## Review of Bank-Fund Debt Sustainability Framework for Low Income Countries: Public Consultation on Staff Reform Proposals

## Source details

**Canonical URL:** [Review of Bank-Fund Debt Sustainability Framework for Low Income Countries: Public Consultation on Staff Reform Proposals](https://www.imf.org/-/media/files/about/partners/cso/2026-lic-dsf-review-presentation-for-public-consultation.pdf)

## Other formats

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

---

### LIC-DSF Primer — purpose, outputs, and core mechanics
- Purpose and use
  - Analytical tool for assessing debt vulnerabilities.
  - Jointly developed by IMF and WB in 2005 and periodically updated.
  - Used by countries eligible for IMF/WB concessional financing.
  - Objective: Guides borrowing and lending decisions.
- Outputs
  - Assessments of risk of public debt distress and sustainability.
  - Public debt distress: Materialization of acute fiscal or market pressures, without presupposing how it can be resolved.
  - Debt unsustainability: When public debt distress cannot be resolved without recourse to debt restructuring.
- Core structure and process
  - Mechanical tools generate risks of external & overall public debt distress.
  - Thresholds for three DCC categories: Weak, Medium, Strong.
  - Baseline macro and debt projections; debt carrying capacity based on CPIA and other fundamentals.
  - Final outputs: Risks of external & overall public debt distress; standardized and tailored stress tests; market financing module; judgment guided by Guidance Note.
  - Final classification: Sustainable or Unsustainable.
- Composite Indicator of Debt-Carrying Capacity (DCC)
  - Institutional quality: WB Country Policy and Institutional Assessment (CPIA) Index.
  - Economic fundamentals: Real GDP growth.
  - External buffers: Reserve Coverage (in level and squared), Remittances-to-GDP.
  - Exposure to global conditions: World growth.
  - Countries classified into Weak, Medium, Strong based on relative ranking of Composite Indicator; calculated twice-a-year using Spring and Fall WEO data.
- Debt burden indicators with DCC-specific thresholds (as presented)
  - Solvency
    - PV of Total Public Debt-to-GDP: Weak 355, Medium 570, Strong 355570
    - PV of External Debt-to-GDP: 30 40 55
    - PV of External Debt-to-Exports: 140 180 240
  - Liquidity
    - External Debt Service-to-Revenue: 14 18 23
    - External Debt Service-to-Exports: 10 15 21
  - Note: Ratios for PV of total public debt-to-GDP are referred to as “benchmarks” (noise-to-signal approach).
- Mechanical risk signals process
  - Compare five debt burden indicators to thresholds by DCC across baseline and stress scenarios over a 10-year horizon.
  - Threshold breaches over any horizon raise red flags.
- External vs Overall mechanical signals
  - External risk signal (primary output): breaches of four external debt burden indicators over 10-year timeframe.
  - Overall risk signal (supplementary): also considers PV of total public debt-to-GDP.
  - Risk flags matrix (Baseline Breaches / Shock Breaches):
    - Low: 0 / 0
    - Moderate: 0 / 1+
    - High: 1+ / 1+
- Role of judgment
  - Judgment brings country-specific considerations not fully captured by models (guided by Guidance Note).
  - Overall debt sustainability determined by judgment, informed by significant or sustained breaches.
- Strengths
  - Ease of interpretation and uniformity via discrete thresholds and consistent models.
  - Single-horizon risk ratings for overall and external debt.
  - Combination of model signals and judgment offers flexibility; since rollout all “In Distress” cases have been preceded by “High Risk” signal.
  - Separate external risk assessment; accounts for concessional borrowing and external debt service importance.
  - Near-complete targeted coverage of public debt.
- Trade-offs / limitations
  - DCC-specific thresholds do not fully reflect LIC heterogeneity.
  - Risks of optimism bias in projections; realism tools and review process expected to mitigate.
  - 10-year horizon conflates near-, medium-, and long-term risks into a single rating.
  - Lack of mechanical signal for debt sustainability reduces transparency.
  - Focus on external risk assessment may limit insight into domestic debt vulnerabilities.
  - Differences in data availability limit debt coverage in practice.

### Impact of changing debt landscape on LIC-DSF implementation
- Observed changes (2000–24)
  - Debt burdens have increased across-the-board over the past 15 years.
  - Importance of domestic debt has increased as share of FX-currency in total general government debt has fallen across low-income countries.
  - Since 2010, frontier and other wealthier LICs decoupled from other LICs in reliance on external borrowing from private creditors and resemble lower income EMs more than other LICs.
- Emerging implementation challenges
  - Potential inability to fully capture increased country heterogeneity could unduly limit access to financing or risk over-borrowing.
  - Effectiveness of realism tools strained as forecast uncertainty increases with higher frequency & magnitude of shocks.
  - Lack of fully developed framework inhibits quality of incorporation of domestic debt vulnerabilities in overall risk assessment.
  - Lack of time-horizon differentiation complicates interpretation of risk ratings.
  - High risk countries have different risk profiles (by risk type and time horizon).
  - Absence of clear framework for sustainability assessment could impede consensus among creditors in debt restructurings.
  - Cross-country differences in debt coverage raise uniformity of treatment concerns.

### Backward-looking analysis of current LIC-DSF
- Overall assessment
  - LIC-DSF remains “Fit-for-Purpose” but can be improved to “Future-Proof” it.
- Main findings (1)
  - LIC-DSF accurately flags debt distress with somewhat high rate of false positives (in line with 2017 calibration).
  - Increased LIC heterogeneity challenging model performance on re-estimation.
  - Grouping countries by DCC useful to keep framework easy to use.
  - “Cliff Effects” can cause policy discontinuity if unanticipated; size generally not big enough to affect final risk rating.
  - Frontier LICs DCC generally aligns well with market signals; more analysis of country characteristics warranted.
  - Use of two frameworks (LIC-DSF and MAC-SRDSF), while justified, could lead to discontinuity in risk assessments.
  - Need to strengthen assessment of overall public debt distress; sustainability assessments would benefit from model-based signals.
- Main findings (2)
  - Comprehensive debt coverage remains key; differences in coverage can affect quality of risk signals.
  - Judgment improves outcomes but adds to false positives when used to capture long-term risks.
  - Coverage of climate risks and policies is evolving but uneven in scope and depth.
  - Short- and medium-term forecasts appear overall fairly accurate, but analysis complicated by COVID-19.
  - Some optimism bias in revenue and exports forecasts over longer horizons.

### Reform objectives and headline proposals
- Reform objectives
  - More resilient predictive accuracy.
  - Better reflect country heterogeneity.
  - Elevate role of overall debt risk model and better inform sustainability assessment.
  - More structured coverage of long-term issues (including climate and development policies).
- Proposed specific reforms (overview)
  - Re-affirm setting of discount rate used in PV calculation.
  - Re-estimation of External Debt Stress model with revised and updated data.
  - Recalibration of predictive power of model.
  - Refined Composite DCC Indicator including new variables capturing LIC heterogeneity and exposure to global macro-financial conditions.
  - Recalibration of rules for grouping countries by DCC.
  - New Overall Debt Stress model with thresholds.
  - New toolkit for analysis of domestic debt vulnerabilities.
  - New mechanical risk signal for debt unsustainability.
  - New long-term module for modelling impact of development challenges and policies.
- Further enhancements
  - Encourage debt data comprehensiveness, quality, and transparency.
  - New debt data confidence flag encouraging broader debt coverage and transparency.
  - Adjustments to baseline debt-burden indicators to address data gaps.
  - Planned enhanced guidance on debt perimeter in new Guidance Note.
  - Add granularity of High-risk rating by time and type of risk.
  - Strengthened realism tools and recalibrated stress tests.
  - Better alignment between LIC-DSF and MAC-SRDSF.
- Structure of retained elements and enhancements
  - External & Overall Stress
    - Retained: External and Overall Debt Burden Indicators (DBIs) in baseline and stress; DCC determined by baseline macro assumptions.
    - Enhancements: Strengthened realism tools and stress tests; adjustment to baseline DBIs to close gaps in SOE data.
  - Mechanical Signal
    - Retained: Thresholds for DBIs by DCC; breaches trigger risk flags.
    - Enhancements: Re-estimation with revised/updated data; updated DCC variable set; new thresholds for Overall DBIs; new domestic debt risk module.
  - Final Assessment of Risk of Debt Stress
    - Enhancements: Long-term risk module to assess LT challenges/policies; modelling of climate risks and development policies; refined country-specific approach to derive stress signal; confidence flag on debt data coverage and transparency; enhanced guidance for application of judgment in Guidance Note; granularity of High-Risk rating by time and type of risk.
    - Retained: Mechanical risk signals combined with structured judgment; separate risk ratings for Public and External public debt stress; granularity of Moderate risk ratings.

### Composite DCC Indicator and DBI re-estimation
- Composite DCC Indicator refinements
  - Institutional quality: WB CPIA Index (retained).
  - Economic fundamentals: Real GDP growth and Output volatility (new).
  - External buffers: Reserve Coverage (log) and Remittances-to-GDP.
  - Exposure to global macro-financial conditions: Interest Rate on New External Loans interacted with Share of External in Overall Public Debt (new).
  - Note: World growth removed due to loss of statistical significance; squared term of reserve coverage replaced by log.
- Re-estimated thresholds and simplified indicator set
  - External DBIs simplified to 3 indicators:
    - Solvency: PV of External Public Debt-to-GDP.
    - Liquidity: External Public Debt Service as Ratio to Revenue (excl. Grants) and Exports.
    - Predictive power comparable to 2017 Review.
  - Overall DBIs: 3 indicators:
    - Solvency: FV of Overall Public Debt-to-GDP.
    - Liquidity: Gross Financing Needs-to-GDP.
    - Interest Payments on Overall Public Debt-to-Revenues (excl. Grants).
    - Similar predictive power to external model.

### New sustainability-focused model and domestic debt module
- New sustainability-focused model
  - Mechanical signal: Country-specific predicted probability (at end of medium term).
  - Probability cut-offs target lower FPR than in stress models.
  - Breaches of probability cut-offs in baseline trigger risk flags.
  - Final sustainability assessment combines mechanical signal with structured judgment informed by:
    - Measure of fiscal effort required to stabilize overall public debt dynamics.
    - Severity and timing of projected threshold breaches in debt stress models.
- New domestic debt risk module (intro)
  - Will use indicators such as Domestic Public Debt-to-GDP ratio and Domestic Public Debt Service-to-Revenue ratio (excl. Grants) to strengthen analysis of domestic vulnerabilities.
  - Module includes reference values for solvency and liquidity debt-burden indicators calibrated on past episodes of domestic debt stress.
  - Depth of analysis reflects severity of risks and is informed by benchmarking of forecasts to:
    - Medians across all LICs
    - Medians in run-up to past stress events
  - Realism tools for domestic financing assumptions:
    - Identify pressures stemming from domestic borrowing plan
    - Assess consistency of projected take-up of new domestic government debt with macroeconomic stability and absorptive capacity of financial system
    - Evaluate build-up of risks from Sovereign-Bank nexus and potential for crowding out of private sector
    - Analysis would also consider risk signals from primary/secondary markets

### Long-term modules: climate-related risks and development policies
- Long-Term Module — Climate-Related Risks and Adaptation Policies
  - Complementary tool, applied where relevant, for assessing public debt stress and sustainability implications of policies and investments to address development and other long-term challenges, including those stemming from climate adaptation.
  - Resulting insights would not generate any mechanical signals and would serve to inform policy choices; they could also be incorporated in final risk and sustainability assessments, where relevant, through application of judgment.
  - Assessment features:
    - Impact on Overall Public Debt and GFN to GDP ratios
    - Climate risk adjusted projections
    - Confidence band
    - Policy scenarios
    - Financing mix and government share in spending consistent with unchanged risk outlook
  - Plug & Play interface:
    - Integration of climate modelling, including from WB (CCDR, MFMoD CC) and IMF (Q-CRAFT and DIGNAD)
  - Tailored natural disasters stress test with updated calibration
  - Note: The WB “Country Climate and Development Report (CCDR) adaptation” scenario assumes an average adaptation investment of 1% of GDP per year fully borne by the government.
- Long-Term Module — Development Policies
  - Tool assesses benefit-cost of development policies:
    - Public investments in human and physical capital, structural reforms, and sectoral interventions can boost growth but come at a cost that adds to debt
  - Assessment of impact on Overall Public Debt and GFN to GDP ratios using mechanics similar to the climate module
  - Notes: The simulations are based on the IMF Sustainable Development Goals Financing Tool. The scenario assumes spending on human capital by 1% of GDP per year for ten years.

### Granularity of “High-Risk” ratings and visualization features
- Granularity approach
  - Heatmap: Size of breaches (in percent of threshold) averaged by time horizon for each DBI; Solvency and maximum across liquidity ones compared to interquartile range of breaches over 2007-25.
  - Disk size: Share of years with projected breaches by time horizon.
  - Proposed granularity of “High Risk” ratings (by percentile bands):
    - Above 75th percentile
    - Below 75th percentile and above median
    - Below median and above 25th percentile
    - Below 25th percentile
  - Share of maximum possible breaches (%) scale shown as: 0, 33, 67, 100
  - Time horizons defined as:
    - Near term (1-2 years ahead)
    - Medium term (3-to-5 years ahead)
    - Longer term (6-to-10 years)
  - Risk dimensions called out: Solvency and Liquidity

### Confidence Flag on Debt Data to incentivize better coverage and reliability
- Components of the confidence flag (traffic light chart)
  - Comprehensiveness of DSA debt perimeter
  - Composite signal of reliability
  - Revisions in historical data in DSAs
  - Discrepancies between debt data in DSAs and published in WB IDS database
  - Relative standing in WB Debt Reporting Heatmap
  - Across nine indicators, incl. coverage, frequency, timeliness, and public availability of debt statistics and debt management documents
- Proposed mechanics and notes
  - Debt data coverage classification reflects the public sector perimeter used in the DSA (proposed to be based on the most recent DSA)
  - The composite measure of debt data reliability combines information about:
    - (i) Revisions of historical external and domestic debt data between the most recent DSA and the DSA vintage from two-years ago.
    - (ii) Comparison of USD discrepancies between WB IDS (DRS) and DSA data.
  - The indicator on debt transparency is based on World Bank Debt Reporting Heatmap (updated annually)

### Core architecture and next steps
- Core framework components to be retained and augmented
  - Staff judgment; Inputs/Assumptions; Debt data & projections; Thresholds and stress tests.
  - Auxiliary modules and signals: Debt burden indicators (DBIs); Debt Carrying capacity (DCC) variables; Realism tools; Mechanical signals (1. Public debt stress, 2. External public debt stress, 3. Public debt sustainability); Domestic debt, Climate, Development modules; Data confidence flag; Granularity features (Debt stabilizing primary balance; Severity and timing of breaches; Cut-offs probabilities).
- Implementation timeline
  - Formal approval of new LIC-DSF expected in Autumn
  - Rollout in 2027, following development of new Template, Guidance Note, and further outreach and training

*Source: Bank-Fund Debt Sustainability Framework for Low Income Countries (LIC-DSF) — staff presentation.*

### Section 1

### Review of Bank-Fund Debt Sustainability Framework for Low Income Countries: Public Consultation on Staff Reform Proposals

### LIC-DSF Primer
- What is the LIC-DSF?
  - Analytical tool for assessing debt vulnerabilities.
  - Jointly developed by IMF and WB in 2005 and periodically updated.
  - Used by countries eligible for IMF/WB concessional financing.
  - Objective: Guides borrowing and lending decisions.
- Outputs
  - Assessments of risk of public debt distress and sustainability.
  - Public debt distress: Materialization of acute fiscal or market pressures, without presupposing how it can be resolved.
  - Debt unsustainability: When public debt distress cannot be resolved without recourse to debt restructuring.
- Bird’s eye view / current structure
  - Mechanical tools: Risks of external & overall public debt distress.
  - Thresholds for three DCC categories: Weak, Medium, Strong.
  - Baseline macro and debt projections; debt carrying capacity based on CPIA and other fundamentals.
  - Final outputs: Risks of external & overall public debt distress; stress tests (standardized and tailored); market financing module; judgment guided by Guidance Note.
  - Final classification: Sustainable or Unsustainable.
- Composite Indicator of Debt-Carrying Capacity (DCC)
  - Institutional quality: WB Country Policy and Institutional Assessment (CPIA) Index.
  - Economic fundamentals: Real GDP growth.
  - External buffers: Reserve Coverage (in level and squared), Remittances-to-GDP.
  - Exposure to global conditions: World growth.
  - Countries classified into Weak, Medium, Strong based on relative ranking of Composite Indicator; calculated twice-a-year using Spring and Fall WEO data.
- Debt burden indicators with DCC-specific thresholds (as presented)
  - Solvency
    - PV of Total Public Debt-to-GDP: Weak 355, Medium 570, Strong (blank in source layout but presented as 0? — keep as listed: 355570)
    - PV of External Debt-to-GDP: 30 40 55
    - PV of External Debt-to-Exports: 140 180 240
  - Liquidity
    - External Debt Service-to-Revenue: 14 18 23
    - External Debt Service-to-Exports: 10 15 21
  - Note: Ratios for PV of total public debt-to-GDP are referred to as “benchmarks” (noise-to-signal approach).
- Mechanical risk signals process
  - Compare five debt burden indicators to thresholds by DCC across baseline and stress scenarios over a 10-year horizon.
  - Threshold breaches over any horizon raise red flags.
- External vs Overall mechanical signals
  - External risk signal (primary output): breaches of four external debt burden indicators over 10-year timeframe.
  - Overall risk signal (supplementary): also considers PV of total public debt-to-GDP.
  - Risk flags matrix (Baseline Breaches / Shock Breaches):
    - Low: 0 / 0
    - Moderate: 0 / 1+
    - High: 1+ / 1+
- Role of judgment
  - Judgment brings country-specific considerations not fully captured by models (guided by Guidance Note).
  - Overall debt sustainability determined by judgment, informed by significant or sustained breaches.
- Strengths
  - Ease of interpretation and uniformity via discrete thresholds and consistent models.
  - Single-horizon risk ratings for overall and external debt.
  - Combination of model signals and judgment offers flexibility; since rollout all “In Distress” cases have been preceded by “High Risk” signal.
  - Separate external risk assessment; accounts for concessional borrowing and external debt service importance.
  - Near-complete targeted coverage of public debt.
- Trade-offs / limitations
  - DCC-specific thresholds do not fully reflect LIC heterogeneity.
  - Risks of optimism bias in projections; realism tools and review process expected to mitigate.
  - 10-year horizon conflates near-, medium-, and long-term risks into a single rating.
  - Lack of mechanical signal for debt sustainability reduces transparency.
  - Focus on external risk assessment may limit insight into domestic debt vulnerabilities.
  - Differences in data availability limit debt coverage in practice.

### Impact of Changing Debt Landscape on LIC-DSF Implementation
- Observed changes (2000–24; indicators shown in presentation)
  - General Government Debt-to-GDP Ratio (percent) — distributions depicted by boxplots (no exact numeric medians provided in text).
  - Share of FX-Currency Debt in General Government Debt (percent) — distributions depicted.
  - Share of Debt to Private Creditors in External Public and Publicly Guaranteed Debt (percent) — distributions depicted.
- Key contextual findings
  - Debt burdens have increased across-the-board over the past 15 years.
  - Importance of domestic debt has increased as share of FX-currency in total general government debt has fallen across low-income countries.
  - Since 2010, frontier and other wealthier LICs decoupled from other LICs in reliance on external borrowing from private creditors and resemble lower income EMs more than other LICs.
- Emerging implementation challenges
  - Potential inability to fully capture increased country heterogeneity could unduly limit access to financing or risk over-borrowing.
  - Effectiveness of realism tools strained as forecast uncertainty increases with higher frequency & magnitude of shocks.
  - Lack of fully developed framework inhibits quality of incorporation of domestic debt vulnerabilities in overall risk assessment.
  - Lack of time-horizon differentiation complicates interpretation of risk ratings.
  - High risk countries have different risk profiles (by risk type and time horizon).
  - Absence of clear framework for sustainability assessment could impede consensus among creditors in debt restructurings.
  - Cross-country differences in debt coverage raise uniformity of treatment concerns.

### Backward-Looking Analysis of Current LIC-DSF
- Overall assessment
  - LIC-DSF remains “Fit-for-Purpose” but can be improved to “Future-Proof” it.
- Main findings (1)
  - LIC-DSF accurately flags debt distress with somewhat high rate of false positives (in line with 2017 calibration).
  - Increased LIC heterogeneity challenging model performance on re-estimation.
  - Grouping countries by DCC useful to keep framework easy to use.
  - “Cliff Effects” can cause policy discontinuity if unanticipated; size generally not big enough to affect final risk rating.
  - Frontier LICs DCC generally aligns well with market signals; more analysis of country characteristics warranted.
  - Use of two frameworks (LIC-DSF and MAC-SRDSF), while justified, could lead to discontinuity in risk assessments.
  - Need to strengthen assessment of overall public debt distress; sustainability assessments would benefit from model-based signals.
- Main findings (2)
  - Comprehensive debt coverage remains key; differences in coverage can affect quality of risk signals.
  - Judgment improves outcomes but adds to false positives when used to capture long-term risks.
  - Coverage of climate risks and policies is evolving but uneven in scope and depth.
  - Short- and medium-term forecasts appear overall fairly accurate, but analysis complicated by COVID-19.
  - Some optimism bias in revenue and exports forecasts over longer horizons.

### Reform Proposals and Key Features
- Reform objectives
  - More resilient predictive accuracy.
  - Better reflect country heterogeneity.
  - Elevate role of overall debt risk model and better inform sustainability assessment.
  - More structured coverage of long-term issues (including climate and development policies).
- Proposed specific reforms (overview)
  - Re-affirm setting of discount rate used in PV calculation.
  - Re-estimation of External Debt Stress model with revised and updated data.
  - Recalibration of predictive power of model.
  - Refined Composite DCC Indicator including new variables capturing LIC heterogeneity and exposure to global macro-financial conditions.
  - Recalibration of rules for grouping countries by DCC.
  - New Overall Debt Stress model with thresholds.
  - New toolkit for analysis of domestic debt vulnerabilities.
  - New mechanical risk signal for debt unsustainability.
  - New long-term module for modelling impact of development challenges and policies.
- Further enhancements
  - Encourage debt data comprehensiveness, quality, and transparency.
  - New debt data confidence flag encouraging broader debt coverage and transparency.
  - Adjustments to baseline debt-burden indicators to address data gaps.
  - Planned enhanced guidance on debt perimeter in new Guidance Note.
  - Add granularity of High-risk rating by time and type of risk.
  - Strengthened realism tools and recalibrated stress tests.
  - Better alignment between LIC-DSF and MAC-SRDSF.
- Structure of retained elements and enhancements
  - External & Overall Stress
    - Key elements retained: External and Overall Debt Burden Indicators (DBIs) in baseline and stress; DCC determined by baseline macro assumptions.
    - Enhancements: Strengthened realism tools and stress tests; adjustment to baseline DBIs to close gaps in SOE data.
  - Mechanical Signal
    - Key elements retained: Thresholds for DBIs by DCC; breaches trigger risk flags.
    - Enhancements: Re-estimation with revised/updated data; updated DCC variable set; new thresholds for Overall DBIs; new domestic debt risk module.
  - Final Assessment of Risk of Debt Stress
    - Enhancements: Long-term risk module to assess LT challenges/policies; modelling of climate risks and development policies; refined country-specific approach to derive stress signal; confidence flag on debt data coverage and transparency; enhanced guidance for application of judgment in Guidance Note; granularity of High-Risk rating by time and type of risk.
    - Key elements retained: Mechanical risk signals combined with structured judgment; separate risk ratings for Public and External public debt stress; granularity of Moderate risk ratings.
- Composite DCC Indicator refinements
  - Institutional quality: WB CPIA Index (retained).
  - Economic fundamentals: Real GDP growth and Output volatility (new).
  - External buffers: Reserve Coverage (log) and Remittances-to-GDP.
  - Exposure to global macro-financial conditions: Interest Rate on New External Loans interacted with Share of External in Overall Public Debt (new).
  - Note: World growth removed due to loss of statistical significance; squared term of reserve coverage replaced by log.
- Re-estimated thresholds and simplified indicator set
  - External DBIs simplified to 3 indicators:
    - Solvency: PV of External Public Debt-to-GDP.
    - Liquidity: External Public Debt Service as Ratio to Revenue (excl. Grants) and Exports.
    - Predictive power comparable to 2017 Review.
  - Overall DBIs: 3 indicators:
    - Solvency: FV of Overall Public Debt-to-GDP.
    - Liquidity: Gross Financing Needs-to-GDP.
    - Interest Payments on Overall Public Debt-to-Revenues (excl. Grants).
    - Similar predictive power to external model.
- New sustainability-focused model
  - Mechanical signal: Country-specific predicted probability (at end of medium term).
  - Probability cut-offs target lower FPR than in stress models.
  - Breaches of probability cut-offs in baseline trigger risk flags.
  - Final sustainability assessment combines mechanical signal with structured judgment informed by:
    - Measure of fiscal effort required to stabilize overall public debt dynamics.
    - Severity and timing of projected threshold breaches in debt stress models.
- New domestic debt risk module (intro)
  - Will use indicators such as Domestic Public Debt-to-GDP ratio and Domestic Public Debt Service-to-Revenue ratio (excl. Grants) to strengthen analysis of domestic vulnerabilities.

*Source: Bank-Fund Debt Sustainability Framework for Low Income Countries (LIC-DSF) — staff presentation.*

### Section 2

### New LIC-DSF: Module Enhancements and Long-Term Tools

### Domestic Debt Risk Module
- New module includes reference values for solvency and liquidity debt-burden indicators calibrated on past episodes of domestic debt stress.
- Depth of analysis reflects severity of risks and is informed by benchmarking of forecasts of solvency and liquidity indicators to:
  - Medians across all LICs
  - Medians in run-up to past stress events
- Realism tools for domestic financing assumptions:
  - Identify pressures stemming from domestic borrowing plan
  - Assess consistency of projected take-up of new domestic government debt with macroeconomic stability and absorptive capacity of financial system
  - Evaluate build-up of risks from Sovereign-Bank nexus and potential for crowding out of private sector
  - Analysis would also consider risk signals from primary/secondary markets

### Long-Term Module — Climate-Related Risks and Adaptation Policies
- Complementary tool, applied where relevant, for assessing public debt stress and sustainability implications of policies and investments to address development and other long-term challenges, including those stemming from climate adaptation.
- Resulting insights would not generate any mechanical signals and would serve to inform policy choices; they could also be incorporated in final risk and sustainability assessments, where relevant, through application of judgment.
- Assessment features:
  - Impact on Overall Public Debt and GFN to GDP ratios
  - Climate risk adjusted projections
  - Confidence band
  - Policy scenarios
  - Financing mix and government share in spending consistent with unchanged risk outlook
- Plug & Play interface:
  - Integration of climate modelling, including from WB (CCDR, MFMoD CC) and IMF (Q-CRAFT and DIGNAD)
- Tailored natural disasters stress test with updated calibration
- Note: The WB “Country Climate and Development Report (CCDR) adaptation” scenario assumes an average adaptation investment of 1% of GDP per year fully borne by the government.

### Long-Term Module — Development Policies
- Tool assesses benefit-cost of development policies:
  - Public investments in human and physical capital, structural reforms, and sectoral interventions can boost growth but come at a cost that adds to debt
- Assessment of impact on Overall Public Debt and GFN to GDP ratios using mechanics similar to the climate module
- Notes: The simulations are based on the IMF Sustainable Development Goals Financing Tool. The scenario assumes spending on human capital by 1% of GDP per year for ten years.

### Granularity of “High-Risk” Ratings: Intensity, Persistence, and Types of Vulnerabilities
- Heatmap:
  - Size of breaches (in percent of threshold) averaged by time horizon for each DBI
  - Solvency and maximum across liquidity ones compared to interquartile range of breaches over 2007-25
- Disk size:
  - Share of years with projected breaches by time horizon
- Proposed granularity of “High Risk” ratings (by percentile bands):
  - Above 75th percentile
  - Below 75th percentile and above median
  - Below median and above 25th percentile
  - Below 25th percentile
- Share of maximum possible breaches (%) scale shown as: 0, 33, 67, 100
- Time horizons defined as:
  - Near term (1-2 years ahead)
  - Medium term (3-to-5 years ahead)
  - Longer term (6-to-10 years)
- Risk dimensions called out: Solvency and Liquidity

### Confidence Flag on Debt Data to Incentivize Better Data Coverage, Transparency, and Reliability
- Traffic light chart components:
  - Comprehensiveness of DSA debt perimeter
  - Composite signal of reliability
  - Revisions in historical data in DSAs
  - Discrepancies between debt data in DSAs and published in WB IDS database
  - Relative standing in WB Debt Reporting Heatmap
  - Across nine indicators, incl. coverage, frequency, timeliness, and public availability of debt statistics and debt management documents
- Proposed Confidence Flag on Debt Data notes:
  - Debt data coverage classification reflects the public sector perimeter used in the DSA (proposed to be based on the most recent DSA)
  - The composite measure of debt data reliability combines information about:
    - (i) Revisions of historical external and domestic debt data between the most recent DSA and the DSA vintage from two-years ago.
    - (ii) Comparison of USD discrepancies between WB IDS (DRS) and DSA data.
  - The indicator on debt transparency is based on World Bank Debt Reporting Heatmap (updated annually)

### New Features to Enhance Core Structure and Future-Proof the LIC-DSF
- Core framework components:
  - Staff judgment
  - Inputs/Assumptions
  - Debt data & projections
  - Thresholds and stress tests
- Auxiliary modules and signals:
  - Debt burden indicators (DBIs)
  - Debt Carrying capacity (DCC) variables
  - Realism tools
  - Mechanical signals:
    1. Public debt stress
    2. External public debt stress
    3. Public debt sustainability
  - Domestic debt, Climate, Development modules
  - Data confidence flag
  - Granularity features:
    - Debt stabilizing primary balance
    - Severity and timing of breaches
    - Cut-offs probabilities

### Next Steps
- Formal approval of new LIC-DSF expected in Autumn
- Rollout in 2027, following development of new Template, Guidance Note, and further outreach and training

---


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