## _wp05223

## Source details

**Canonical URL:** [_wp05223](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05223.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05223.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05223.pdf.json)

---

### Introduction and research question
- Research objective: investigate the effect of indebtedness on growth, resource flows, and investment patterns in a panel of about 80 developing countries, including more than 30 HIPCs, focusing on differences between HIPCs and non-HIPCs and on how country characteristics (indebtedness, policies, institutions) alter the debt–growth relationship.
- Context facts:
  - Sample period and scope: panel of 79 developing countries over 1970–2002; three-year averages yield 11 time periods (1970–73 to 2000–02).
  - HIPC implementation: through July 2005, debt reduction packages approved for 28 of the 38 eligible countries, 23 in sub-Saharan Africa; packages have provided $38.2 billion in debt relief over time, in present value terms.
- Conceptual framing:
  - Distinguish “marginal debt overhang” (negative marginal effect of debt on growth) and “average debt overhang” (country could enjoy higher growth if indebtedness were decreased sufficiently).

### Data, sample, and measurement
- Sample: unbalanced panel of 79 developing countries, 1970–2002; three-year averaging.
- Primary debt measures:
  - Nominal debt (Debt) from Global Development Finance (GDF).
  - Proxy for net present value of total external debt (NPV): NPV of public debt series as in Kraay and Nehru (2004) updated through 2002; private NPV proxied by private nominal debt from GDF.
- Key control variables: initial GDP per capita, population growth, terms of trade growth, secondary school enrollment, investment, central government balance, inflation, net private capital flows, openness, aid (grants and concessional loans), debt service.
- Cross-sectional variables: rule of law, legal origin, ethnic fractionalization, distance from equator, CPIA (1–5), settler mortality.
- Endogeneity and smoothing:
  - Use SGMM (system GMM) with lags as instruments.
  - Use Hodrick-Prescott filtered GDP denominators to address reverse causality concerns.

### Methodology
- Regression specifications:
  - Quadratic specification of debt (levels and squared) interacted with HIPC or high-debt dummy.
  - Spline specification with two breaks (low, medium, high debt ranges); breakpoints chosen by best fit over candidate ranges between about 10 and 80 percent of GDP.
  - Threshold estimation (Hansen) applied sequentially to detect endogenous thresholds and confidence intervals (used as robustness test; threshold estimation does not correct for endogeneity as SGMM does).
- Controls consistently include net transfers (aid) and debt service to isolate marginal debt overhang effects holding net flows constant.
- Diagnostics reported across tables: Observations, Number of Countries, AR(1) P-values, AR(2) P-values, Hansen/Sargan P-values, R-squared.

### Key empirical findings — debt and growth
- General pattern:
  - For non-HIPCs (and low-debt countries): evidence of a concave (quadratic) relationship — debt overhang present beyond a threshold.
  - For HIPCs (and high-debt countries): no significant marginal relationship between debt and growth over the sample period.
- Thresholds and nonlinearities (NPV of debt unless otherwise noted):
  - Spline (whole sample): debt overhang threshold between 10 and 35 percent of GDP; debt irrelevance threshold around 50 percent of GDP.
  - Countries with good policies and institutions:
    - Debt overhang threshold between 23 and 35 percent of GDP.
    - Debt irrelevance threshold around 80 percent of GDP (top of the searched range).
  - Countries with bad policies and institutions:
    - Debt overhang threshold around (10) percent of GDP (possibly below 10).
    - Debt irrelevance threshold between 15 and 53 percent of GDP.
  - Quadratic/Text Table A ranges:
    - Non-HIPC NPV overhang thresholds: Full sample 10–12; Split samples 15–16.
    - Non-HIPC nominal debt overhang thresholds: Full sample 20–21; Split samples 26–30.
  - Threshold estimation (Hansen, sequential) (NPV-based):
    - Full sample: first threshold 17 percent (confidence interval [9,46]); second threshold 37 percent (confidence interval [34,79]).
    - Good-CPIA countries: first threshold 17.6 percent (CI [14,30]); second threshold 71.7 percent (CI [69,77]).
    - Bad-CPIA countries: first threshold 17.3 percent (CI [5,37]); second threshold 35.1 percent (CI [34,39]).
- Magnitudes:
  - Spline estimates: combined overhang coefficient (β2+β3) ranges between -1 and -2.7 (OLS and SGMM), implying that reducing debt by a third (within the overhang–irrelevance interval) could enhance per capita growth by about one-third to a full percentage point.
- Robust result summary:
  - Debt overhang (negative marginal effect) present for non-HIPCs and low-debt countries.
  - Marginal effect of debt is not significant for HIPCs and high-debt countries.

### Determinants of resource transfers and investment — channels explaining heterogeneity
- Aggregate net transfers (ANT/GDP):
  - In non-HIPCs: net transfers decline with higher debt (negative and significant relationship).
  - In HIPCs: no significant negative relationship between debt and net transfers.
  - Appendix Table 2 sample means (selected):
    - Aggregate net transfer to GDP (ALL): 5.5 (1970–1979), 6.0 (1980–1989), 5.7 (1990–1995), 3.7 (1996–2002), 5.4 (1970–2002).
    - Aggregate net transfer to GDP (HIPC): 7.2, 10.2, 12.5, 8.4, 9.4.
    - Aggregate net transfer to GDP (Non-HIPC): 4.3, 3.3, 1.5, 1.5, 3.0.
- Concessionality of new lending (Average Grant Element):
  - In non-HIPCs: average grant element of new borrowing falls with higher debt (negative relationship).
  - In HIPCs: average grant element rises with higher debt (positive relationship), indicating more concessional lending as debt rises.
- Investment (INV/GDP):
  - Pooling all countries: negative and significant relationship between investment and debt.
  - Allowing different slopes for HIPCs vs non-HIPCs:
    - In HIPCs, investment is not affected by indebtedness levels.
    - In non-HIPCs, investment declines with higher debt.
- Interpretation:
  - Donors/creditors appear to offset crowding-out in HIPCs by maintaining or increasing net official transfers and concessionality, explaining why high debt does not show a marginal negative effect on growth or investment in HIPCs.
  - In non-HIPCs, private capital and terms respond to indebtedness, contributing to debt overhang via reduced private flows.

### Robustness, diagnostics, and subgroup findings
- Methods compared: OLS and sys-GMM (SGMM); lagged dependent variables included in dynamics.
- Threshold and spline analyses performed across subgroups:
  - High K flows vs Low K flows, Low settler mortality vs High settler mortality, Good CPIA vs Bad CPIA.
- Representative threshold cutoffs (from tables):
  - Log NPV/GDP <2.84 (NPV/GDP <17); 2.84>Log NPV/GDP >3.62 (17>NPV/GDP >37.4); Log NPV/GDP >3.62 (NPV/GDP >37.4).
  - Good-CPIA example: Log NPV/GDP cutoffs include <2.87 (NPV/GDP <17.6) and >4.273 (NPV/GDP >71.7).

### Policy implications and conclusions
- Debt relief effectiveness conditional on scale and context:
  - Debt relief is growth-enhancing only if sufficiently large to move indebtedness below the relevant debt overhang threshold for the country.
  - Countries must present good economic and policy conditions to convert debt relief into higher investment and growth.
- For countries with poor policies and institutions:
  - Evidence does not clearly support that additional debt relief will raise growth; the debt–growth relationship may be flat or thresholds may be at low debt levels such that marginal relief may not help.
- On HIPC-targeted relief:
  - The absence of a marginal debt–growth relationship in HIPCs historically reflects high indebtedness combined with continued donor support; marginal debt cancellation may have limited impact unless it is large and accompanied by better policies/institutions and access to private capital.
- Operational recommendations:
  - Use country characteristics (quality of policies and institutions) to guide debt relief decisions rather than a one-size-fits-all approach.
  - Make new lending after relief contingent on improvements in policy environment.
  - Consider extending eligibility for relief to non-HIPC low-income countries with similar debt levels and policy quality.
- Cautions:
  - Threshold estimates have wide confidence intervals in some cases; use of point estimates as policy benchmarks should be cautious.
  - Results depend on NPV vs nominal debt measures and on controlling for net transfers and endogeneity; robustness checks (SGMM, HP-filtered denominators, threshold estimation) yield broadly consistent qualitative conclusions.

### Supporting observations and contextual notes
- Findings consistent with Spring 2005 IMF African Regional Economic Outlook: lower debt associated with the occurrence of sustained growth acceleration.
- Observed increase in litigation by commercial creditors in HIPCs suggests creditors recognize improved conditions and potential for repayment.
- Appendix Table 2 selected macro means (1970–2002):
  - Growth per capita (ALL): 1.9, 0.2, 0.6, 1.2, 1.0 (by period and overall).
  - NPV of debt to GDP (ALL): 12.6, 40.8, 59.3, 44.8, 35.8.
  - Investment to GDP (ALL): 20.2, 19.8, 19.9, 20.3, 20.0.
  - Official net transfers to GDP (ALL): 4.2, 6.6, 6.0, 2.8, 5.1.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05223.pdf*

### References..............................................................................................................

### References................................................................................................................................24

### Figures
- 1. HIPC: Total Debt to GDP and Net Official Transfers to GDP............................................27
- 2. BRADY: Total Debt to GDP and Aggregate Net Transfer to GDP ....................................27
- 3. Net Official Transfers to GDP .............................................................................................28
- 4a. HIPC: Debt Shares .............................................................................................................29
- 4b. BRADY: Debt Shares ........................................................................................................29
- 5. Debt and Growth..................................................................................................................30

### Text Tables
- A. Debt Overhang Thresholds (% GDP) .................................................................................13
- B. Debt Overhang and Irrelevance Thresholds........................................................................16

### Tables
- 1. Growth Regressions: HIPCs vs. Non-HIPCs.......................................................................31
- 2. Growth Regressions: HIPCs vs. Non HIPCs (sub samples) ................................................32
- 3. Growth Regressions: High Debt vs. Low Debt ...................................................................33
- 4. Growth Regressions: High Debt vs. Low Debt HIPCs (sub samples).................................34
- 5. Spline Growth Regression, NPV of Debt, OLS and sys-GMM ..........................................35
- 6. Spline Growth Regression, NPV of Debt, OLS and sys-GMM (continued) .......................36
- 7. Spline Growth Regression, Nominal Debt, OLS and sys-GMM.........................................37
- 8. Spline Growth Regression, Nominal Debt, OLS and sys-GMM (continued) .....................38
- 9. Threshold Estimation Regressions, NPV of Debt, OLS, Full Sample.................................39
- 10. Threshold Estimation Regressions, NPV of Debt, OLS, Good CPIA ...............................40
- 11. Threshold Estimation Regressions, NPV of Debt, OLS, Bad CPIA .................................41
- 12. Determinants of Resource Flows, NPV Debt ....................................................................42
- 13. Determinants of Resource Flows, Debt .............................................................................43
- 14. Determinants of Lending Terms ........................................................................................44
- 15. Determinants of Total Investment, NPV Debt...................................................................45
- 16. Determinants of Total Investment, Debt............................................................................46

### Appendix Tables
- 1. List and Sources of Variables ..............................................................................................47
- 2. Summary Statistics: Mean Values .......................................................................................49
- 3. Country Classification by HIPC Status and Debt Levels. ...................................................51
- 4. Country Classification by HIPC Status and CPIA Index.....................................................52

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05223.pdf*

### 5. Country Classification by HIPC Status and IDA Status in 2004.........................................53

### 5. Country Classification by HIPC Status and IDA Status in 2004.........................................53

### Introduction and research question
- Research objective: investigate the effect of indebtedness on growth, resource flows, and investment patterns in a panel of about 80 developing countries, including more than 30 HIPCs, focusing on differences between HIPCs and non-HIPCs and on how country characteristics (indebtedness, policies, institutions) alter the debt–growth relationship.
- Context: large share of HIPCs’ stock of external debt cancelled through the HIPC initiative; through July 2005, debt reduction packages had been approved for 28 of the 38 eligible countries, 23 of which in sub-Saharan Africa. These packages have provided $38.2 billion in debt relief over time, in present value terms.
- Conceptual framing: distinguish “marginal debt overhang” (negative marginal effect of debt on growth) and “average debt overhang” (country could enjoy higher growth if indebtedness were decreased sufficiently).

### Data, sample, and measurement
- Sample: unbalanced panel of 79 developing countries over the period 1970–2002; three-year averages yield 11 time periods (1970–73 to 2000–02).
- Primary debt measures:
  - Nominal debt (Debt) from Global Development Finance (GDF).
  - Proxy for net present value of total external debt (NPV), sum of public and private: NPV of public debt series as in Kraay and Nehru (2004) updated through 2002; private NPV proxied by private nominal debt from GDF.
- Key control variables: initial GDP per capita, population growth, terms of trade growth, secondary school enrollment, investment, central government balance, inflation, net private capital flows, openness, aid (grants and concessional loans), debt service.
- Cross-sectional variables: rule of law, legal origin, ethnic fractionalization, distance from equator, CPIA (1–5), settler mortality.
- Endogeneity and smoothing: use SGMM (system GMM) with lags as instruments; use Hodrick-Prescott filtered GDP denominators to address reverse causality concerns.

### Methodology
- Regression forms used:
  - Quadratic specification of debt (levels and squared) interacted with HIPC or high-debt dummy.
  - Spline specification with two breaks (low, medium, high debt ranges), choosing breakpoints by best fit over candidate ranges between about 10 and 80 percent of GDP.
  - Threshold estimation (Hansen) applied sequentially to detect endogenous thresholds and confidence intervals (note: threshold estimation does not correct for endogeneity as SGMM does; used as robustness test).
- Controls consistently include net transfers (aid) and debt service to isolate marginal debt overhang effects holding net flows constant.

### Key empirical findings — debt and growth
- General pattern:
  - For non-HIPCs (and low-debt countries): evidence of a concave (quadratic) relationship — debt overhang present beyond a threshold.
  - For HIPCs (and high-debt countries): no significant marginal relationship between debt and growth over the sample period.
- Thresholds and nonlinearities (NPV of debt unless otherwise noted):
  - Spline (whole sample): debt overhang threshold between 10 and 35 percent of GDP; debt irrelevance threshold around 50 percent of GDP.
  - Countries with good policies and institutions:
    - Debt overhang threshold between 23 and 35 percent of GDP.
    - Debt irrelevance threshold around 80 percent of GDP (top of the searched range).
  - Countries with bad policies and institutions:
    - Debt overhang threshold around (10) percent of GDP (possibly below 10).
    - Debt irrelevance threshold between 15 and 53 percent of GDP.
  - Text Table A (quadratic results, ranges):
    - Non-HIPC NPV overhang thresholds: Full sample 10–12; Split samples 15–16.
    - Non-HIPC nominal debt overhang thresholds: Full sample 20–21; Split samples 26–30.
    - HIPC thresholds often blank or lower and not statistically robust.
  - Threshold estimation (Hansen, sequential):
    - For full sample using NPV: first threshold 17 percent (confidence interval [9,46]); second threshold 37 percent (confidence interval [34,79]).
    - For good-CPIA countries: first threshold 17.6 percent (CI [14,30]); second threshold 71.7 percent (CI [69,77]).
    - For bad-CPIA countries: first threshold 17.3 percent (CI [5,37]); second threshold 35.1 percent (CI [34,39]).
- Magnitudes:
  - Spline estimates: combined overhang coefficient (β2+β3) ranges between -1 and -2.7 (OLS and SGMM), implying that reducing debt by a third (within the overhang–irrelevance interval) could enhance per capita growth by about one-third to a full percentage point.
- Robust result: debt overhang (negative marginal effect) is present for non-HIPCs and low-debt countries; marginal effect of debt is not significant for HIPCs and high-debt countries.

### Determinants of resource transfers and investment — channels explaining heterogeneity
- Net transfers (aggregate) behavior:
  - In non-HIPCs: net transfers decline with higher debt (negative and significant relationship between debt and net transfers).
  - In HIPCs: no significant negative relationship between debt and net transfers; donors/creditors continue to provide resources as indebtedness rises.
- Concessionality of new lending:
  - In non-HIPCs: average grant element of new borrowing falls with higher debt (negative relationship).
  - In HIPCs: average grant element rises with higher debt (positive relationship), indicating more concessional lending as debt rises.
- Investment:
  - When pooling all countries: negative and significant relationship between investment and debt.
  - Allowing different slopes for HIPCs vs non-HIPCs: in HIPCs, investment is not affected by indebtedness levels; in non-HIPCs, investment declines with higher debt.
- Interpretation: donors appear to offset crowding-out effects in HIPCs by maintaining or increasing net official transfers and concessionality, which can explain why high debt does not show a marginal negative effect on growth or investment in HIPCs. In non-HIPCs, private capital and terms respond to indebtedness, contributing to debt overhang through reduced private flows.

### Policy implications and conclusions
- Debt relief is growth-enhancing only under certain conditions:
  - Debt relief must be sufficiently large to move indebtedness below the debt overhang threshold relevant for the country’s characteristics.
  - Countries must present good economic and policy conditions to convert debt relief into higher investment and growth.
- For countries with poor policies and institutions:
  - Evidence does not offer clear support that additional debt relief will raise growth; debt-growth relationship may be flat or have thresholds at low debt levels such that marginal relief may not help.
- On HIPC-targeted relief:
  - The absence of a marginal debt–growth relationship in HIPCs over the sample reflects historically high indebtedness combined with continued donor support; thus marginal debt cancellation may have limited impact unless it is large and accompanied by better policies/institutions and access to private capital.
- Cautions:
  - Threshold estimates have wide confidence intervals in some cases, implying uncertainty about exact threshold values; use of point estimates as policy benchmarks should be cautious.
  - Results hinge on distinguishing NPV vs nominal debt measures and on controlling for net transfers and endogeneity; robustness checks (SGMM, HP-filtered denominators, threshold estimation) yield broadly consistent qualitative conclusions.

*Source: 5. Country Classification by HIPC Status and IDA Status in 2004.*

### Appendix Table 5) to draw inference on the independent effect of income.

### _wp05223 - Appendix Table 5) to draw inference on the independent effect of income.

### Evidence on the effect of debt on growth
- There is some evidence of a positive effect at low levels; this effect, however, is not robust.
- The fact that high levels of debt do not matter for growth does not imply that intermediate levels of debt do not matter either: debt that is not too small (not below the debt overhang threshold) and not too big (not above the debt irrelevance threshold), does negatively affect growth.
- The absence of marginal debt overhang at high levels of debt does not imply the absence of average debt overhang: even if small reductions in debt may have no impact on growth, large reductions might still significantly improve growth performance.

### Thresholds and heterogeneity by country characteristics
- For countries with better institutions, better policies, and easier access to private capital:
  - Debt overhang thresholds appear when debt exceeds 15-30 percent of GDP.
  - Marginal debt overhang persists until debt reaches 70-80 percent of GDP (debt irrelevance thresholds).
  - For debt above these very high levels, the marginal effect of debt on growth is zero (no marginal debt overhang), even though the overall effect of debt remains negative (average debt overhang).
- For countries with worse conditions:
  - Debt overhang thresholds mainly around 0 and 20 percent of GDP.
  - Debt irrelevance thresholds between 15 and 53 percent of GDP.
  - The relationship between debt and growth for this group is much less robust, casting doubts on whether debt matters at all.
- Note on uncertainty: The thresholds identified should be viewed cautiously; they are affected by considerable estimation uncertainty and might reflect factors not fully captured in the analysis.

### Implications for debt relief and subsequent lending
- One cannot dismiss debt relief as an effective way of providing resources to highly indebted countries solely because high debt levels did not matter historically.
- The fact that a debt irrelevance threshold was found at debt levels much lower than those the HIPCs used to have implies that past experience is not necessarily a reliable guide for current policy trade-offs.
- It is possible that the debt relief initiative has moved HIPCs within the marginal debt overhang region, where debt relief could be effective in fostering growth.
- Because the effectiveness of debt relief depends on countries’ characteristics, a one-size-fits-all debt relief policy might not be appropriate: the choice of providing debt relief should depend on the same factors (quality of policies and institutions).
- After debt relief is granted, new lending to these countries should be made contingent on improvements in the quality of policy environment.
- Eligibility for further debt relief should be extended to other non-HIPC low-income countries in a manner similar to HIPCs when those countries have similar levels of debt and quality of policies and institutions, since these are the key determinants of the effect of debt relief on growth.

### Supporting observations cited in the analysis
- The findings are consistent with the Spring 2005 IMF African Regional Economic Outlook, which found that lower debt was associated with the occurrence of sustained growth acceleration.
- An observed increase in litigation by commercial creditors in HIPCs suggests that creditors are recognizing that conditions have improved, and expect that there may now be some possibility of repayment.

*Source: _wp05223 - Appendix Table 5) to draw inference on the independent effect of income.*

### REFERENCES

### _wp05223 - REFERENCES

### Bibliographic references
- Comprehensive list of academic and policy sources cited, including but not limited to:
  - Abiad, A., and J. Ostry, 2005, “Primary Surpluses and Sustainable Debt Levels in Emerging Market Countries,” IMF Policy Discussion Paper 05/6.
  - Acemoglu, D., S. Johnson, and T. Robinson, 2001, “Colonial Origins of Comparative Development: An Empirical Investigation,” American Economic Review, No. 91, pp. 1369–1401.
  - Arellano, M., and O. Bover, 1995, “Another Look at the Instrumental Variable Estimation of Error-Components Models,” Journal of Econometrics, Vol. 68, pp. 29–51.
  - Blundell, R., and S. Bond, 1998, “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models,” Journal of Econometrics, Vol. 87, pp. 115–43.
  - Borenzstein, E., 1990, “Debt Overhang, Credit Rationing and Investment,” Journal of Development Economics, 32, pp. 315–35.
  - Clements, B.J., R. Bhattacharya, and T. Q. Nguyen, (2003), “External Debt, Public Investment, and Growth in Low-Income Countries,” IMF Working Paper 03/249.
  - Easterly, W. R., 2001, “Growth Implosions and Debt Explosions: Do Growth Slowdowns Cause Public Debt Crises?” Contributions to Macroeconomics, Vol. 1, No. 1, Article 1, pp. 1–24.
  - Pattillo, C., H. Poirson, and L.A. Ricci, 2002, “External Debt and Growth?” IMF Working Paper 02/69.
  - Reinhart, C., K. Rogoff, and M. Savastano, 2003, “Debt Intolerance,” Brookings Papers on Economic Activity, Vol. 1, pp. 1–74.
  - Sachs, J. D., 1988, “The Debt Overhang of Developing Countries,” in Debt Stabilization and Development: Essays in Memory of Carlos Diaz-Alejandro.
  - Warner, A.M., 1992, “Did the Debt Crisis Cause the Investment Crisis?” Quarterly Journal of Economics, Vol. 107, No. 4.
- The references list includes working papers, journal articles, IMF and World Bank publications, and conference papers addressing debt overhang, growth, aid, lending terms, and methodology for dynamic panel estimation.

### Figures (titles)
- Figure 1. HIPC: Total Debt to GDP and Net Official Transfers to GDP
- Figure 2. BRADY: Total Debt to GDP and Aggregate Net Transfers to GDP
- Figure 3. Net Official Transfers to GDP
- Figure 4a. HIPC: Debt Shares
- Figure 4b. BRADY: Debt Shares
- Figure 5. Debt and Growth
  - Note: “Source: Tables 5 and 7; averages across methods and debt measures. Data are normalized to have the same growth at the debt overhang threshold. For a definition of ‘good’ and ‘bad’ conditions, see the text.”

### Tables (titles and selected structural notes)
- Table 1. Growth Regressions: HIPCs vs. Non-HIPCs
  - Dependent Variable: Per Capita GDP Growth (PCGDP Gr)
  - Methods: OLS and SGMM; reported coefficients, robust standard errors in parentheses; significance indicators: * 10 percent; ** 5 percent; *** 1 percent; All regressions include time dummies.
  - Observations: 703 (OLS) / 734 (SGMM); Number of Countries: 79.
- Table 2. Growth Regressions: HIPCs vs. Non HIPCs (subsamples)
  - Dependent Variable: Per Capita GDP Growth (PCGDP Gr)
  - Observations vary (e.g., 424, 279, 446, 288); Number of Countries reported for subsamples.
- Table 3. Growth Regressions: High Debt vs. Low Debt
  - Observations: 703 (OLS) / 734 (SGMM); Number of Countries: 79.
- Table 4. Growth Regressions: High Debt vs. Low Debt (subsamples)
  - Subsample columns: Low Debt / High Debt for Nominal Debt and NPV Debt; Observations and Number of Countries reported.
- Table 5. Spline Growth Regression, NPV of Debt, OLS and sys-GMM
  - Panels: ALL, Good Policies, Bad Policies; includes coefficients (a) Log NPV/GDP, (b) Log NPV/GDP*D1, (c) Log NPV/GDP*D2 and tests for (a)+(b)=0 and (a)+(b)+(c)=0.
  - Debt overhang threshold entries include numeric values such as 35, 23, 23, (10), (10).
  - Observations: 703 (OLS) / 734 (sys-GMM) for ALL; other panels report 346, 360, 357, 374.
- Tables 6–9. Spline and Threshold regressions
  - Provide subgroup analyses (High K flows, Low K flows, Low settl. Mort, High settl. Mort).
  - Table 9 reports threshold ranges for Log NPV/GDP with bracketed NPV/GDP cutoffs: Log NPV/GDP <2.84 (NPV/GDP <17), 2.84>Log NPV/GDP >3.62 (17>NPV/GDP >37.4), Log NPV/GDP >3.62 (NPV/GDP >37.4).
- Tables 10–11. Threshold Estimation Regressions by CPIA (Good CPIA, Bad CPIA)
  - Provide separate threshold regressions and R-squared values for subgroups; examples of thresholds include Log NPV/GDP <2.87 (NPV/GDP<17.6) and Log NPV/GDP>4.273 (NPV/GDP>71.7).
- Tables 12–13. Determinants of Resource Flows (NPV Debt and Debt)
  - Dependent Variable: Aggregate Net Transfers (ANT/GDP).
  - Models: OLS and SGMM. Report coefficients for ANT/GDP-1, Log InGDP, ToTGr-1, LogPopGr-1, (a) Log NPV/GDP-1 or (a) Log Debt/GDP-1, and interaction H* Log NPV/GDP or H* Log Debt/GDP where H denotes HIPC.
  - Observations reported: 655 in SGMM columns; Number of Countries: 79.
- Table 14. Determinants of Lending Terms
  - Dependent Variable: Average Grant Element (GREL/GDP).
  - Coefficients include (a) Log NPV/GDP-1 and (b) H* Log NPV/GDP-1; HIPC (H) coefficients and significance reported.
  - Observations: 628 (OLS columns), 655 (SGMM columns).
- Tables 15–16. Determinants of Total Investment (INV/GDP) for NPV Debt and Debt
  - Dependent Variable: Total Investment (INV/GDP).
  - Key regressors: INV/GDP-1, PCGDP Gr-1, ToTGr-1, Open, (a) Log NPV/GDP-1 or (a) Log Debt/GDP-1, (b) H* Log NPV/GDP-1 or (b) H* Log Debt/GDP-1, HIPC (H).
  - Observations: 628 (OLS) / 655 (SGMM) in various specifications.
  - Sargan/Hansen p-values and AR(2) P-Values reported in tables.

### Appendix: Data, variables, summary statistics, and country classifications
- Table 1 (Appendix). List and Sources of Variables — variable definitions and data sources include:
  - Per capita real GDP: GDP per capita in constant 2000 US dollars. Source: WDI.
  - Per capita real GDP growth (PCGDP Gr): Log difference of per capita real GDP. Source: WDI.
  - Population growth (L PopGr): Annual population growth (in %). Source: WDI.
  - Secondary school enrollment (L SecEnr): Secondary school enrollment (in % gross). Source: WDI.
  - Terms of trade growth (ToT Gr): Log difference of terms of trade. Source: WEO.
  - Investment to GDP (LINV/GDP): The ratio of gross fixed capital formation to GDP. Source: WEO.
  - Debt stock to GDP (LDebt/GDP): Ratio of total debt stocks to GDP. Source: GDF (debt) and WDI (GDP).
  - NPV debt to GDP (LNPV/GDP): Ratio of net present value of debt to GDP. Source: World Bank data (NPV debt) and WDI (GDP).
  - Debt service to exports (DebtSer/EXP): Ratio of total debt service paid to exports. Source: GDF (debt) and WEO (exports).
  - Aid to GDP (Aid/GDP): Sum of grants (excluding technical cooperation), multilateral and bilateral concessional disbursements and IMF purchases divided by GDP. Source: GDF and WDI.
  - Aggregate net transfers to GDP (ANT/GDP): Ratio of aggregate net transfers to GDP. Source: GDF and WDI.
  - CPIA index: Country policy and institutional assessment index (ranges from 0 to 1). Source: World Bank.
  - Legal origin, Ethnic fractionalization, Distance to Equator: sources referenced (La Porta et al; Easterly and Levine; Rodrik et al); governance dataset referenced (Kaufmann et al).
- Appendix Table 2. Summary Statistics: Mean Values (selected entries preserved exactly)
  - Aggregate net transfer to GDP (ALL): 5.5 (1970–1979), 6.0 (1980–1989), 5.7 (1990–1995), 3.7 (1996–2002), 5.4 (1970–2002).
  - Aggregate net transfer to GDP (HIPC): 7.2, 10.2, 12.5, 8.4, 9.4 (same period breakdown).
  - Aggregate net transfer to GDP (Non-HIPC): 4.3, 3.3, 1.5, 1.5, 3.0.
  - Central government balance to GDP (ALL): -5.2, -5.7, -3.6, -3.2, -4.8.
  - CPIA (ALL): 2.6, 2.9, 3.0, 3.3, 2.9.
  - Grants to GDP (ALL): 3.6, 5.5, 7.0, 4.1, 4.9.
  - Growth per capita (ALL): 1.9, 0.2, 0.6, 1.2, 1.0.
  - Inflation (ALL): 13.6, 22.6, 25.3, 9.4, 17.9.
  - Investment to GDP (ALL): 20.2, 19.8, 19.9, 20.3, 20.0.
  - Net Private Capital Flows to GDP (ALL): 2.1, 2.0, 0.6, 1.7, 1.7.
  - NPV of debt to GDP (ALL): 12.6, 40.8, 59.3, 44.8, 35.8.
  - Official net transfers to GDP (ALL): 4.2, 6.6, 6.0, 2.8, 5.1.
  - Openness to GDP (ALL): 57.3, 56.7, 62.5, 69.1, 59.9.
  - Additional series: population growth, private investment to GDP, private net transfers to GDP, public investment to GDP, secondary school enrolment, terms of trade growth, total NPV of debt to GDP, total debt service to GDP, total debt to GDP — all reported in Appendix Table 2 with period-specific means.
- Appendix Tables 3–5. Country classifications
  - Table 3: Country Classification by HIPC Status and Debt Levels (lists specific countries by HIPC/non-HIPC and whether debt is above or below median).
  - Table 4: Country Classification by HIPC Status and CPIA Index (Good CPIA vs. Bad CPIA lists).
  - Table 5: Country Classification by HIPC Status and IDA Status in 2004 (IDA vs. Non-IDA lists).

### Key empirical elements and diagnostics reported across tables
- Regression methodologies: OLS and system-GMM (SGMM / sys-GMM) dynamic panel estimations.
- Diagnostic statistics regularly reported: Observations, Number of Countries, AR(1) test, AR(2) test P-values, Hansen/Sargan P-values, R-squared.
- Significance notation: * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent.
- Threshold findings: multiple tables report estimated debt overhang thresholds and “debt irrelevance” thresholds with exact numeric values presented in tables (examples include thresholds at 23, 35, 53, 80, and bracketed values noted as at the limit of the range).
- Interaction effects: regressions include interactions between HIPC indicator (H) or High Debt indicator (HD) and debt measures (Log Debt/GDP or Log NPV/GDP) to capture heterogeneous effects.
- Outcome linkages reported: debt (nominal and NPV) and resource flows (ANT/GDP), lending terms (GREL/GDP), and investment (INV/GDP) are analyzed with dynamic controls (lagged dependent variables, growth, openness, inflation, budget balance, aid, debt service ratios).
- Robust standard errors are consistently reported in parentheses for all coefficient estimates.

*Source: _wp05223 - REFERENCES (IMF PDF content unit).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05223.pdf_