## 2018. Issuances have increased not only for emerging

## Source details

**Canonical URL:** [2018. Issuances have increased not only for emerging](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020230-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020230-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020230-print-pdf.pdf.json)

---

### Research question and context
- Empirical question: whether access to external private financing (international bonds and external commercial loans) affects low-income and emerging market economies’ ability to collect tax revenue.
- Focus: external private financing (bond and commercial loans), not domestic private financing.
- Period studied: 2004–2018.
- Sample: 72 emerging market and developing economies.

### Data and methodology
- Dataset: balanced panel of 72 emerging market and developing economies over 2004–2018.
- Tax revenue data: IMF World Economic Outlook.
- International bond issuance and external commercial debt disbursements: World Bank’s International Debt Statistics (IDS) database (coverage includes public and publicly guaranteed debt; country reporting may differ).
- Main empirical approach: entropy balancing (Hainmueller 2012) to compare market-access and non-market-access countries with similar observable characteristics.
- Robustness checks: panel fixed-effect and GMM estimations.
- Analysis extension: includes external commercial loans in the market-access concept and covers 2004–18.

### Key empirical findings and descriptive statistics
- Total private financing (disbursements from international bond market issuance and external commercial debt) rose from USD 72 billion in 2004 to a peak of US$ 296 billion in 2017.
- Increase driven mainly by a rise in international bonds.
- Number of countries issuing bonds:
  - averaged 18 per year in 2004–6;
  - averaged 29 countries per year in the last three years of the sample period.
- Market access via debut bond market issuance:
  - 24 countries in the sample gained market access via a debut bond market issuance over 2004–18;
  - 21 of those made subsequent issues in the years following their debut.
  - Debut issuances averaged 2.7 percent of GDP;
  - Sample average of all issues averaged 1.8 percent of GDP.
- Tax revenue trends:
  - Median tax revenue rose from 13.7 percent of GDP in 2013 to 15.3 percent of GDP in 2018.
  - Countries issuing bonds regularly (at least two thirds of the sample years) have average tax ratios of 19.4 percent of GDP;
  - Occasional issuers average 15.6 percent of GDP;
  - Non-issuers average 13.4 percent of GDP.
- Correlations and dynamics:
  - Tax levels are positively correlated with the level of private financing (bonds and external commercial loans as a share of GDP) from the previous year.
  - Tax levels tend to increase following bond market issuances with a reported median cumulative change (figure referenced).

### Principal empirical conclusion
- No evidence that access to bond markets or external commercial loans undermines countries’ efforts to collect tax revenue.
- Access to markets is found to have a positive impact on domestic revenue mobilization in the analysis.

### Quantitative treatment-effect estimates (entropy balancing results)
- Estimated average treatment effect of market access on DRM is positive and economically meaningful: about 1 percentage point of GDP on average.
- Entropy balancing estimates (Table 3) for the Market Access dummy:
  - Column [1]: 1.044 ** (standard error (0.444))
  - Column [2]: 0.943 *** (standard error (0.352))
  - Column [3]: 0.928 *** (standard error (0.353))
  - Column [4]: 1.114 *** (standard error (0.354))
- Robustness and alternative entropy balancing specifications (examples from Table 4 and Table 5) include significant positive coefficients:
  - Market Access dummy examples: 1.027 ***, 1.006 ***, 0.991 ***, 1.320 ***, 2.884 ***, 3.663 ***, 3.913 ***, 3.459 ***, 3.358 ***, 3.318 ***
  - Alternative treatment definitions (Table 5): 1.263 *** (0.370); 0.969 *** (0.359); 1.633 *** (0.355); 1.829 *** (0.348)
- Panel regressions provide mixed support:
  - Fixed effects and GMM estimations often yield smaller and sometimes statistically insignificant positive effects compared with entropy balancing estimates.
  - Example fixed-effect private finance/GDP (t-1) coefficients: 0.201** (standard error (0.03)); 0.206* (standard error (0.05)).
- Descriptive sample moments (pre-weighting means, Table 2):
  - Grants-to-GDP (t-1): Market Access 2.538; No Market Access 3.883; Difference 1.345 (t-test p-val 0.000)
  - GDP per capita, log (t-1): Market Access 8.041; No Market Access 7.071; Difference -0.969 (t-test p-val 0.000)
  - Trade openness, log (t-1): Market Access 4.288; No Market Access 4.262; Difference -0.026 (t-test p-val 0.222)
  - Agriculture-to-GDP (t-1): Market Access 22.434; No Market Access 22.343; Difference 9.909 (t-test p-val 0.000)
  - Political risk (t-1): Market Access 67.403; No Market Access 61.976; Difference -5.426 (t-test p-val 0.000)
  - Inflation (t-1): Market Access 6.368; No Market Access 7.190; Difference 0.822 (t-test p-val 0.060)
  - Observations: Market Access 832; No Market Access 681
- After entropy balancing, synthetic control shows no significant pretreatment differences remain.

### Heterogeneity results (selected estimates, Table 8)
- By country income:
  - Market Access, non-LICs: 1.603 *** (0.415)
  - Market Access, LICs: 0.182 (0.576)
- By type of market access:
  - International bond markets: 3.890 *** (0.502)
  - External commercial debt: 0.650 (0.435)
- By issuance frequency:
  - Regular bond issuers: 3.240 *** (0.584)
  - Occasional bond issuers: -0.650 (0.435)
- By commodity dependence:
  - Market Access, commodity exporters: 1.811 *** (0.531)
  - Market Access, non-commodity exporters: 0.805 ** (0.395)
- By political cycle:
  - Market Access in non-election periods: 1.075 *** (0.425)
  - Market Access in election periods: -0.229 (0.786)

### Interpretation and plausible channels
- Market discipline channel: international bond issuances—and regular, transparent issuances—appear to strengthen incentives for DRM, consistent with bond vigilante effects and demands for creditworthiness.
- Institutional and macroeconomic channels: higher institutional quality (ICRG Political Risk index), macroeconomic stability (lower inflation), and stronger fundamentals are associated with higher tax revenue performance and with market access.
- Cross-sectional differences imply that market access effects depend on country characteristics (income level, bond market participation, issuance regularity, commodity dependence, election timing).
- Caveats: results are conditioned on empirical strategies (entropy balancing, panel FE, GMM) and the sample/period (2004–18); reverse causality and endogeneity remain identification challenges.

### Policy implications and recommendations
- Access to private financing should not be assumed to weaken tax collection incentives; in many cases, market access is associated with higher DRM.
- Countries should:
  - continue enhancing tax administration and tax policy to increase domestic revenue mobilization;
  - continue strengthening debt management practices to manage debt risks associated with market access, particularly exchange rate and interest rate risks.
- Recognize that market access may impose discipline that supports DRM, particularly for regular international bond issuers and non-LICs.
- Caution: given mounting concerns over the sustainability of public debt in many countries, market access is not an unambiguous substitute for robust domestic revenue and prudent debt management.

### Areas for further research (from the source)
- Analyze impact of market access on subcomponents of tax revenue (e.g., income tax, trade taxes).
- Study timing: how long after market access do revenue mobilization effects materialize.
- Consider business cycle effects (e.g., cyclically adjusted revenues).
- Note on Sub-Saharan Africa (Table 9): some significance found when controlling for external debt levels, but results influenced by small sample size and outliers.

*Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020230-print-pdf.pdf (wpiea2020230-print-pdf).*

### 2018. Issuances have increased not only for emerging

### wpiea2020230-print-pdf - 2018. Issuances have increased not only for emerging

### Research question and context
- Empirical question: whether access to external private financing (international bonds and external commercial loans) affects low-income and emerging market economies’ ability to collect tax revenue.
- Focus: external private financing (bond and commercial loans), not domestic private financing.
- Period studied: 2004–2018.
- Sample: 72 emerging market and developing economies.

### Data and methodology
- Dataset: balanced panel of 72 emerging market and developing economies over 2004–2018.
- Tax revenue data: IMF World Economic Outlook.
- International bond issuance and external commercial debt disbursements: World Bank’s International Debt Statistics (IDS) database (coverage includes public and publicly guaranteed debt; country reporting may differ).
- Main empirical approach: entropy balancing (Hainmueller 2012) to compare market-access and non-market-access countries with similar observable characteristics.
- Robustness checks: panel fixed-effect and GMM estimations.
- Additional notes: analysis extends prior work by covering 2004–18 and by including external commercial loans in the market-access concept.

### Key empirical findings and descriptive statistics
- Total private financing (disbursements from international bond market issuance and external commercial debt) rose from USD 72 billion in 2004 to a peak of US$ 296 billion in 2017.
- Increase driven mainly by a rise in international bonds.
- Number of countries issuing bonds:
  - averaged 18 per year in 2004–6 (first three years of sample period);
  - averaged 29 countries per year in the last three years of the sample period.
- Market access via debut bond market issuance:
  - 24 countries in the sample gained market access via a debut bond market issuance over 2004–18;
  - 21 of those made subsequent issues in the years following their debut.
  - Debut issuances averaged 2.7 percent of GDP;
  - Sample average of all issues averaged 1.8 percent of GDP.
- Tax revenue trends:
  - Median tax revenue rose from 13.7 percent of GDP in 2013 to 15.3 percent of GDP in 2018.
  - Countries issuing bonds regularly (at least two thirds of the sample years) have average tax ratios of 19.4 percent of GDP;
  - Occasional issuers average 15.6 percent of GDP;
  - Non-issuers average 13.4 percent of GDP.
- Correlations and dynamics:
  - Tax levels are positively correlated with the level of private financing (bonds and external commercial loans as a share of GDP) from the previous year.
  - Tax levels tend to increase following bond market issuances with a reported median cumulative change (figure referenced).

### Principal empirical conclusion
- No evidence that access to bond markets or external commercial loans undermines countries’ efforts to collect tax revenue.
- Access to markets is found to have a positive impact on domestic revenue mobilization in the analysis.

### Interpretation and caveats
- Access to private financing may disincentivize domestic revenue mobilization if perceived as an easier, politically less costly source of revenue; alternatively, access may incentivize fiscal discipline (Market Discipline Hypothesis) by raising sovereign debt risk premia in response to higher deficits or public debt.
- The paper’s results are conditioned on the empirical strategies used (entropy balancing, panel FE, GMM) and the sample/period chosen (2004–18).
- The analysis extends prior literature by including external commercial loans and focusing on the recent rapid increase in sovereign bond issuances by developing countries.

### Policy implications and recommendations
- Low-income and emerging market economies should:
  - continue enhancing tax administration and tax policy to increase domestic revenue mobilization;
  - continue strengthening debt management practices to manage debt risks associated with market access, particularly exchange rate and interest rate risks.
- The findings serve as a “wake-up call” given mounting concerns over the sustainability of public debt in many countries.

*Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020230-print-pdf.pdf (wpiea2020230-print-pdf - 2018. Issuances have increased not only for emerging).*

### 0.1 percentage points of GDP in the two years following. However, this is somewhat lower

### wpiea2020230-print-pdf - 0.1 percentage points of GDP in the two years following. However, this is somewhat lower

### Key findings
- Access to private external financing (international bond markets and external commercial loans) is not found to undermine domestic revenue mobilization (DRM).  
- Estimated average treatment effect of market access on DRM is positive and economically meaningful: about 1 percentage point of GDP on average.
- Entropy balancing estimates (Table 3) for the Market Access dummy:
  - Column [1]: 1.044 ** (standard error (0.444))
  - Column [2]: 0.943 *** (standard error (0.352))
  - Column [3]: 0.928 *** (standard error (0.353))
  - Column [4]: 1.114 *** (standard error (0.354))
- Panel regressions provide some support but generally yield smaller and often statistically insignificant positive effects compared with entropy balancing estimates.
- Heterogeneity results (Table 8):
  - Market Access, non-LICs: 1.603 *** (0.415)
  - Market Access, LICs: 0.182 (0.576) — not statistically different from zero
  - International bond markets: 3.890 *** (0.502)
  - External commercial debt: 0.650 (0.435) — not statistically significant
  - Regular bond issuers: 3.240 *** (0.584)
  - Occasional bond issuers: -0.650 (0.435)
  - Market Access, commodity exporters: 1.811 *** (0.531)
  - Market Access, non-commodity exporters: 0.805 ** (0.395)
  - Market Access in non-election periods: 1.075 *** (0.425)
  - Market Access in election periods: -0.229 (0.786)

### Estimation strategy and methodology
- Main identification challenge: reverse causality and endogeneity between market access and tax revenues.
- Primary identification approach: impact assessment via entropy balancing (Hainmueller, 2012) to estimate the average treatment effect on the treated (ATT, 흉).
- Treatment variable: market access dummy = 1 if a country issues international bonds or borrows external commercial debt in a given year; 0 otherwise.
- Outcome variable: tax revenue-to-GDP (DRM).
- Matching covariates (lagged by one year): GDP per capita, agriculture value-added, trade openness, institutional quality (ICRG Political Risk index), foreign aid (grants-to-GDP), and inflation.
- Entropy balancing implementation:
  - Step 1: reweight control (no-market-access) observations so pretreatment covariate means match treated group.
  - Step 2: weighted regression (weighted least squares) of DRM on Market Access dummy (with variants including covariates, year fixed effects, regional fixed effects).
- Also used panel fixed effects, difference-GMM and system-GMM for robustness.

### Descriptive sample moments (pre- and post-weighting)
- Table 2 (pre-weighting means, market access vs no market access):
  - Grants-to-GDP (t-1): Market Access 2.538; No Market Access 3.883; Difference 1.345 (t-test p-val 0.000)
  - GDP per capita, log (t-1): Market Access 8.041; No Market Access 7.071; Difference -0.969 (t-test p-val 0.000)
  - Trade openness, log (t-1): Market Access 4.288; No Market Access 4.262; Difference -0.026 (t-test p-val 0.222)
  - Agriculture-to-GDP (t-1): Market Access 22.434; No Market Access 22.343; Difference 9.909 (t-test p-val 0.000)
  - Political risk (t-1): Market Access 67.403; No Market Access 61.976; Difference -5.426 (t-test p-val 0.000)
  - Inflation (t-1): Market Access 6.368; No Market Access 7.190; Difference 0.822 (t-test p-val 0.060)
  - Observations: Market Access 832; No Market Access 681
- After entropy balancing (synthetic control), column [5] shows no significant differences remain between treated and synthetic control groups.

### Robustness checks and alternative specifications
- Alternative entropy balancing specifications adding covariates (Table 4) keep the estimated effect positive; some specifications yield larger coefficients (examples from Table 4):
  - Market Access dummy coefficients across columns include: 1.027 ***, 1.006 ***, 0.991 ***, 1.320 ***, 2.884 ***, 3.663 ***, 3.913 ***, 3.459 ***, 3.358 ***, 3.318 *** with associated standard errors shown in Table 4.
  - R2 values vary across specifications, e.g., 0.411, 0.413, 0.414, 0.458, 0.677, 0.667, 0.68, 0.685, 0.690, 0.691 and observations vary across columns (978, 970, 968, 899, 783, 523, 523, 523, 480, 480).
- Alternative treatment definitions (Table 5):
  - Treatment = 1 if market financing above a country-specific standard deviation: Market Access dummy 1.263 *** (0.370)
  - Treatment = 1 if market financing above 30th percentile of non-null observations: 0.969 *** (0.359)
  - Reweighting to match 1st and 2nd moments: 1.633 *** (0.355)
  - Reweighting to match 1st, 2nd, and 3rd moments: 1.829 *** (0.348)
  - Observations in Table 5: 1,007; R2 values: 0.334, 0.445, 0.383, 0.389
- Panel regressions (Table 6 and Table 7) provide mixed support:
  - Fixed effects regression examples: Market access dummy (t-1) estimates not consistently significant; private finance/GDP (t-1) shows positive coefficients in some specifications (e.g., 0.201** with standard error (0.03) and 0.206* with (0.05) in some columns).
  - System and difference GMM results (Table 7) present AR(1), AR(2) p-values and overidentification p-values across specifications; many coefficients for Market access dummy remain not statistically significant in these GMM specifications.

### Heterogeneity analysis
- By country income classification:
  - Positive effect concentrated in non-LICs (Market Access, non-LICs 1.603 ***) while LICs show no significant effect (0.182).
- By type of market access:
  - International bond market access yields a strong positive effect (3.890 ***) while external commercial debt alone shows no significant effect (0.650).
- By issuance frequency:
  - Regular bond issuers (issued bonds at least two-thirds of sample years): 3.240 ***.
  - Occasional bond issuers: -0.650 (not significant).
- By commodity dependence:
  - Commodity exporters: 1.811 ***; Non-commodity exporters: 0.805 **.
- By political cycle:
  - Positive effect concentrated in non-election periods: 1.075 ***; no significant effect during election periods: -0.229.

### Interpretation and plausible channels
- Market discipline channel: international bond issuances—and regular, transparent issuances—appear to strengthen incentives for DRM, consistent with bond vigilante effects and demands for creditworthiness.
- Institutional and macroeconomic channels: higher institutional quality (ICRG political risk index), macroeconomic stability (lower inflation), and stronger fundamentals are associated with higher tax revenue performance and with market access.
- Cross-sectional differences imply that market access effects depend on country characteristics (income level, bond market participation, issuance regularity, commodity dependence, election timing).

### Policy implications and recommendations
- Access to private financing should not be assumed to weaken tax collection incentives; in many cases, market access is associated with higher DRM.
- Countries should:
  - Continue enhancing tax administration and policy to increase DRM.
  - Strengthen debt management practices to manage debt risks, including exchange rate and interest rate risks associated with market access.
  - Recognize that market access may impose discipline that supports DRM, particularly for regular international bond issuers and non-LICs.
- Caution: given mounting concerns over public debt sustainability, market access is not an unambiguous substitute for robust domestic revenue and prudent debt management.

### Areas for further research (as noted in the source)
- Analyze impact of market access on subcomponents of tax revenue (e.g., income tax, trade taxes) to clarify channels.
- Study timing: how long after market access do revenue mobilization effects materialize.
- Consider business cycle effects (e.g., cyclically adjusted revenues) to further test the relationship between private external financing and tax revenues.
- Note on Sub-Saharan Africa (Table 9): some significance found when controlling for external debt levels, but results influenced by small sample size and outliers; significance may be sensitive to removal of specific countries.

*Source: IMF working paper (text extracted from wpiea2020230-print-pdf).*

### References

### wpiea2020230-print-pdf - References

### Key referenced topics and studies
- Political cycles and macroeconomy: Alesina, A., Roubini, N., Cohen, G.D., 1997. Political cycles and the macroeconomy, MIT Press.
- Market discipline and sovereign borrowing: Bayoumi, T., Goldstein, M., & Woglom, G. (1995); Lane, T. (1993); Bulut, L. (2012).
- Bond markets initiation and tax revenue mobilization: Balima, W.H., Combes, J-L., Minea, A., 2016. Bond markets initiation and tax revenue mobilization in developing countries, Southern Economic Journal, 83, 2, 550–572.
- Sovereign debt crises and credit default swaps: Balima, W.H., Combes, J-L., Minea, A., 2018. The ‘dark side’ of credit default swaps initiation: A close look at sovereign debt crises. Macroeconomic Dynamics, forthcoming.
- Bailouts and sovereign debt crises probability: Balima, W.H., Sy, A. 2019. The impact of bailouts on the probability of sovereign debt crises: Evidence from IMF-supported programs, IMF wp, 19/ 2.
- Tax revenue, trade liberalization and tax capacity: Baunsgaard, T., Keen, M., 2010; Gaspar, V., Jaramillo, L., and Wingender, P., 2016. “Tax Capacity and Growth: Is there a Tipping Point?” IMF Working Paper, No. 16/234.
- Foreign aid and revenue effects: Benedek, D., Crivelli, E., Gupta, S., Muthoora, P., 2014; Benedek, D., Crivelli, E., Gupta, S., and Muthoora P., 2012; Gupta, A.S., Clements, B., Pivovarsky, A., Tiongson, E., 2003.
- Determinants of tax revenue efforts: Gupta, A.S. 2007. Determinants of tax revenue efforts in developing countries, IMF wp, 07/184, 1-39.
- Fiscal policy over election cycles and in Sub-Saharan Africa: Ebeke, C., Olcer, D., 2013. Fiscal Policy over the Election Cycle in Low-Income Countries, IMF wp, 13/153, 1–18; Lledo, V. and Poplawski-Ribeiro M., 2013. Fiscal Policy Implementation in Sub-Saharan Africa. World Development Vol.46, pp 79-91, 2013.
- Methods for causal inference: Hainmueller, J., 2012. Entropy balancing for causal effects: A multivariate reweighting method to produce balanced samples in observational studies. Political Analysis 20, 25–46.
- Donor behavior and sanctions: Fuchs, A., Dreher, A., Nunnenkamp, P., 2014; Neuenkirch, M., Neumeier, F., 2016.
- Sovereign borrowing access determinants: Gelos, R.G., Sahay, R., Sandleris, G., 2011.
- Country spreads and emerging markets interactions: Uribe, M., Yue, V.Z., 2006.

### Annex — Table A1: Countries by Issuer Group
- Non-issuer:
  - Algeria, Bangladesh, Botswana, Burkina Faso, The Gambia, Guinea, Guinea-Bissau, Guyana, Haiti, Iran, Liberia, Madagascar, Malawi, Mali, Moldova, Myanmar, Nicaragua, Niger, Sierra Leone, Sudan, Tanzania, Togo, Uganda, Yemen, Zimbabwe
- Occasional Issuer (issued 1-9 years in sample period):
  - Albania*, Angola*, Armenia*, Azerbaijan*, Belarus*, Bolivia*, Bulgaria, Cameroon*, Costa Rica, Cote d'Ivoire*, Dominican Republic, Ecuador*, Egypt, El Salvador, Ethiopia*, Gabon*, Ghana*, Guatemala, Honduras*, Jordan*, Kazakhstan, Kenya*, Morocco, Mozambique*, Nigeria*, Pakistan*, Papua New Guinea*, Paraguay*, Senegal*, Vietnam*, Zambia*
  - Note: *Indicates country had debut bond market issuance during 2004–18
- Regular Issuer (issued >9 years in sample period):
  - Brazil, China, Colombia, India, Indonesia, Jamaica, Mexico, Peru, Philippines, Romania, Russia, South Africa, Thailand, Tunisia*, Turkey, Ukraine

### Annex — Table A2: Balima et al. (2016) benchmark result of bond markets participation on domestic tax revenue
- Dependent variable: Tax revenue ratio
- Estimation approaches and baseline ATT estimates (bootstrapped standard errors in parenthesis; based on 500 replications):
  - N Nearest Neighbor Matching:
    - N=1: ATT = 1.225** (0.614)
    - N=2: ATT = 1.473*** (0.584)
    - N=3: ATT = 1.565*** (0.544)
  - Radius Matching:
    - r=0.005: ATT = 1.565*** (0.514)
    - r=0.01: ATT = 1.669*** (0.496)
    - r=0.05: ATT = 1.758*** (0.443)
  - Kernel Matching: ATT = 1.769*** (0.476)
  - Local linear Matching: ATT = 1.727*** (0.467)
  - Stratification Matching: ATT = 1.555*** (0.479)
- Treated/Untreated/Total observations reported by specification:
  - N=1, N=2, N=3: 511/1300/1811
  - r=0.005: 465/1300/1765
  - r=0.01: 505/1300/1805
  - r=0.05: 511/1300/1811
  - Kernel/Local linear/Stratification: 511/1300/1811
  - Alternate stratification listing: 587/1160/1747
- Significance notation:
  - ***. **. * respectively represent significance thresholds of 1%. 5% et 10%

*Content from wpiea2020230-print-pdf - References*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020230-print-pdf.pdf_