## 4. Spreads and Country Performance Before Issuance

## Source details

**Canonical URL:** [4. Spreads and Country Performance Before Issuance](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15275.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15275.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15275.pdf.json)

---

### Introduction and scope
- Sample: 104 EMDEs, 1995-2013; 49 issued international sovereign bonds at least once; 55 did not issue.
- Since 2005, 14 LIDCs have issued sovereign bonds in international markets; 10 in Sub-Saharan Africa.
- LIDCs issued US$4 billion in 2013; Côte d’Ivoire, Ethiopia, Ghana, Kenya, Senegal, Vietnam, and Zambia issued about US$8 billion in 2014.
- Research questions:
  - What determines LIDCs’ ability to issue bonds internationally?
  - What factors influence spreads at issuance?
  - What lessons can LIDCs draw from EMDEs with market access?
- Contribution: joint estimation of sovereign bond issuance (SBI) and spreads at issue using a two-step Heckman framework; inclusion of recent first-time LIDC issuances.

### Stylized facts and dataset
- Coverage and data sources:
  - 104 countries over 1995–2013; bond data from Bloomberg; macro and institutional data from WEO and WDI.
  - Domestic variables winsorized at 1st and 99th percentiles.
- Issuance patterns:
  - Sovereign bond issuances rose from an annual average of 8 per year in the late 1990s, to 12 in the 2000s and 20 since 2010.
  - During 1995-2013, LIDCs issued 20 sovereign bonds (8.4 percent of all SBIs in the sample).
  - Geographic concentration of issuances: 110 from Europe and Central Asia, 89 from Latin America and the Caribbean, 16 from Sub-Saharan Africa, 11 from Middle East and North Africa, 6 from South Asia, 5 from East Asia and Pacific.
- Spreads:
  - Average primary spread for the 18 LIDCs that issued international bonds: 434 basis points.
  - Average spread for EMDE issuers: 310 basis points; difference statistically significant.
  - Spreads declined during 2005-09 and rose thereafter; frontier markets face higher spreads than average EMDE.

### Empirical framework
- Two-step Heckman model:
  - Selection equation (SBI = 1): probit for probability of issuing in year t. Explanatory sets: GLOBAL (US 10-year treasury yield, VIX), DOMESTIC (real per capita GDP, GDP growth, inflation, current account/GDP, reserves in months, fiscal balance/GDP, external public debt/GDP, government effectiveness, IMF program), and POPULATION (log population as exclusion restriction).
  - Outcome equation (SPREAD): primary spread at issue observed only when SBI = 1; linear function of GLOBAL and DOMESTIC factors plus inverse Mills ratio (IMR).
- Identification and controls:
  - Population excluded from outcome equation (fixed cost rationale).
  - Regional dummies for 6 regions; global shocks alternatively modeled by US 10-year yield and VIX or year-fixed effects.
- Additional choices:
  - Persistence dummy = 1 if country issued at least once in previous three years.
  - Non-linear fiscal effect: interaction fiscal balance/GDP × real per capita GDP.
  - Robustness: alternative debt measures, financial development, resource-rich dummy, Chinn and Ito capital account openness, aid inflows/GDP, and domestic variables measured at t-1.

### Main empirical findings — selection and spreads
- Sample selection:
  - Wald test rejects orthogonality; inverse Mills ratio coefficient (ρ) negative and significant, indicating non-random participation.
  - Population (log) is a significant predictor of issuance — validates exclusion restriction.
  - Regional dummies and global factors significant — spillovers and global demand/supply matter.
- Global factors:
  - VIX: one standard deviation change in VIX corresponds to a 29 basis points change in spread (higher VIX → higher spreads) (Table notes: VIX coefficient examples include 4.782*** with SE 1.503 in column (1)).
  - US 10-year yield: a 1 percentage point reduction in the US rate translates into a 29 basis-point increase in primary spread; lower US rates associated with higher probability of issuance (examples: US rates coefficient -29.561*** (SE 9.523) in SPREAD column).
- Domestic macro and institutional determinants:
  - Real per capita GDP: higher levels increase likelihood of issuance (examples: SBI coefficient 0.873*** (SE 0.078) in column (2)); associated with lower spreads in some specifications (e.g., SPREAD -121.760*** (SE 18.878) in column (1)).
  - GDP growth: lower-growth countries issue at higher spreads (example SPREAD coefficient -12.538*** (SE 4.315)).
  - Inflation: generally not significant in baseline; positive association with spreads when domestic variables measured at t-1 in some specs.
  - IMF programs: IMF-supported lending arrangement in previous three years increases likelihood of issuance (catalytic role); example SBI coefficients positive and significant (e.g., 0.594** (SE 0.268) and robustness example 0.807*** (SE 0.267)). IMF program presence may be associated with higher spreads in some specifications (e.g., SPREAD 94.664** (SE 44.310) in column (1)).
  - External sector and liquidity:
    - Current account: lower deficits (more positive current account) associated with lower spreads (example SPREAD -5.631*** (SE 2.060)).
    - Reserves (months of imports): higher reserves associated with lower spreads (example SPREAD -11.813*** (SE 3.605)); reserves show weak or non-significant negative correlation with issuance probability in some specifications.
  - Fiscal position:
    - Public external debt (%GDP): higher external debt reduces probability of issuance and is associated with higher spreads when issuing (examples: SBI -0.004** (SE 0.002) and SPREAD 1.033** (SE 0.489) in column (9)).
    - Fiscal balance (%GDP): improvement in fiscal balance increases likelihood of issuance and lowers spreads; interaction with per capita GDP highlights stronger importance of fiscal discipline for poorer countries (examples: average one percentage point improvement → 3 basis-point drop in spreads on average; effect more than five times larger for lower income countries with per capita GDP around US$400 per contextual statement).
    - Table estimates include large negative SPREAD coefficients for fiscal balance in robustness (e.g., -96.671*** (SE 32.105)).
  - Past issuance/persistence:
    - Issuing countries are more likely to issue again and at lower spreads; repeat issuers have about 75 basis points lower spreads compared to non-recent issuers (examples: SBI in previous 3 years SPREAD -74.553** (SE 29.527); SBI coefficient 0.727*** (SE 0.124)).
  - Government effectiveness:
    - Greater government effectiveness increases probability of issuance and lowers spreads; one standard deviation increase associated with about 98–164 basis-point reduction in spreads across specifications (examples: SPREAD -163.807*** (SE 27.521) in column (7); robustness SPREAD -173.623*** (SE 27.718)).
    - Once institutional quality is controlled for, per capita GDP and IMF program effects on spreads lose significance in some specifications.
- Robustness results (selected):
  - Total public debt similar to external public debt: higher debt → lower issuance likelihood and higher spreads.
  - Financial development: higher private credit/GDP associated with lower probability of issuance (substitution away from external financing).
  - Resource-rich dummy: inclusion does not change baseline results materially.
  - Capital account openness (Chinn and Ito): more open capital accounts → significantly lower spreads.
  - Aid inflows/GDP: higher aid dependence associated with larger spreads.
  - Measuring domestic variables at t-1 yields similar qualitative results; inflation more often positive and significant for spreads in lagged specifications.

### Key statistics and summary moments (Tables 3 and 4; simple averages)
- Table 3 (sample means, 104 countries, 1995–2013; many variables winsorized):
  - SBI (0/1): Mean 0.109, St. Dev. 0.312, Obs 1,749.
  - SPREAD: Mean 309.590, St. Dev. 171.972, Min 21.600, Max 825.000, Obs 191.
  - Real per capita GDP (log): Mean 7.172, St. Dev. 1.119, Min 4.775, Max 9.445, Obs 1,749.
  - GDP growth: Mean 4.329, St. Dev. 3.621, Obs 1,749.
  - Inflation (CPI %): Mean 17.401, St. Dev. 76.563, Min -4.148, Max 1265.734, Obs 1,749.
  - PPG external debt (%GDP): Mean 58.491, St. Dev. 45.712, Obs 1,749.
  - Fiscal balance (%GDP): Mean -2.287, St. Dev. 4.223, Obs 1,749.
  - Current account (%GDP): Mean -5.369, St. Dev. 8.086, Obs 1,749.
  - Reserves (months): Mean 4.698, St. Dev. 4.257, Obs 1,749.
  - IMF program previous 3 years (dummy): Mean 0.150, St. Dev. 0.199, Obs 1,749.
  - Government effectiveness (WGI): Mean -0.482, St. Dev. 0.599, Obs 1,614.
  - US 10-year yield: Mean 4.320, St. Dev. 1.310, Obs 1,749.
  - VIX index: Mean 21.316, St. Dev. 6.088, Obs 1,749.
- Table 4 (country groups, 1995–2013 simple averages):
  - Real per capita GDP (log): No-issuance 6.72 (N 55); Occasional issuers 7.40 (N 31); Regular issuers 8.31 (N 18).
  - GDP growth: No-issuance 4.58; Occasional 4.49; Regular 3.53.
  - Private credit (%GDP): No-issuance 31.38; Occasional 40.86; Regular 49.99.
  - PPG external debt (%GDP): No-issuance 64.23; Occasional 54.73; Regular 49.46.
  - Current account (%GDP): No-issuance -6.88; Occasional -4.91; Regular -3.43.
  - Reserves (months): No-issuance 5.27; Occasional 3.86; Regular 5.08.
  - Population (logs): No-issuance 15.24; Occasional 15.44; Regular 16.48.
  - Government effectiveness: No-issuance -0.67; Occasional -0.40; Regular 0.05.

### Selected regression coefficient examples (Tables 5 and 6; two-step Heckman)
- US 10-year yield (SPREAD): -29.561*** (SE 9.523) in column (1); SBI coefficients show negative effects on issuance probability in some columns (e.g., -0.107*** (SE 0.040) column (2)).
- VIX (SPREAD): 4.782*** (SE 1.503) in column (1).
- Real per capita GDP (SPREAD): -121.760*** (SE 18.878) in column (1); SBI: 0.873*** (SE 0.078) in column (2).
- GDP growth (SPREAD): -12.538*** (SE 4.315) in column (1).
- PPG external debt (%GDP): SBI -0.004** (SE 0.002) in column (2); SPREAD 1.033** (SE 0.489) in column (9).
- Fiscal balance (%GDP): SPREAD 9.699* (SE 5.500) in column (1); SBI -0.085*** (SE 0.023) in column (2); robustness shows large negative SPREAD coefficients for fiscal balance in some columns (e.g., -96.671*** (SE 32.105)).
- Reserves (months) (SPREAD): -11.813*** (SE 3.605) in column (1).
- IMF program (SPREAD): 94.664** (SE 44.310) in column (1); SBI 0.594** (SE 0.268) in column (2).
- Government effectiveness (SPREAD): -163.807*** (SE 27.521) in column (7); SBI 0.226* (SE 0.129) in column (8).
- SBI in previous 3 years (SPREAD): -74.553** (SE 29.527) in column (5); SBI 0.727*** (SE 0.124) in column (6).
- Observations typically 1,749; for specifications with government effectiveness, observations drop to 1,614.

### Conclusions — messages for LIDCs
- Determinants of spreads and market access:
  - Stronger budget balances and stronger external positions (current account and international reserves) are associated with lower spreads.
  - Lower global market volatility (VIX) and favorable global conditions reduce spreads; lower US rates increase issuance probability but can be associated with higher primary spreads.
  - Fiscal discipline matters for both market access and costs: improved fiscal balance increases likelihood of issuance and lowers spreads; higher public debt reduces likelihood and raises spreads (example: a 10 percent increase in external debt-to-GDP associated with a 10–12 basis-point increase in primary spread in some specifications).
  - Government effectiveness strongly linked to higher issuance probability and substantially lower spreads (one standard deviation improvement associated with roughly 98–174 basis-point reduction across specifications).
  - Larger countries and those with higher per capita GDP have greater market access and, in many specifications, lower spreads.
  - Past issuance lowers subsequent spreads (persistence effect: repeat issuance associated with about 75 basis points lower spreads).
- Policy implications for LIDCs (implied by findings):
  - Strengthen fiscal balances and manage public debt to improve market access and reduce borrowing costs.
  - Build international reserves and external buffers to lower spreads.
  - Improve institutional quality and government effectiveness to broaden market access and reduce spreads.
  - Consider the catalytic role of IMF-supported programs for achieving initial market access, while recognizing potential short-term spread effects.

*Source: _wp15275 - 4. Spreads and Country Performance Before Issuance (1995–2013 sample analysis).*

### 1. Sovereign Bond Issuances by LIDCs ..................................................... 20

### 1. Sovereign Bond Issuances by LIDCs ..................................................... 20

### 2. Sample
- "Sample" section listed on page 21.

### 3. Variables’ Definition, Sources, and Summary Statistics
- "Variables’ Definition, Sources, and Summary Statistics" section listed on page 22.

### 4. Country Characteristics by Access Frequency Groups
- "Country Characteristics by Access Frequency Groups" section listed on page 23.

### 5. Regression Results: Baseline
- "Regression Results: Baseline" section listed on page 24.

### 6. Regression Results: Robustness
- "Regression Results: Robustness" section listed on page 25.

### List of Figures
- Figure 1: "Sovereign Bonds and Syndicated Loans to the Public Sector" — page 26.
- Figure 2: "Sovereign Bond Issuances and Global Conditions, 1995-2013" — page 26.
- Figure 3: "The Distribution of Bond Spread at Issue, 1995-2013" — page 26.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15275.pdf*

### 4. Spreads and Country Performance Before Issuance....................................

### 4. Spreads and Country Performance Before Issuance....................................

### Introduction
- Context and scope
  - Since 2005, some 14 LIDCs have issued sovereign bonds in the international capital markets, 10 of which are in Sub-Saharan Africa.
  - This paper examines 104 EDMEs during the period 1995-2013, including 49 that had issued sovereign bonds at least once in the international capital markets and 55 that never issued.
  - In 2013, LIDCs issued sovereign bonds amounting to US$4 billion. In 2014, Côte d’Ivoire, Ethiopia, Ghana, Kenya, Senegal, Vietnam, and Zambia issued bonds totaling about US$8 billion.
- Research aims
  - What determines the ability of LIDCs to issue bonds internationally?
  - What factors influence spreads on these bonds?
  - What lessons can LIDCs draw from EMDEs with market access?
- Contributions
  - Joint estimation of determinants of sovereign bond issuance (SBI) and the spreads on those bonds.
  - Inclusion of recent first-time issuances by several LIDCs in the analysis.

### Stylized facts and dataset
- Dataset and coverage
  - Sample: 104 EDMEs, 1995-2013; 49 issued international sovereign bonds at least once; 55 used as control group (countries with GDP per capita lower than US$10,000 in 1995 that did not issue).
  - Data on sovereign bonds from Bloomberg (date of issuance, maturity, amount, yield and spread at issue).
  - Macroeconomic and institutional variables from the World Economic Outlook database and the World Development Indicators.
  - Domestic variables winsorized at the 1st and 99th percentiles: GDP growth, inflation, private credit, total debt, fiscal balance, current account, reserves.
- Issuance patterns
  - Sovereign bond issuances rose from an annual average of 8 per year in the late 1990s, to 12 in the 2000s and 20 since 2010.
  - During 1995-2013, LIDCs issued 20 sovereign bonds (8.4 percent of all SBIs in the sample).
  - Geographic concentration of issuances: 110 issuances from Europe and Central Asia, 89 from Latin America and the Caribbean, 16 from Sub-Saharan Africa, 11 from Middle East and North Africa, 6 from South Asia, 5 from East Asia and Pacific.
- Spreads and frontier markets
  - Average primary spread for the 18 LIDCs that issued international bonds was 434 basis points.
  - Average spread of the bonds issued by EMDEs was 310 basis points. The difference is statistically significant.
  - Spreads at issue declined during 2005-09 and picked up thereafter, with frontier markets placing bonds at higher spreads than average EMDE.
- Cross-country comparisons (three groups)
  - No-issuance countries (NI): 55 countries that did not issue sovereign bonds during 1995-2013.
  - Occasional issuers (OI): 31 countries that issued sovereign bonds for less than 5 years.
  - Regular issuers (RI): 18 countries that issued sovereign bonds for 5 or more years during 1995-2013.
- Simple correlations (three-year averages prior to issuance)
  - Lower primary yield spreads are associated with:
    - Faster growth
    - Higher per capita real GDP
    - Stronger institutions (government effectiveness)
    - Current account surpluses
    - Higher international reserves
  - Limited simple correlation between primary spreads and fiscal position (government balance or external public debt).

### Empirical analysis — specification
- Two-step framework (Heckman sample selection)
  - Selection equation: probit for probability of issuing a sovereign bond in year t (SBI = 1).
    - Explanatory sets: GLOBAL (10-year US treasury yield, VIX index), DOMESTIC (per capita real GDP, GDP growth, inflation, current account/GDP, reserves in months of imports, fiscal balance/GDP, external public and publicly guaranteed debt/GDP, government effectiveness, IMF program participation), and POPULATION (log of population as exclusion restriction).
    - Domestic variables measured as averages in the 3-year period prior to year of issuance.
  - Outcome equation: primary spread at issue (SPREAD) observed only when SBI = 1; linear function of same global and domestic factors plus inverse Mills ratio (IMR) from selection equation.
- Identification assumptions and controls
  - Population included in selection but excluded from outcome equation (fixed cost argument).
  - Regional dummy variables for East Asia and Pacific, Europe and Central Asia, Latin America and Caribbean, Middle East and North Africa, South Asia, and Sub-Saharan Africa.
  - Global shocks alternatively modeled by contemporaneous US 10-year yield and VIX or by year-fixed effects.
- Additional modeling choices
  - Persistence: dummy = 1 if country issued at least once in previous three years.
  - Non-linear fiscal effect: interaction between fiscal balance/GDP and per capita real GDP.
  - Robustness: alternative debt measures, financial development, resource-rich dummy, Chinn and Ito de jure financial openness, aid inflows/GDP, and measuring domestic variables at t-1.

### Empirical analysis — main results
- Sample selection and overall findings
  - Wald test rejects orthogonality of errors; inverse Mills ratio coefficient (ρ) is negative and statistically significant, indicating non-random participation and selection bias.
  - Country size (population) is a significant predictor of sovereign bond issuance, validating exclusion restriction.
  - Regional dummies and global factors statistically significant — spillovers and global demand/supply factors matter.
- Global factors
  - Market volatility (VIX): a one standard deviation change in the VIX index corresponds to a 29 basis points change in spread (higher volatility → higher spreads).
  - US 10-year treasury yield: a 1 percentage point reduction in the US rate translates into a 29 basis-point increase in the primary spread; lower US rates associated with a higher probability of issuance.
- Domestic macro and institutional determinants
  - Per capita real GDP: higher levels increase likelihood of issuance.
  - GDP growth: lower-growth countries issue at higher spreads.
  - Inflation: coefficients generally not statistically significant in baseline; in some specifications positive association with spreads when domestic variables measured at t-1.
  - IMF programs: countries with an IMF-supported lending arrangement in the previous three years are more likely to issue (catalytic role).
  - External sector and liquidity:
    - Lower current account deficits and higher international reserves → lower spreads.
    - No strong evidence that external position robustly affects probability of issuance; reserves show a non-significant negative correlation with issuance probability.
  - Fiscal position:
    - Public external debt-to-GDP: more indebted countries are less likely to issue and pay higher spreads. A 10 percent increase in the external debt-to-GDP ratio is associated with 10-12 basis-point increase in the primary spread (columns 7-9).
    - Government budget balance: in selection equation, a negative coefficient on budget balance suggests higher demand for external borrowing when fiscal deficit is larger; interaction with per capita GDP shows fiscal discipline especially important for poorer countries.
    - Non-linearity: a one percentage point improvement in fiscal balance → 3 basis-point drop in spreads on average; effect is more than five times larger for lower income countries (with a per capita GDP of about US$400).
  - Past issuance/persistence:
    - After first issuance, countries are more likely to issue again and at lower spreads. Reduction in spreads for repeat issuers is about 75 basis points compared to countries that did not issue in the previous three years.
  - Government effectiveness:
    - Greater government effectiveness increases probability of issuance and lowers cost. A one standard deviation increase in government effectiveness is associated with a 98 basis-point reduction in primary spreads.
    - Once controlling for institutional quality, per capita GDP and past IMF program effects on spreads are not significant.
- Robustness checks
  - Total public debt vs external public debt: similar effects — higher debt ratios → lower likelihood to issue and higher spreads.
  - Financial development: higher private credit/GDP associated with a lower probability of issuance (less reliance on external financing).
  - Resource-rich dummy inclusion does not alter baseline results.
  - Chinn and Ito de jure financial openness: more open capital accounts → significantly lower spreads.
  - Aid inflows/GDP: higher aid dependence associated with larger spreads (markets may perceive aid recipients as riskier).
  - Measuring domestic variables at t-1 yields similar results; inflation shows significant positive association with spreads in that specification.

### Conclusions — messages for low-income developing countries (LIDCs)
- Bond spreads
  - Stronger budget balance positions and stronger external positions (current account balance and international reserves) are associated with lower spreads.
  - Spreads are lower in periods of declining global market volatility.
- Bond issuance (market access)
  - Fiscal position matters for market access: an improved government budget balance increases the likelihood of issuance.
  - Higher public debt levels reduce likelihood of issuance and, if issuing, countries pay a premium.
  - Strong government effectiveness is associated with both greater likelihood of issuance and lower spreads.
  - Larger countries and countries with higher per capita GDP have greater market access than smaller and less developed peers.

*Source: _wp15275 - 4. Spreads and Country Performance Before Issuance (PDF chapter/section), 1995-2013 sample analysis*

### REFERENCES

### _wp15275 - REFERENCES

### Major cited works
- Abbas, S.A., Belhocine, N., El-Ganainy, A., and Horton, M., (2011). “Historical Patterns and Dynamics of Public Debt—Evidence From A New Database”, IMF Economic Review, 59(4): 717-742.
- AfDB, OECD and UNDP, (2014). “African Economic Outlook”, African Development Bank, Organization for Economic Co-operation and Development, United Nations Development Program, Tunis, Paris and New York.
- Alessandro, M., Sandleris, G., and Van der Ghote, A., (2011). “Sovereign Defaults and the Political Economy of Market Re-access,” Universidad Torcuato Di Tella Working Paper, No. 08/2011.
- Arora, V., and Cerisola, M., (2001). “How Does U.S. Monetary Policy Influence Sovereign Spreads in Emerging Markets?”, IMF Staff Papers, 48(3): 474-498.
- Bellas D., Papaioannou, M. G., and Petrova, I., (2010). “Determinants of Emerging markets sovereign bond spreads: fundamentals vs financial stress”, IMF Working Paper, No. 10/281.
- (Full list of references continues as provided in the source.)

### Table 1 — International Sovereign Bond Issuances by LIDCs (selected entries; Source: Bloomberg; Updated to end-December 2014)
- Moldova 6/6/97: Yield at issue 9.88, Tenor 5, Amt (USD mn) 75, Spread (bps) 340.0, S&P rating at issue Not rated
- Vietnam 10/27/05: Yield at issue 7.25, Tenor 10, Amt (USD mn) 750, Spread (bps) 256.4, S&P rating at issue BB-
- Ghana 9/27/07: Yield at issue 8.50, Tenor 10, Amt (USD mn) 750, Spread (bps) 387.0, S&P rating at issue B+
- Côte d'Ivoire 4/8/10: Yield at issue 17.35, Tenor 23, Amt (USD mn) 2330, Spread (bps) 393.0, S&P rating at issue Not rated
- Zambia 9/13/12: Yield at issue 5.63, Tenor 10, Amt (USD mn) 750, Spread (bps) 383.6, S&P rating at issue B+
- Kenya 6/24/14: Yield at issue 6.88, Tenor 10, Amt (USD mn) 1500, Spread (bps) 429.0, S&P rating at issue B+
- Ghana 9/11/14: Yield at issue 8.25, Tenor 11, Amt (USD mn) 1000, Spread (bps) 572.0, S&P rating at issue B-
- Vietnam 11/6/14: Yield at issue 4.80, Tenor 10, Amt (USD mn) 1000, Spread (bps) 238.7, S&P rating at issue BB-
- Ethiopia 12/4/14: Yield at issue 6.625, Tenor 10, Amt (USD mn) 1000, Spread (bps) 435.6, S&P rating at issue B

### Table 2 — Sample (SBI indicator)
- SBI = 1 for countries that issued at least a sovereign bond between 1995 and 2013; SBI = 0 otherwise.
- Examples:
  - Albania: 1
  - Algeria: 0
  - Bangladesh: 0
  - Bolivia: 1
  - Brazil: 1
  - Côte D'Ivoire: 1
  - Kenya: 0 (in this table)
  - Nigeria: 1
  - Vietnam: 1
- Notes: Table lists SBI status for 104 countries over 1995–2013.

### Table 3 — Variables’ Definition, Sources and Summary Statistics (annual data for 104 countries, 1995–2013; many variables winsorized at 1st and 99th percentiles)
- SBI (0/1): Mean 0.109, St. Dev. 0.312, Min 0, Max 1, Obs 1,749 (Bloomberg)
- SPREAD: Mean 309.590, St. Dev. 171.972, Min 21.600, Max 825.000, Obs 191 (Bloomberg)
- Real per capital GDP (log): Mean 7.172, St. Dev. 1.119, Min 4.775, Max 9.445, Obs 1,749 (WDI)
- GDP growth: Mean 4.329, St. Dev. 3.621, Min -11.967, Max 21.060, Obs 1,749 (WEO)
- Inflation (CPI, annual % change): Mean 17.401, St. Dev. 76.563, Min -4.148, Max 1265.734, Obs 1,749 (WEO)
- PPG external debt (%GDP): Mean 58.491, St. Dev. 45.712, Min 2.203, Max 304.203, Obs 1,749 (WDI)
- Fiscal balance (%GDP): Mean -2.287, St. Dev. 4.223, Min -16.403, Max 20.123, Obs 1,749 (WEO and country reports)
- Current account (%GDP): Mean -5.369, St. Dev. 8.086, Min -34.795, Max 21.405, Obs 1,749 (WEO)
- Reserves (months of imports): Mean 4.698, St. Dev. 4.257, Min 0.031, Max 30.307, Obs 1,749 (WDI, WEO and country reports)
- IMF program in the previous 3 year (dummy): Mean 0.150, St. Dev. 0.199, Min 0, Max 1, Obs 1,749 (IMF historical data set)
- Resource rich dummy: Mean 0.200, St. Dev. 0.400, Min 0, Max 1, Obs 1,749 (IMF (2012))
- Population (logs): Mean 15.516, St. Dev. 1.843, Min 10.642, Max 20.923, Obs 1,749 (WDI)
- Government effectiveness (WGI): Mean -0.482, St. Dev. 0.599, Min -1.929, Max 1.278, Obs 1,614
- Total debt (%GDP): Mean 65.306, St. Dev. 50.900, Min 5.474, Max 328.583, Obs 1,747 (WDI)
- Private credit (%GDP): Mean 36.479, St. Dev. 30.205, Min -40.464, Max 149.421, Obs 1,745 (Abbas and others (2011))
- Aid (% GDP): Mean 7.532, St. Dev. 9.032, Min -0.016, Max 100.386, Obs 1,708 (WDI)
- US rates (10-year yield): Mean 4.320, St. Dev. 1.310, Min 1.800, Max 6.570, Obs 1,749 (Federal Reserve)
- VIX index: Mean 21.316, St. Dev. 6.088, Min 12.389, Max 32.693, Obs 1,749 (CBOE)

### Table 4 — Country Characteristics by Access Frequency Groups (simple averages 1995–2013 across 104 countries)
- Per capita real GDP: All countries Mean 7.20 (N 104); No-issuance Mean 6.72 (N 55); Occasional issuers Mean 7.40 (N 31); Regular issuers Mean 8.31 (N 18). Tests: NI-OI ***, NI-RI ***, OI-RI ***.
- GDP growth: All Mean 4.37; No-issuance 4.58; Occasional 4.49; Regular 3.53. Tests: NI-RI * ; OI-RI *.
- Inflation: All Mean 12.37; No-issuance 14.05; Occasional 11.15; Regular 9.38.
- Private credit (% GDP): All Mean 37.43; No-issuance 31.38; Occasional 40.86; Regular 49.99. Test OI-RI **.
- PPG External debt (% GDP): All Mean 58.84; No-issuance 64.23; Occasional 54.73; Regular 49.46.
- Current account (% GDP): All Mean -5.69; No-issuance -6.88; Occasional -4.91; Regular -3.43. Test OI-RI **.
- Reserves (months of imports): All Mean 4.81; No-issuance 5.27; Occasional 3.86; Regular 5.08. Test NI-RI *.
- Resource-rich country: All Mean 0.21; No-issuance 0.26; Occasional 0.26; Regular 0.00. Tests: NI-OI ** ; NI-RI ** ; OI-RI **.
- Population (logs): All Mean 15.52; No-issuance 15.24; Occasional 15.44; Regular 16.48. Tests: NI-OI ** ; NI-RI **.
- IMF program in previous 3 years: All Mean 0.14; No-issuance 0.12; Occasional 0.15; Regular 0.17. Test NI-RI *.
- Government effectiveness: All Mean -0.47; No-issuance -0.67; Occasional -0.40; Regular 0.05. Tests: NI-OI ** ; NI-RI *** ; OI-RI ***.

### Table 5 — Regression Results: Baseline (two-step Heckman; dependent variables SPREAD and SBI; coefficients and robust standard errors shown)
- Specification summary (columns 1–10 present various specifications; Area FE included; some columns include Year FE):
  - US rates: Coefficient examples:
    - Column (1) SPREAD: -29.561*** (SE 9.523)
    - Column (2) SBI: -0.107*** (SE 0.040)
    - Column (9) SPREAD: -10.853 (SE 8.827)
    - Column (10) SBI: -0.054 (SE 0.047)
  - VIX index:
    - Column (1) SPREAD: 4.782*** (SE 1.503)
    - Column (2) SBI: -0.002 (SE 0.008)
  - Real per capita GDP:
    - Column (1) SPREAD: -121.760*** (SE 18.878)
    - Column (2) SBI: 0.873*** (SE 0.078)
    - Column (9) SPREAD: -12.059 (SE 22.822); Column (10) SBI: 0.480*** (SE 0.130)
  - GDP growth:
    - Column (1) SPREAD: -12.538*** (SE 4.315)
    - Column (2) SBI: -0.017 (SE 0.018)
  - PPG external debt (%GDP):
    - Column (1) SPREAD: 0.323 (SE 0.467)
    - Column (2) SBI: -0.004** (SE 0.002)
    - Column (9) SPREAD: 1.033** (SE 0.489)
  - Fiscal balance (%GDP):
    - Column (1) SPREAD: 9.699* (SE 5.500)
    - Column (2) SBI: -0.085*** (SE 0.023)
    - Column (3) SPREAD: -88.350*** (SE 30.611); Column (4) SBI: 0.356** (SE 0.154)
  - Current account (%GDP):
    - Column (1) SPREAD: -5.631*** (SE 2.060)
    - Column (2) SBI: -0.008 (SE 0.009)
  - Reserves (months): negative association with SPREAD:
    - Column (1) SPREAD: -11.813*** (SE 3.605)
    - Column (2) SBI: -0.027* (SE 0.015)
  - IMF program in previous 3 years:
    - Column (1) SPREAD: 94.664** (SE 44.310)
    - Column (2) SBI: 0.594** (SE 0.268)
  - Population:
    - Positive and significant for SBI in multiple columns, e.g., Column (2): 0.466*** (no SE shown in table snippet)
  - Interaction Fiscal balance * real pc GDP:
    - Column (3) SPREAD: 12.006*** (SE 3.708); Column (4) SBI: -0.056*** (SE 0.019)
  - SBI in previous 3 years:
    - Column (5) SPREAD: -74.553** (SE 29.527); Column (6) SBI: 0.727*** (SE 0.124)
  - Government effectiveness:
    - Column (7) SPREAD: -163.807*** (SE 27.521); Column (8) SBI: 0.226* (SE 0.129)
- Observations: commonly 1,749 (lower in specifications using Government effectiveness: 1,614)
- ρ reported (sample): e.g., -0.414 (col 1); Wald test p-values provided (e.g., 0.018 for column 1)
- Notes: Dependent variable selection equation SBI; outcome equation SPREAD. US yield and VIX measured at time t; other variables are averages between t-3 and t-1. Model estimated by two-step Heckman.

### Table 6 — Regression Results: Robustness (multiple alternative specifications; two-step Heckman)
- Real per capita GDP (examples across columns):
  - Column (1) SPREAD: -1.962 (SE 25.309); Column (2) SBI: 0.516*** (SE 0.112)
  - Column (7) SPREAD: -39.002** (SE 18.235); Column (8) SBI: 0.695*** (SE 0.106)
- GDP growth: consistently negative and significant for SPREAD (e.g., Column (1) -12.107*** (SE 3.754))
- Inflation: mixed signs; some SBI coefficients negative and significant at 10% (e.g., Column (2) -0.003*)
- PPG External debt (%GDP): positive and significant for SPREAD in many columns (e.g., Column (1) 1.278*** (SE 0.446)); negative and significant for SBI (e.g., -0.005**)
- Fiscal balance (%GDP): large negative coefficients for SPREAD in many columns (e.g., Column (1) -96.671*** (SE 32.105)) and positive for SBI (e.g., 0.447***)
- Fiscal balance * real pc GDP: positive and significant for SPREAD and negative for SBI across several columns (e.g., 11.704*** and -0.061***)
- Reserves: negative and significant for SPREAD (e.g., Column (1) -9.386*** (SE 3.003)) and negative for SBI in some columns (e.g., -0.030*)
- Government effectiveness: consistently negative and significant for SPREAD (e.g., -173.623*** (SE 27.718)); positive for SBI but significance varies
- IMF program in previous 3 year: SBI coefficients positive and strongly significant in many specifications (e.g., Column (2) 0.807*** (SE 0.267))
- SBI in previous 3 year: negative and significant for SPREAD in several columns (e.g., -58.861** (SE 29.239)) and positive for SBI (e.g., 0.747***)
- Additional robustness checks include Total debt (%GDP), Credit (%GDP), Natural resource-rich dummy, Capital account openness, Aid (%GDP) in various columns with reported coefficients (see table for exact values).
- Observations vary by specification (e.g., 1,614; 1,612; 1,605)
- Year and Area fixed effects included in robustness specifications; Wald test p-values reported (many p-values 0.000 for FE tests).

### Figures (descriptive notes)
- Figure 1: International Sovereign Bonds and Syndicated Loans to the Public Sector — Notes: annual data for 104 emerging markets and developing economies, over 1995-2013. GDP-weighted averages. Syndicated loans include central government and state/provincial authorities. Source: Dealogic Loan Analytics and Bloomberg.
- Figure 2: International Sovereign Bond Issuances and Global Conditions, 1995-2013 — Notes: annual data for 104 developing countries (49 issuers), over 1995-2013.
- Figure 3: The Distribution of Bond Spread at Issue, 1995-2013 — Notes: data for 202 SBIs (by 49 countries), over 1995-2013.
- Figure 4: Spreads and Country Performance Before Issuance — Notes: data for 185 SBIs (by 49 countries), over 1995-2013.

*Source: _wp15275 - REFERENCES.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15275.pdf_