## IMF WORKING PAPERS Peer Pressure: How Relative Debt Drives Emerging Market Sovereign Spreads

## Source details

**Canonical URL:** [IMF WORKING PAPERS Peer Pressure: How Relative Debt Drives Emerging Market Sovereign Spreads](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026110-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026110-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026110-source-pdf.pdf.json)

---

### Introduction and framing
- Sovereign bond spreads are defined as the yield differential between sovereign bonds and benchmark securities, typically U.S. Treasuries of comparable maturity.
- Public debt is a prominent determinant of sovereign spreads; investor assessments are comparative and benchmark countries against peers with similar income levels, regional location, or economic structures.
- The paper introduces "relative public debt" as a complementary/explanatory variable to traditional gross debt measures for emerging market and developing economies (EMDEs).
- Relative public debt is intended to capture market benchmarking behavior rather than structural sustainability thresholds.

### Empirical approach, data, and definitions
- Relative public debt definition:
  - Relative debt_{i,t} = ln(Total Public Debt-to-GDP_{i,t-1}) − ln(Median Total Public Debt-to-GDP_{j,t-1}), where j is the peer group (income, region, commodity-exporter status).
- Peer groups:
  - Income classification: World Bank (low, medium, high income).
  - Region: Moody’s regional classification.
  - Commodity exporters: World Bank 2024 Global Economic Prospects list (Venezuela added).
- Sample and dependent variable:
  - J.P. Morgan EMBIG sovereign spreads for 83 emerging market countries from 1993 to 2024; annual average spreads; sample restricted to observations ≤ 1,500 basis points.
- Controls:
  - Domestic: Real GDP growth, inflation.
  - Global: VIX, U.S. 10-year Treasury yield.
- Baseline estimation:
  - Fixed effects panel regressions with clustered standard errors; main specification uses lagged debt measures and lagged domestic factors; VIX and US 10-year included contemporaneously.

### Core empirical findings
- Relative debt adds significant explanatory power to sovereign spread models, especially during tranquil global financial conditions.
- When both absolute (gross) debt and relative debt are included, relative debt often dominates and gross debt can lose statistical significance or become unstable.
- Key coefficients and model fit (selected):
  - Univariate comparisons (Text Table 1):
    - Debt coefficient: 0.232*** (SE 0.087) in column (1); Observations: 1,197; Number of countries: 83; R-squared: 0.020.
    - Relative debt (Income classification): 0.468*** (SE 0.101) in column (2); R-squared: 0.065.
  - Baseline multivariate (Text Table 4, core specification Model 6 by income classification):
    - Relative debt: 0.379*** (SE 0.098); Observations 1,189; R-squared 0.250; Number of countries 83.
    - VIX example: 0.617*** (SE 0.045) in baseline.
  - Predicted magnitude:
    - "A ten percent increase in relative debt is predicted to increase spreads by 3.8 percent" in core specification (Model 6).
  - Standardized impact:
    - A one-standard-deviation increase in relative debt raises sovereign spreads by roughly 0.2–0.3 standard deviations.
- Control variable behavior examples:
  - Real GDP growth coefficients ~ -0.020 to -0.024 (significant).
  - Inflation coefficients ~ 0.014 to 0.016 (significant).
  - US 10-year Treasury yield: insignificant and negative in some specifications (example -0.031, SE 0.064).

### Heterogeneity: when relative debt matters more (split-sample evidence)
- Six split dimensions: Institutional strength (WGI), Reserve adequacy (months of imports), Access to concessional finance (World Bank Blend/IDA status), History of arrears (PPG arrears in past five years), Global risk aversion (VIX threshold 20), Global liquidity (global M2-to-GDP gap ≥ 0 vs < 0).
- Selected relative-debt coefficient estimates (from Text Table 5):
  - Baseline: 0.379*** (SE 0.098).
  - WGI ≥ 0: 0.629*** (SE 0.225).
  - WGI < 0: 0.344** (SE 0.155).
  - Reserves-to-imports ≥ 3: 0.282** (SE 0.110).
  - Reserves-to-imports < 3: 0.776*** (SE 0.180).
  - Concessional lending = 1: 0.708*** (SE 0.174).
  - Concessional lending = 0: 0.330*** (SE 0.107).
  - Arrears = 1: 0.245 (SE 0.174) — not significant.
  - Arrears = 0: 0.400*** (SE 0.110).
  - VIX ≥ 20: 0.192* (SE 0.115).
  - VIX < 20: 0.556*** (SE 0.122).
  - M2 gap ≥ 0: 0.427*** (SE 0.109).
  - M2 gap < 0: 0.267** (SE 0.119).
- Interpretation:
  - Relative debt matters more when institutional quality is higher, reserves are low, access to concessional finance exists, no recent arrears, during low VIX (“tranquil” periods), and when global liquidity is in excess.

### Robustness and alternative specifications
- Alternative relative-debt benchmarks and robustness results:
  - Debt frontier benchmark (Text Table 6): Relative debt coefficient 0.207** (SE 0.092).
  - Three-year rolling window: 0.186*** (SE 0.068).
  - Quadratic term (relative debt^2) statistically insignificant (Model 4) — supports linear specification.
- Robustness exercises (Text Table 7) show persistent positive significance:
  - Removing outliers: 0.337*** (SE 0.093).
  - Winsorization: 0.465*** (SE 0.107).
  - Pre-GFC (1993–2010) sample: 0.261* (SE 0.132).
  - Post-GFC (2011 onwards) sample: 0.608*** (SE 0.140).
  - Driscoll-Kraay SEs: 0.379*** (SE 0.099).
- Interaction results (Text Table 8):
  - VIX × Relative debt: -0.184*** (SE 0.048) — implies relative debt matters less when VIX is high.
  - M2 gap main effect: 0.014** (SE 0.006); M2 gap × Relative debt not significant.
  - Other interaction terms generally not significant.

### Additional controls: domestic and external factors
- Adding domestic controls (Text Table 9):
  - GDP per capita, WGI, Reserves-to-GDP, Primary balance-to-GDP, Exports-to-GDP — relative debt remains significant.
  - WGI example: -0.735*** (SE 0.199) when included.
  - Reserves-to-GDP example: -0.014** (SE 0.006).
- Adding external controls (Text Table 10):
  - Global M2-to-GDP, world GDP growth, Global EPU, Brent oil price — relative debt remains significant.
  - World GDP growth example: -0.046*** (SE 0.008).
  - Brent oil price example: -0.173*** (SE 0.057).
  - M2 not statistically significant in these specifications; VIX remains dominant external driver.

### Policy-relevant implications and recommendations
- Incorporate relative metrics in fiscal surveillance and debt sustainability analysis because investors price debt comparatively.
- Authorities should benchmark debt against peer medians or structural debt frontiers to assess market perceptions and potential borrowing-cost penalties.
- Policy frameworks relying solely on absolute fiscal metrics can miss market-driven reputational and peer-comparison effects, especially during tranquil periods when investors discriminate across borrowers.
- Strengthening institutional quality and external buffers (reserves) can reduce sensitivity of spreads to relative debt positioning.
- Be aware that during periods of abundant global liquidity, peer comparisons can amplify market reactions to relative debt; during risk-off episodes (high VIX), global sentiment can dominate and reduce the marginal role of relative comparisons.

### Contributions, extensions, and future research directions
- Reframes sovereign risk pricing through a relative lens aligned with investor behavior in integrated capital markets.
- Identifies domestic and global conditions that amplify the relevance of relative debt assessments.
- Suggests extensions: apply the relative framework to other drivers of sovereign spreads (growth, inflation, institutional quality) and examine differences between short-term and long-term effects.
- Proposes investigating how different types of debt arrears (official versus private) and histories of default affect investor pricing.

*IMF Working Paper: Peer Pressure: How Relative Debt Drives Emerging Market Sovereign Spreads (chapter 3 text extracted from the source PDF).*

### Introduction ...........................................................................................................

### IMF WORKING PAPERS Peer Pressure: How Relative Debt Drives Emerging Market Sovereign Spreads

### Introduction
- Sovereign bond spreads are market-based measures of sovereign credit risk, defined as the yield differential between sovereign bonds and benchmark securities, typically U.S. Treasuries of comparable maturity.
- Public debt is a prominent determinant of sovereign spreads; higher debt is conventionally associated with higher borrowing costs due to increased risk of fiscal stress or default.
- The paper emphasizes that investor assessments of sovereign debt risk are comparative: investors benchmark countries against peers with similar income levels, regional location, or economic structures, rather than assessing debt solely in absolute terms.
- The paper introduces "relative public debt" as a complementary/explanatory variable to traditional gross debt measures for emerging market and developing economies (EMDEs).

### Empirical approach and definitions
- Relative public debt is defined as a country’s deviation from the median debt level of a structurally similar peer group.
- Peer groups are classified by:
  - income per capita,
  - regional location,
  - commodity export exposure.
- The approach is distinct from debt-limit, fiscal-space, or “safe debt” concepts, which are grounded in structural or model-based assessments of sustainability; instead, it aims to capture the comparative benchmarking dimension used by investors and rating agencies.

### Key empirical statements tested
- The paper seeks to provide empirical evidence supporting the following statements:
  - Higher gross (absolute) debt levels are associated with higher sovereign spreads for EMDEs.
  - Higher relative debt (compared to peers) is associated with higher sovereign spreads for EMDEs.

### Framing and contribution
- The contribution lies in introducing and testing relative public debt as a novel explanatory variable for sovereign spreads, reflecting the hypothesis that sovereign risk pricing is inherently relative in interconnected global markets.
- The methodology is positioned as empirically driven and intended to capture market benchmarking behavior rather than structural sustainability thresholds.

*Source: IMF Working Paper — Introduction section*

### 3. The impact of relative debt on EMDEs’ sovereign spreads depends on domestic and global conditions.

### 3. The impact of relative debt on EMDEs’ sovereign spreads depends on domestic and global conditions.

### Key takeaways
- Relative debt—deviation of a country’s public debt-to-GDP ratio from a peer-group median—adds significant explanatory power to sovereign spread models, especially during “normal” or tranquil global financial conditions.
- Sovereign spreads reflect both absolute debt and relative positioning; when both are included, relative debt often dominates and gross debt can lose statistical significance.
- A ten percent increase in relative debt is predicted to increase spreads by 3.8 percent in the core specification (Model 6 by income classification), all else equal.
- Relative debt’s impact is heterogeneous: it matters more in countries with stronger institutions, low reserves, access to concessional finance, no recent arrears, during low VIX (risk-on) periods, and during periods of excess global liquidity.

### Literature and motivation
- Earlier panel studies find global financial conditions (e.g., U.S. interest rates, global risk aversion) and domestic fundamentals (public debt, fiscal balance) drive spreads.
- Recent work emphasizes the global financial cycle, index-driven investor flows, and peer benchmarking: examples include findings that a 10-percentage point increase in net debt-to-GDP is associated with a roughly 100 basis point rise in spreads.
- Index composition and flow shocks can move prices independently of fundamentals (example: a one percentage point reduction in supply can raise prices by 33 basis points).
- Empirical gap: limited direct research on relative debt metrics (deviations from regional or income-group medians) motivates this chapter.

### Theoretical intuition and conceptual framework
- Investors use comparative heuristics (peer benchmarking) and global portfolio allocation logic: sovereign bonds compete globally and investor differentiation varies with liquidity and risk appetite.
- Three reinforcing mechanisms:
  - Peer benchmarking and information constraints (investors compare to structurally similar countries).
  - Global portfolio allocation (search-for-yield in abundant liquidity; reduced differentiation during flight-to-quality).
  - Policy signaling and market discipline (debt below peers can signal prudence; exceeding peers raises rollover concerns).
- Empirical model frames spreads as function of absolute debt (D) and relative debt (ln(D) – ln(D*)), where D* is a benchmark debt-to-GDP derived from peers or a debt frontier.

### Stylized facts (empirical patterns)
- Stylized Fact #1: Relative positioning by income level shapes sovereign risk — median spread for low-income countries exceeds middle- and high-income peers; greater dispersion for low-income countries.
- Stylized Fact #2: Economic structure anchors risk perception — for commodity exporters, sovereign spreads move with oil prices (Figure 2).
- Stylized Fact #3: Positive correlation between relative debt and sovereign spreads (Figure 3).
- Stylized Fact #4: Relative debt outperforms absolute debt in univariate panel regressions (Text Table 1):
  - Debt coefficient: 0.232*** (SE 0.087) in column (1).
  - Relative debt (Income classification): 0.468*** (SE 0.101) in column (2).
  - Relative debt (Region): 0.280*** (SE 0.100) in column (3).
  - Relative debt (Commodity status): 0.355*** (SE 0.099) in column (4).
  - Observations: 1,197; Number of countries: 83; R-squared: 0.020, 0.065, 0.024, 0.035 respectively.

### Data and methodology
- Sample: J.P. Morgan EMBIG sovereign spreads for 83 emerging market countries from 1993 to 2024; annual average spreads; sample restricted to observations ≤ 1,500 basis points (approximately the 95th percentile).
- Relative debt definition (Equation 1):
  - Relative debt_{i,t} = ln(Total Public Debt-to-GDP_{i,t-1}) − ln(Median Total Public Debt-to-GDP_{j,t-1}), where j is the peer group (income, region, commodity-exporter status).
- Peer groups:
  - Income classification: World Bank (low, medium, high income).
  - Region: Moody’s regional classification (Text Table 2 lists 83 countries by region).
  - Commodity exporters: World Bank 2024 Global Economic Prospects list (Venezuela added).
- Controls:
  - Domestic: Real GDP growth (expected negative), inflation (expected positive).
  - Global: VIX (positive), U.S. 10-year Treasury yield (expected positive but empirical relationship varies).
- Baseline panel regressions: fixed effects (Hausman test), clustered standard errors; main specification (Equation 2) uses lagged debt measures and lagged domestic factors; VIX and US 10-year included contemporaneously.

### Baseline empirical results (Text Table 4 highlights)
- Core findings:
  - Relative debt (income classification) coefficient in baseline Model 6: 0.379*** (SE 0.098); Observations 1,189; R-squared 0.250; Number of countries 83.
  - When absolute debt and relative debt are included simultaneously (Models 9–11), relative debt remains robustly positive and significant while gross debt often loses significance or becomes unstable.
  - Control variables in Model 5 behavior:
    - Public debt: positively and significantly linked with spreads.
    - Real GDP growth: reduces spreads (example coefficients across models ~ -0.020 to -0.024, significant).
    - Inflation: increases spreads (coefficients ~ 0.014 to 0.016, significant).
    - VIX: positively and significantly related to spreads (e.g., 0.617***, SE 0.045).
    - US 10-year Treasury yield: insignificant and negative in some specifications (example -0.031, SE 0.064).
- Standardized impacts (Figure 4): a one-standard-deviation increase in relative debt raises sovereign spreads by roughly 0.2–0.3 standard deviations — comparable to global risk indicators.

### When does relative debt matter more? (split-sample evidence; Text Table 5)
- Six dimensions used to split sample: Institutional strength (WGI ≥ 0 vs WGI < 0), Reserve adequacy (months of imports ≥ 3 vs < 3), Access to concessional finance (World Bank Blend/IDA status), History of arrears (PPG arrears in past five years), Global risk aversion (VIX threshold 20), Global liquidity (global M2-to-GDP gap ≥ 0 vs < 0).
- Key split-sample coefficient estimates (relative debt):
  - Baseline: 0.379*** (SE 0.098).
  - WGI ≥ 0: 0.629*** (SE 0.225).
  - WGI < 0: 0.344** (SE 0.155).
  - Reserves-to-imports ≥ 3: 0.282** (SE 0.110).
  - Reserves-to-imports < 3: 0.776*** (SE 0.180).
  - Concessional lending = 1: 0.708*** (SE 0.174).
  - Concessional lending = 0: 0.330*** (SE 0.107).
  - Arrears = 1: 0.245 (SE 0.174) — not significant.
  - Arrears = 0: 0.400*** (SE 0.110).
  - VIX ≥ 20: 0.192* (SE 0.115).
  - VIX < 20: 0.556*** (SE 0.122).
  - M2 gap ≥ 0: 0.427*** (SE 0.109).
  - M2 gap < 0: 0.267** (SE 0.119).
- Interpretation:
  - Relative debt matters more when institutional quality is higher, reserves are low, access to concessional finance exists, no recent arrears, during low VIX (“tranquil” periods), and when global liquidity is in excess.

### Alternative measures and robustness checks
- Alternative relative-debt definitions:
  - Calculated debt frontier (Text Table 6): Relative debt coefficient 0.207** (SE 0.092) when benchmarked to a structural frontier; three-year rolling window: 0.186*** (SE 0.068).
  - Quadratic term (relative debt^2) is statistically insignificant (Model 4), supporting linear specification.
- Robustness exercises (Text Table 7):
  - Removing outliers, winsorization, using all historical observations, dependent variable ln(1+spreads), pre-GFC (1993–2010) and post-GFC (2011 onwards) samples, three-year averages, Driscoll-Kraay SEs, adding time trends — relative debt remains positive and typically statistically significant.
  - Examples of relative debt coefficients across robustness models:
    - Removing outliers: 0.337*** (SE 0.093).
    - Winsorization: 0.465*** (SE 0.107).
    - Pre-GFC sample: 0.261* (SE 0.132).
    - Post-GFC sample: 0.608*** (SE 0.140).
    - Driscoll-Kraay SEs: 0.379*** (SE 0.099).
- Interaction models (Text Table 8):
  - Relative debt remains significant across interaction specifications.
  - Significant interactions:
    - VIX × Relative debt: -0.184*** (SE 0.048) — implies relative debt matters less when VIX is high.
    - M2 gap: 0.014** (SE 0.006) positive main effect; M2 gap × Relative debt not significant.
  - Interaction terms generally not significant for other dimensions, suggesting nonlinearity or more complex environments.

### Additional controls: domestic and external factors
- Domestic controls (Text Table 9):
  - Adding GDP per capita, WGI, Reserves-to-GDP, Primary balance-to-GDP, Exports-to-GDP — relative debt remains significant.
  - WGI coefficient example: -0.735*** (SE 0.199) when included.
  - Reserves-to-GDP: -0.014** (SE 0.006) in one specification.
- External controls (Text Table 10):
  - Adding global M2-to-GDP, world GDP growth, Global EPU, Brent oil price — relative debt remains significant.
  - World GDP growth: negative and significant (example -0.046***, SE 0.008).
  - Brent oil price: associated with narrower spreads (example -0.173***, SE 0.057).
  - M2 not statistically significant in these specifications; VIX remains a dominant external driver.

### Empirical magnitudes and model fit highlights
- Text Table 4 (baseline Model 6): Relative debt 0.379***; Observations 1,189; R-squared 0.250.
- Text Table 1 univariate comparison shows higher R-squared for relative debt (Income classification: 0.065) than for Debt alone (0.020).
- In core specification (Model 6), a ten percent increase in relative debt → predicted 3.8 percent increase in spreads.
- VIX consistently large and significant driver: e.g., 0.617*** (SE 0.045) in baseline.

### Policy-relevant implications (as derived from findings)
- Fiscal surveillance and debt sustainability analysis should incorporate relative metrics in addition to absolute measures because investors price debt comparatively.
- Authorities may benefit from benchmarking debt against peer medians or structural debt frontiers to understand market perceptions and potential borrowing-cost penalties.
- Policy frameworks relying solely on absolute fiscal metrics can miss market-driven reputational and peer-comparison effects, particularly in integrated capital markets and during tranquil periods when investors discriminate across borrowers.
- Maintaining or improving institutional strength and external buffers (reserves) can reduce the sensitivity of spreads to relative debt positioning.
- During periods of abundant global liquidity, peer comparisons can amplify market reactions to relative debt; conversely, during risk-off episodes (high VIX), global sentiment can dominate and reduce the marginal role of relative comparisons.

*Source: IMF Working Paper — Peer Pressure: How Relative Debt Drives Emerging Market Sovereign Spreads (chapter 3 text extracted from the source PDF).*

### Conclusion

### Conclusion

### Main findings: relative debt and sovereign spreads
- Investors price sovereign risk comparatively, evaluating debt sustainability relative to peer groups rather than in isolation.
- Gross debt remains statistically significant, but relative debt—benchmarked by income classification, region, or commodity-exporter status—offers greater explanatory power.
- Income-based benchmarking consistently emerges as the most robust and intuitive specification.

### Robustness and temporal dynamics
- The income-based approach yields the highest within R², including when compared to models incorporating multiple structural characteristics.
- No evidence of non-linearity was found in the relationship between relative debt and spreads.
- The importance of relative debt has grown in the post-global financial crisis period, as rising global debt levels have made investors more selective.
- The coefficient on relative debt remains stable across specifications that control for macroeconomic and institutional variables.

### Contextual amplifiers (when relative debt matters more)
- Stronger institutions: relative debt matters most in countries with stronger institutions, where investors are more likely to differentiate based on fundamentals.
- Weak reserve buffers: impact is amplified when reserve buffers are weak, increasing concerns over liquidity and repayment capacity.
- Access to concessional finance: relative debt is more relevant for countries with access to concessional finance, where economic strength is typically lower and a greater share of liabilities may be non-restructurable.
- Global financial environment:
  - Relative debt plays a larger role during periods of low or moderate global risk aversion, when investors adopt a more fundamentals-based approach.
  - Relative debt is more influential during episodes of abundant global liquidity, consistent with search-for-yield behavior.

### Contributions to the literature
- Reframes sovereign risk pricing through a relative lens, aligning analysis more closely with investor behavior in integrated capital markets.
- Identifies domestic and global conditions that amplify the relevance of relative debt assessments.
- Notes that the framework, though focused on debt, could be extended to other drivers of sovereign spreads such as growth, inflation, or institutional quality.

### Directions for future research
- Explore potential differences between short-term and long-term effects of relative debt on sovereign spreads.
- Investigate how different types of debt arrears (arrears to official creditors versus arrears to the private sector) and a history of debt default drive investors’ pricing of sovereign debt risk.

### Annex note
- Annex I provides the peer-group classification used for commodity (resource-rich) exporters versus non-commodity exporters.

*IMF Working Paper: Peer Pressure: How Relative Debt Drives Emerging Market Sovereign Spreads (Working Paper No. WP/2026/110).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026110-source-pdf.pdf_