programs.
- New checks to cover projections for growth, exports, revenues (excluding grants), and financing assumptions:
  - Real GDP Growth — compare baseline projections to 10-year historical averages to flag unusually high or low projected growth.
  - Changes in exports-to-GDP and revenues (excl. grants)-to-GDP — flag large projected five-year changes relative to country history and LICs; if flagged, assess plausibility given country-specific factors.
  - More systematic scrutiny of financing assumptions, including diagnostics on maturity structure and cost of new domestic borrowing to identify understated gross financing needs.
- Empirical basis for realism tools: LIC DSAs database and staffs’ calculations; histograms based on all LIC-DSAs over 2015-2024 with bottom and top 2.5% trimmed to exclude outliers. Projected 5-year changes are relative to latest data outturns; latest 5-year changes capture difference between variable outturns in periods t and t - 5.

### Stress testing framework: standardized, tailored, and customized tests
- Framework retains enhanced versions of standardized, tailored, and customized stress tests; mechanical signals remain for “Low” and “Moderate” risk ratings.
- Standardized stress tests remain activated for all LIC-DSF users; tailored stress tests triggered only for relevant countries.
- Stress tests applied to both external and overall public DBIs; a new triggered banking crisis stress test applies only to overall DBIs (modelled assuming new debt created is financed domestically).
- Recalibration and updates using revised and extended dataset compiled in the Review.
- Proposed enhancements to tailored stress tests:
  - Natural disasters:
    - country-specific triggers revised to reflect enhanced data and analytical advances;
    - calibration improved to make growth impacts country- and source-specific, capturing effects of five types of climate-related natural disasters — storms, floods, droughts, heatwaves, and cold snaps — and non-climate-related disasters (earthquakes);
    - country teams retain option to adjust scenario parameters for country-specific circumstances.
  - Banking crisis stress test:
    - replace standardized uniform-exposure assumption with a triggered test activated when macro-financial indicators signal heightened risks (similar to MAC-SRDSF);
    - default calibration includes a direct, first-round effect on the primary deficit of 10 percent of GDP (benchmarked by median fiscal cost of systemic banking crises in LICs reported in Laeven and Valencia (2020)) and a second-round effect on real GDP growth, inflation, and interest rates; users can tailor calibration further.
  - Domestic financing stress test:
    - new tailored test for LICs with non-zero domestic debt to assess impact of a deterioration in availability of external financing that results in a shift towards shorter-term domestic debt; calibration tailored to country circumstances.
- Additional updates:
  - Commodity price shock and market financing shock will be updated and recalibrated.
  - Contingent liability stress test remains unchanged.
  - Customized alternative scenarios continue to allow assessment of other country-specific risks not covered by standardized and tailored tests.

*Source: Staff analysis.*

### 45.      The structured and transparent use of judgment to capture country-specific

### 45.      The structured and transparent use of judgment to capture country-specific considerations in the final assessments of risk and sustainability

### Use of judgment in assessments
- The structured and transparent use of judgment will remain an integral part of the LIC-DSF final assessments of risk and sustainability.
- As in the current LIC-DSF, the application of judgment would be expected to be highlighted in the DSA summary and justified in the DSA write-up.
- Under staff’s proposal:
  - The use of judgment would be better grounded on signals from the new auxiliary modules (Section D).
  - The new Guidance Note will provide further elaboration on its structured use to account for country-specific considerations.
- The assessment would be expected to take account of the main risks to the assessment, including those related to factors such as data quality, macroeconomic uncertainty, policy implementation risks, and global conditions, among others.

### Final risk ratings
- Final stress and sustainability ratings would continue to combine information from mechanical risk signals and the structured and transparent application of judgment.
  - Mechanical risk signals provide a first pass in determining the final rating.
  - Judgment— informed by the auxiliary modules and country-specific considerations—would be used, where appropriate, in the final decision.
- The final rating of the risk of public debt stress would be expected to continue to be at least as conservative as the rating of external public debt stress.
  - Rationale: the continued importance for LICs of external financing in contributing to macroeconomic stability while also representing a key source of risk.
- The ratings scale would remain unchanged: Low Risk, Moderate Risk, High Risk, or In Debt Distress.
- For cases of a high risk of public debt stress or in debt distress, the public debt sustainability would be assessed as either Sustainable or Unsustainable.
- Further guidance on the use of the different ratings will be provided in the new Guidance Note.

### Granularity of stress ratings

- “Moderate Risk” rating
  - Moderate Risk ratings will provide granularity on available fiscal space; this would now be anchored on the overall public DBIs.
  - The estimate of fiscal space continues to be derived from the minimum distance (from below) between debt-burden indicators and their DCC-specific thresholds.
  - This distance from below to thresholds—or fiscal space to absorb shocks—would continue to be assessed against the distribution of shocks that have historically been associated with downgrades from Moderate to High Risk of debt stress.
  - Classification for countries at Moderate Risk of public debt stress:
    - “Limited fiscal space to absorb shocks” ⎯ if a median-size observed shock would lead to threshold breaches, excluding cases of marginal and/or temporary breaches. This would be the case if at least one debt-burden indicator lacks sufficient distance to withstand a shock equivalent to the median observed shock.
    - “Substantial fiscal space to absorb shocks” ⎯ if threshold breaches would occur only under shocks in the top quartile of the observed distribution. This would be the case if all debt-burden indicators have sufficient distance to withstand all but the top-quartile observed shock.
  - The analysis would primarily focus on the overall public debt-to-GDP and GFN-to-GDP ratios, with other debt-burden indicators considered only when there are conflicting signals.
  - The new Guidance Note will provide guidance on how to incorporate information from these auxiliary signals and will also establish the implied distances to thresholds using the same empirical methodology applied in the last review.
  - This granularity would help inform discussions of the benefits and costs of scaling-up development spending to address human capital and development gaps.
    - Example: countries rated at Low or Moderate Risk of public debt stress with Substantial Fiscal Space could be good candidates for activation of the long-term module to explore the potential benefits and costs of stepped-up development spending.

- “High Risk” rating
  - Staffs propose to complement the “High Risk” stress ratings with a new snapshot of the intensity and persistence of underlying debt vulnerabilities by time horizon and type of risk.
  - Candidate proposal uses the average size and frequency of threshold breaches to capture the intensity and persistence of debt vulnerabilities (see Figure 9).
  - Expected benefits: enhance the clarity, consistency, and policy relevance of the LIC-DSF as a framework for flagging debt risks ahead of time and guiding the policy adjustments needed to mitigate them.

- Notes on Figure 9 methodology (as described)
  - For each debt burden indicator, threshold breaches (i.e., positive difference between debt burden indicator values and their DCC-specific thresholds expressed in percent) are averaged over three time horizons: near term (1-2 years ahead), medium term (3-to-5 years ahead), and longer term (beyond 5 years).
  - The values for the solvency indicator and the maximum of the values for the two liquidity indicators are compared with the interquartile range of such breaches across all LIC-DSAs with “High Risk, Sustainable” ratings over 2007-25, and shown as heatmap disks.
  - The size of the disks indicates the number of projected breaches, expressed as a percentage of the maximum possible for each time horizon.
  - Legend percentiles and share bands referenced: Above 75th percentile; Below 75th percentile and above median; Below median and above 25th percentile; Below 25th percentile; Below 25th percentile.
  - Share of maximum possible breaches (%) scale: 0, 33, 67, 100.

### Next steps
- Once the new framework is approved, a transition period of approximately 9-12 months will be needed before it can be fully implemented.
  - Purpose: time for staffs to finalize the new template, user manual, and Guidance Note, and to begin rolling out training for staffs and country authorities.

---


_Source: https://www.imf.org/-/media/files/about/partners/cso/2026-lic-dsf-review-background-note-for-public-consultation.pdf_