Preserving Debt Sustainability in Low-Income Countries in the Wake of the Global Crisis

We cannot allow the return of economic stability to signify a return to "business as usual" for the IMF. The crisis exposed huge cracks in the international financial architecture of which the Fund is a key part. We have an historic responsibility to fix them. I urge all of us to recommit to seeing our collective goals to the finish line before reform fatigue sets in.


We cannot allow the return of economic stability to signify a return to "business as usual" for the IMF. The crisis exposed huge cracks in the international financial architecture of which the Fund is a key part. We have an historic responsibility to fix them. I urge all of us to recommit to seeing our collective goals to the finish line before reform fatigue sets in.

Executive Summary

The global financial crisis has had a significant impact on low-income countries (LICs)’ debt vulnerabilities. Recent debt sustainability analyses (DSAs) indicate that external and fiscal financing requirements have increased. In addition, standard measures of a country’s capacity to repay debt?GDP, exports, and fiscal revenue?are expected to be permanently lower. On average, debt ratios are therefore expected to deteriorate in the near term, particularly for public debt.

The global crisis, however, is not expected to result in systemic debt difficulties across LICs. Debt ratios are expected to return to a downward trend by 2011–12, and risk rating downgrades have been rare in post-crisis DSAs. Critical assumptions to achieve this outcome are that: (i) the crisis has no permanent impact on long-term growth; (ii) the recovery will be relatively quick, consistent with the nature of the shock for LICs (mostly external demand) and the expected recovery in industrialized countries; and (iii), as the recovery firms up, LICs restore the policy space that many of them have used to mitigate the impact of the crisis, and continue to have access to adequate financing.

The share of LICs that face higher debt vulnerabilities is significant but has not increased with the crisis. These countries are rated as being either at high risk of external debt distress or in debt distress in their most recent DSAs. A few countries with more favorable risk ratings also have high total public debt.

Sustained implementation of a combination of measures, involving debtors and creditors, should reduce debt vulnerabilities significantly in all these countries over the medium term:

  • In about half of these countries—including all those in debt distress—debt vulnerabilities are expected to be reduced substantially through HIPC/MDRI relief (or will require similar treatment).

  • In the other half:

    • ➢ Options to address debt vulnerabilities include fiscal consolidation and efforts to improve institutions and policies (particularly in the economic and debt management areas) on the debtor side, and more concessional financing terms on the creditor side;

    • ➢ Sustained implementation of a combination of these options should be sufficient to reduce debt vulnerabilities substantially over the medium term in all these countries.

    • ➢ Nevertheless, the need for debt relief in isolated cases at some point in the future cannot be excluded, given the hazard of large negative shocks. The effectiveness of traditional debt relief mechanisms would hinge on the participation of all creditors, as the share of Paris Club debt in total debt is relatively limited in these countries.

I. Introduction1

1. The global financial crisis has had a substantial impact on LICs.2 They have had to cope with a sharp slowdown in external demand, leading to a contraction in export growth, and much reduced non-debt-creating external financing, with a drop in foreign direct investment (FDI) inflows and remittances.3 In response, many LICs have implemented countercyclical macroeconomic policies, often with the financial support of the Fund, the Bank, and other international financial institutions.4 Automatic stabilizers have generally been allowed to operate and, in a number of cases, have been complemented with discretionary fiscal stimulus, mainly on the spending side. With scarce external financing (including aid), some countries have resorted to domestic debt financing to close widening fiscal financing gaps. Overall, economic growth has decreased sharply and external and fiscal borrowing requirements have increased substantially.

2. This paper analyses the extent to which debt vulnerabilities in LICs have risen as a result of the crisis based on a comparison of pre- and post-crisis debt sustainability analyses.5 With higher borrowing requirements and a less favorable evolution of GDP, exports, and revenues, the debt sustainability outlook has clearly deteriorated. A key question is how severe this deterioration is and, in particular, whether the global crisis could lead to a systemic debt crisis in LICs, as some have argued. This issue is taken up in Section II.

3. The paper also discusses options to address cases of high risk of debt distress or actual debt distress. These countries, which we refer to as those with “higher debt vulnerabilities,” face a wide diversity of situations. The options available to each of them to reduce these vulnerabilities vary accordingly. In Section III, the paper seeks to quantify whether the sustained use of these options can reasonably be expected to reduce debt vulnerabilities significantly in the medium term. Section III also considers whether traditional debt relief mechanisms would be effective, should they need to be activated for some LICs with higher debt vulnerabilities. Section IV offers conclusions.

4. The analysis in this paper focuses on public debt to the extent possible. As shown in earlier studies, available information indicates that public and publicly guaranteed external debt?the focus of DSAs conducted with the Debt Sustainability Framework (DSF)?still constitutes the bulk of LICs’ external debt.6 Private external debt is therefore not considered in this paper. However, a number of LICs have sizeable domestic public debt levels, and recourse to domestic debt has increased in the recent period.

II. How has the Debt Sustainability Outlook for LICs been Affected by the Crisis?

5. DSAs performed under the DSF allow for a comprehensive analysis of the debt situation of LICs (Box 1). DSAs are conducted annually for LICs, providing an opportunity to compare their debt situation before and after the crisis. WEO assumptions constitute a critical input for DSAs, as they provide the external environment a given LIC is expected to face during the projection period. These assumptions were revised drastically in early 2009, as it became clear that the financial crisis would have a major negative impact on the world economy. Subsequent revisions were, in comparison, much more limited. For the purpose of this analysis, it is therefore assumed that DSAs issued to the IMF Executive Board after May 1, 2009 include macroeconomic assumptions and frameworks that fully capture the expected impact of the crisis, as reflected in the April 2009 WEO (“post-crisis DSAs”) (Table 1). Conversely, DSAs issued prior to this date are assumed to reflect the pre-crisis situation, as they are based on earlier WEO assumptions (“pre-crisis DSAs”).7

Table 1.

Real Growth, WEO Spring 2008-Fall 2009

(in percent)

article image

Debt Sustainability Analysis1

DSAs conducted under the DSF focus on five debt burden indicators for external public debt: (i) PV of debt-to-GDP; (ii) PV of debt-to-exports; (iii) PV of debt-to-revenues; (iv) debt service-to-revenues; and (v) debt service-to-exports. Each of these indicators has an indicative threshold in the framework that depends on a country’s quality of policies and institutions as measured by the three-year average of the Country Policy and Institutional Assessment (CPIA) index, compiled annually by the World Bank. The specific thresholds are as follows:

Debt Burden Thresholds under the DSF (Applied to external public debt) 1/

article image
1/ CPIA measured as a three-year average.

A rating of the risk of external debt distress is derived by reviewing the evolution of debt burden indicators compared to their indicative policy-dependent debt-burden thresholds under a baseline scenario, alternative scenarios, and stress tests. There are four possible ratings:

  • Low risk. All debt indicators are well below relevant country-specific debt-burden thresholds. Stress testing and country-specific alternative scenarios do not result in indicators significantly breaching thresholds. In cases where only one indicator is above its threshold, judgment is needed to determine whether there is a debt sustainability problem or some other issue, for example, a data problem.

  • Moderate risk. While the baseline scenario does not indicate a breach of thresholds, alternative scenarios or stress tests result in a significant rise in debt-service indicators over the projection period (nearing thresholds) or a breach of debt or debt-service thresholds.

  • High risk. The baseline scenario indicates a protracted breach of debt or debt-service thresholds but the country does not currently face any payment difficulties. Alternative scenarios or stress tests also show protracted threshold breaches.

  • In debt distress. Current debt and debt-service ratios are in significant or sustained breach of thresholds. Actual or impending debt restructuring negotiations or the existence of arrears would generally suggest that a country is in debt distress.

The risk ratings are based on a probabilistic approach. The indicative policy-dependent thresholds correspond to probabilities of debt distress ranging from 18 to 22 percent for CPIA ratings of 3.25, 3.5 and 3.75 (the benchmarks set for strong, medium, and weak performers, respectively).2 Therefore, a high risk rating (unlike an “in debt distress” rating) should not be interpreted as synonymous of an unsustainable debt situation.

While the focus of DSAs is on public and publicly guaranteed external debt, they all also include an analysis of public debt sustainability. The DSF does not include, however, indicative thresholds for total public debt.

1 See “Staff Guidance Note on the Application of the Joint Fund-Bank Debt Sustainability Framework for Low-Income Countries” IMF policy paper (2010), ( See also Barkbu, B., C. H. Beddies, and M-H. Le Manchec (2008), “The Debt Sustainability Framework for Low-Income Countries” IMF Occasional Paper No. 266 (Washington: International Monetary Fund).2 See Box 9 in IMF (2009), “A Review of Some Aspects of the Low-Income Country Debt Sustainability Framework” (

A. The Pre-Financial Crisis Debt Situation8

6. Prior to the financial crisis, LICs had sizeable public debt levels, mostly owed to external creditors. Public debt amounted to 62 percent of GDP on average. Most of this debt was external, with domestic debt averaging about 14 percent of GDP. Debt levels were significantly lower when measured in present value (PV) terms, reflecting the concessionality of LICs’ external debt. In PV terms, total public debt and public external debt respectively averaged 49 and 34 percent of GDP. The concessionality of external debt was also reflected in moderate debt-service ratios.9

7. Beyond these averages, LICs faced very diverse debt sustainability situations. This diversity is well summarized by the distribution of “risk of debt distress” ratings included in DSAs. About a third of the countries in the sample had a low risk rating, about 30 percent had a moderate risk rating, and the rest were classified as either at high risk of debt distress or in debt distress. Countries with comparatively lower debt ratios tended to have received debt relief under the HIPC Initiative and the MDRI; be rich in hydrocarbon resources; and have relatively stronger policies and institutions, as measured by their CPIA score (Table 2).

Table 2.

Pre-Crisis Public Debt Indicators 2007–08 1/

article image
Note: Columns 1-4 are in percent of GDP; columns 5-8 are in percent of unit specified.

According to Guide in Resource Revenue Transparency, 2007 Revised Edition.

Using the IDA income cut-off of US$1,095 of July 1, 2008. Exclude Afghanistan and Myanmar.

Measured by the three-year backward looking CPIA used in the pre-crisis DSAs. No CPIAs are available for Liberia and Myanmar.

B. The Impact of the Crisis

8. The global financial crisis has lowered LICs’ GDP growth, exports, and fiscal revenue, and increased their external and fiscal borrowing requirements. This is reflected in the macroeconomic scenarios underpinning post-crisis DSAs, which are available for 36 LICs (see Annex I).10 Specifically, key changes between macroeconomic scenarios in pre- and post-crisis DSAs are as follows (see Panels 1–2):

  • On average, real GDP growth was revised downwards by 2.5 percentage points in 2009 and 1 percentage point in 2010, but is expected to return to pre-crisis levels in 2011-12. This relatively quick recovery by historical standards is consistent with the nature of the shock for LICs (mostly external demand) and the expected recovery in industrialized countries (see Annex 2), as well as the implementation of supportive policies in LICs.

  • The new growth and price outlook implies a permanent reduction in the level of nominal GDP, by an average of 10.5 percent in the long-term (relative to pre-crisis DSAs); exports and government revenues follow a similar pattern.

  • Current account deficits, excluding interest payments, are larger over a prolonged period. The initial deterioration averages about ¾ percentage point of GDP in 2009 and peaks at about 2 percentage points of GDP in 2012. Current account deficits only return to within one percentage point of GDP of their pre-crisis levels around 2021.The deterioration in current account deficits reflects permanently lower exports and net current transfers, the latter on account of remitttances (see Box 2).

  • FDI, a non-debt creating source of external financing, is substantially reduced in 2009 and 2010 (as a percentage of GDP) but recovers to pre-crisis levels by 2011.

  • Primary fiscal deficits are projected to be larger by an average of about 2 percentage points of GDP in 2009. This reflects both the operation of automatic stabilizers (mostly lower revenues) and, in a number of LICs, discretionary fiscal stimulus (mostly higher non-interest expenditures), both contributing to mitigating the impact of the crisis. The larger deficits are assumed to be financed both externally and domestically.11

  • Primary fiscal deficits gradually return to their pre-crisis levels by 2015. Thus, DSAs assume that the policy space used during the crisis is progressively restored as revenues recover and non-interest expenditures (as a share of GDP) are reduced to levels slightly below those projected in pre-crisis DSAs.

9. On average, post-crisis DSAs show a significant deterioration in debt ratios compared with the pre-crisis projections, particularly over the next five years or so (Panel 3).12 The PV of public debt-to-GDP ratio is expected to be higher by 5-7 percentage points in 2009 and 2010 than projected earlier. The debt service-to-revenue ratio is expected to be permanently higher by 2 percentage points over the projection period. Post-crisis DSAs show a significant increase in the level of debt and debt-service ratios between 2008 and 2009-10. The PV of public debt-to-GDP ratio is estimated to have increased on average by about 5 percent of GDP in 2009 over 2008. Ratios related to public external debt follow broadly similar patterns.

Panel 3.
Panel 3.

How has the crisis affected debt burden indicators?

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Source: Bank and Fund staff estimates.

10. The deterioration in the public debt to GDP ratio tapers off over the long term and this ratio is expected to return to a downward trend in post-crisis DSAs. This development is predicated on a return to less expansionary fiscal policies (reduction in primary deficits) as well as more favorable endogenous debt dynamics (a recovery in growth). By contrast, keeping primary fiscal balances at 2009 levels would lead to unsustainable debt levels over the medium- and long-term. Similarly, lower GDP growth than assumed in baseline scenarios would worsen debt ratios significantly.

Panel 1.
Panel 1.

How has the crisis affected LICs’ repayment capacity?

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Source: Bank and Fund staff estimates.
Panel 2.
Panel 2.

How has the crisis affected LICs borrowing requirements? (in percent of GDP)

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Prospects for Remittances

Remittances are likely to grow slowly in the short run. Although remittance flows tend to be countercyclical with respect to recipient country economic cycles (Chami et al., 2008), the effect of source country economic cycles clearly dominates in the context of the current global downturn. According to World Bank estimates (Ratha et al., 2009), remittances to developing countries may have declined by some 6 percent in 2009, the first time they have declined in a long time. Prospects for the next two years are expected to improve, but growth is likely to be weak and face a number of uncertainties, including with regard to the strength of the economic recovery—particularly in migrant-intensive industries such as construction—potential policy tightening that restricts migrant employment in host countries, and exchange rate movements.

Short-term prospects are likely to vary greatly among recipient countries. Despite the increasing geographical diversification of remittance sources—thanks to the emergence of many emerging markets—most LICs remain highly dependent on industrial countries as the dominant source of remittances. With an expected slower recovery in industrialized countries than in major emerging markets, recipient countries that are more dependent on industrialized countries are expected to have slower growth of remittances than those that have more diverse sources of inflows. Remittances to Latin America, for example, are expected to recover more slowly—despite having suffered a sharp decline in 2009—than those to most other parts of the world because of the region’s greater dependence on the United States and European countries. In contrast, many countries in South and East Asia, which receive a significant share of their remittances from the Middle East, have been less affected by the global financial crisis and are expected to see a more rapid recovery of remittances.

The longer-term prospects for remittances are uncertain. Strong economic growth and rapid global integration—through movement of goods, services, capital, and labor—have underpinned the extraordinary growth of remittances over the past two decades. However, the strong remittance flows may also have been a result of better recording due to improved statistics, reduced transaction costs of remitting, and the shift of remittances from informal channels (e.g., cash-carrying) to formal financial channels as a result of the Anti-Money Laundering and Combating the Financing of Terrorism campaign. It is unclear how these factors will develop in a post-crisis environment. On the positive side, the crisis does not seem to have undermined the process of global integration, and continued diversification of source countries in favor of more rapidly growing emerging markets may provide wider and more stable sources of remittances. On the negative side, the recent tightening of immigration policies in some destination countries (for immigrants) may have a negative impact on remittances in the medium term.

11. The evolution of risk ratings between pre- and post-crisis DSAs provide a complementary measure of the impact of the global financial crisis on the debt situation of LICs. Risk rating downgrades have been rare so far. Only two countries have experienced a downgrade during the last year. In the case of Eritrea, the rating downgrade (from high risk to “in debt distress”) reflected the accumulation of arrears since 2007. The downgrade of Georgia’s rating (from low to moderate) reflected both the impact of the conflict with Russia and the global financial crisis. Meanwhile, the Central African Republic’s and the Republic of Congo’s risk ratings were upgraded from high to moderate, thanks to the delivery of HIPC/MDRI debt relief at completion point.

12. This relatively favorable outcome reflects a number of factors. First, many LICs entered the crisis in a much better situation than in previous downturns. Macroeconomic policies and performance had improved substantially, and debt burdens of many heavily indebted LICs had been relieved though the HIPC and MDR Initiatives. Second, the assessment of debt vulnerabilities in DSAs is based on a long-term perspective: deterioration in debt ratios need not lead to a rating downgrade if it is temporary.13

13. Overall, the above analysis of post-crisis DSAs suggests that, broadly speaking, debt vulnerabilities remain manageable in LICs. Debt ratios have deteriorated in the short term, but are expected to return to a declining trend by 2011-12. Furthermore, the distribution of risks of debt distress has not materially changed.

14. However, this conclusion needs to be treated cautiously as it hinges on a few key assumptions, namely that the crisis will not adversely affect medium- and long-term growth; and that LICs, after having appropriately adopted accommodative fiscal policies in response to the crisis, will reduce their fiscal and current account deficits to more sustainable levels in the medium and long term, while continuing to have access to adequate financing.

15. In addition, these average results mask some important differences across LICs. Post-crisis DSAs indicate that debt ratios in LICs with low or moderate risk ratings converge to low levels over the projection horizon. However, countries at high risk continue to have significantly higher debt ratios, especially in relation to exports, even at the end of the projection period.

Further analysis of debt vulnerabilities in these countries, where such vulnerabilities were already relatively high prior to the financial crisis, is thus warranted. This is the topic of the next section.

III. Addressing Higher Debt Vulnerabilities

16. To complement the analysis of the impact of the crisis on debt sustainability in LICs, this section considers specific options to address the debt situation of LICs with “higher” debt vulnerabilities.14 Specifically, the analysis below is focused on all countries rated as being either at high risk of debt distress or in debt distress in their most recent DSAs. In addition to these countries that have high external debt vulnerabilities, the analysis below also considers the Maldives, St. Lucia, and St. Vincent and the Grenadines, where total public debt is high (see Table 3).

17. The circumstances of LICs with higher debt vulnerabilities are diverse, and this has implications for the options available to individual countries to address debt vulnerabilities. Such diversity relates to the nature and extent of their debt vulnerabilities, their HIPC Initiative status, their macroeconomic circumstances, the composition of their financing, and the structure of their debt.

Table 3.

LICs at High Risk of Debt Distress, or In Debt Distress

article image

Maldives, St. Lucia and St. Vincent and the Grenadines are also included in the sample on account of high total public debt vulnerabilities, defined to be a PV of public debt-GDP ratio in excess of 65 percent. Dominica, whose PV of public debt-GDP ratio experiences only a small and temporary breach of this threshold even under the baseline scenario is, however, not included in the sample

A. LICs Eligible or Potentially Eligible under the HIPC Initiative, and Similar Cases

18. All of the countries that remain eligible for debt relief under the HIPC Initiative are in debt distress, with the exception of Côte d’Ivoire. Nevertheless, there is significant variation in the extent of the debt vulnerabilities of these countries.

  • Extreme debt vulnerabilities characterize Eritrea, Guinea-Bissau, and Sudan, with PV of debt-to-exports ratios in particular ranging from 500 to 700 percent (i.e., from five to seven times their indicative DSF threshold) during the projection period, and with very large breaches of the debt service-to-exports ratios as well in some instances. 15

  • Less extreme, albeit still severe, debt vulnerabilities exist in the Comoros, the D.R.C., and Liberia, with some debt indicators roughly three to four times the indicative threshold. Debt service-to-exports ratios in the D.R.C. and the PV of debt as a ratio to both exports and GDP in Liberia reach particularly high levels.

  • By contrast, debt vulnerabilities are more contained in Guinea and Togo, with peak levels of all debt indicators less than two times the indicative DSF thresholds.

  • Similarly, consistent with its high risk rating, Côte d’Ivoire’s indicators are also somewhat lower, in part because it has already received substantial debt relief from many of its creditors.

19. For these countries, debt relief under the HIPC Initiative and the MDRI is expected to significantly improve their debt outlook. Many of these HIPCs have already reached the decision point, and are implementing Fund-supported programs aimed at getting them to completion point in the near future. Some others, however, have not yet reached the decision point, face ongoing security related risks, and will need to overcome significant challenges to benefit from the Initiatives.16

20. Sound policies will be needed in these countries before and after completion point. A track record of sound macroeconomic policies, as well as measures (“triggers”) focusing on improving public financial management, are some of the main requirements to reach the completion point. Such policies contribute to the reduction of debt vulnerabilities, and they will also need to be pursued after debt relief has been provided at the completion point to keep debt vulnerabilities at a lower level.17

21. Myanmar and Zimbabwe, which are also in debt distress, but not at present HIPC-eligible, are also likely to require comprehensive debt relief to restore debt sustainability. Myanmar could become potentially eligible for HIPC Initiative relief if, when its debt data become available, it can be demonstrated that it meets the HIPC Initiative eligibility criteria based on end-2004 data.18 Zimbabwe is in a different situation vis-à-vis the HIPC Initiative, as it does not meet the World Bank’s income criterion at end-2004 (it was not IDA-only at the time).19 However, Zimbabwe’s debt situation is such that it too is eventually likely to require comprehensive and coordinated debt relief from all its creditors.

B. Non-HIPCs and Post-HIPC-Completion-Point LICs

22. The severity of debt problems also varies across the non-HIPCs and post-HIPC-Completion-Point LICs with higher debt vulnerabilities. While all have a high risk rating, their vulnerabilities are quite different in nature, intensity, and immediacy (Table 4).20, 21

  • Afghanistan, Grenada, Tajikistan, and Tonga face the most severe debt problems, with the breaches of the applicable thresholds that are large and sustained. Breaches in excess of the respective thresholds of 50 percent of exports or 10 percent of GDP persist for at least 5 consecutive years.22 It should be noted that Tajikistan and Tonga enjoy very high levels of remittances, which constitute an important mitigating factor when their debt vulnerabilities are considered.

  • Grenada and Tonga also have high current debt-service ratios, which constitute a more immediate risk to debt sustainability. While Tonga’s debt service-to-exports ratio already exceeds its applicable DSF thresholds substantially and on a protracted basis, in Grenada there is a small breach for only one year, but this ratio remains elevated (exceeding 15 percent) for a number of years.

  • São Tomé and Príncipe also experiences a large breach of a DSF threshold, but here, debt vulnerabilities would be substantially mitigated if oil production comes on stream as expected in the DSA.

  • St. Lucia and St. Vincent and the Grenadines have persistently high levels of public debt over the entire sample period.

  • In the remaining countries debt vulnerabilities are less severe—breaches of thresholds in baseline DSA scenarios are more modest, or temporary, or occur only in the medium or long term.

Table 4.

High Debt Vulnerability Indicators

(In percent; unless otherwise indicated)

article image

In years.

For the PV of public debt to GDP ratios refers to years in excess of 65 percent of GDP.

Options to Address Debt Vulnerabilities

23. A number of options can be considered to address the debt situation of the 14 non-HIPCs and post-completion-point LICs identified as having higher debt vulnerabilities. These include: (a) reforms aimed at improving policy and institutional performance (particularly in the economic and debt management areas), and through them a country’s capacity to carry debt; (b) stronger fiscal positions, including to restore policy space post-crisis; and (c) better financing terms from donors/creditors.

24. These options are explored below through illustrative simulations of individual country DSAs.23 Specifically, the impact on the most recent DSAs of stylized scenarios of improvements in policy and institutional capacity (measured by CPIA scores), fiscal position, and borrowing terms are simulated. These scenarios should not be seen as recommended strategies for any particular country. Rather, they are stylized exercises that aim to provide a sense of the impact that such hypothetical options, individually or in combination, could have on debt vulnerabilities. In these scenarios, vulnerabilities are considered to be substantially reduced by a certain date when all debt indicators are below their applicable thresholds at that date and beyond.

25. Growth-enhancing policies should also be considered to reduce debt vulnerabilities whenever possible. Such policies, which are desirable in and of themselves, are already incorporated to varying degrees in baseline scenarios in DSAs. For this reason, and to err on the side of caution, the impact of such policies is not simulated in this paper. However, the impact of higher growth on debt vulnerabilities should not be underestimated over the long term.

Improvements in Policy and Institutional Capacity

26. Enhancing policy and institutional performance improves countries’ capacity to carry debt, among other benefits. Empirical evidence supports the view that the higher the quality of a country’s policies and institutions, the better its capacity to carry debt.24 This evidence led to the inclusion in the DSF of policy-dependent thresholds which are used to assess a country’s risk of debt distress. In the context of the DSF, better policies and institutions translate into better CPIA scores, which lead to higher indicative DSF thresholds for countries that are not already strong performers. While the DSF thresholds are based on the full CPIA, some sub-categories, particularly those related to macroeconomic and public financial management, are especially relevant for debt sustainability. In this regard, technical assistance from the Bank and the Fund in such areas, including debt management, could contribute to improving capacity.

27. Simulations show that modest but sustained improvements in policies and institutions could significantly reduce debt vulnerabilities in a number of vulnerable LICs. Table 5 below describes recent trends in CPIA scores of vulnerable LICs and shows the impact of a one-half percent annual increase in the CPIA score over 2009–14 on the policy capacity rating.25, 26 Even with such relatively modest improvements, which fall within the range of improvements in several of the vulnerable LICs in recent years, six of the 14 countries could migrate to a higher policy performance rating by 2014.27 This, in turn, could result in two countries moving out of the high risk category as soon as 2014, and more beyond, assuming that fiscal policies remain unchanged from the baseline scenarios (Table 6).

Table 5.

Country Policy and Institutional Assessment Ratings Simulations

article image
Source: The World Bank and staff simulations.

In percent.

Rating based on 3-year average of CPIA scores.

Simulated 3-year average rating based on a 0.5 percent per anum growth of annual CPIA scores over 2009-14.

Based on the simulation, Tajikistan's 3-year average CPIA score in 2014 reaches the threshold for a medium rating, but not the requirement to maintain it for two consecutive years in order to qualify for a rating upgrade.

Table 6.

Impact of Capacity Improvements on Risk of Debt Distress under the Baseline Scenario

article image

Excludes Maldives, St. Lucia, and St. Vincent and the Grenadines, since there are no formal public debt thresholds associated with capacity categories.

This is in relation to the above list of countries with breaches of thresholds at the respective dates.

Macroeconomic performance

28. Substantial strengthening of fiscal and external accounts is already incorporated in the baseline scenarios of most LIC DSAs. The extent of adjustment varies across countries, and reflects their circumstances. As shown in Panel 4, assumed fiscal adjustment tends to be larger in the countries where fiscal deficits are higher.

Panel 4.
Panel 4.

Countries in High Risk of Debt Distress—Macroeconomic Adjustment in the Baseline

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Note: All data as a percent of GDP.Source: Fund staff projections.1/ Fiscal ad justment and current account adjustment defined as the difference between the values of the average fiscal and current account deficits in 2010-28 and 2010-14 respectively and their values in 2009. A positive value indicates lower deficits.2/ For Burundi, averages compared against fiscal deficits in 2008.

29. The impact of a further strengthening of fiscal positions on debt vulnerabilities is modeled in two illustrative scenarios. Specifically, the simulations explore the impact of primary fiscal spending that is 0.5 percent of GDP, and 1 percent of GDP respectively, lower than in the baseline DSA scenario over the entire projection period.28 As emphasized above, these scenarios are meant to be purely illustrative; fiscal consolidation, in practice, could also be obtained through revenue measures. Any adjustment would need to protect priority spending, including expenditures on poverty alleviation.

30. In the near term, higher debt vulnerabilities persist in most countries, but fiscal consolidation produces results over time. Table 7 provides the results of the simulations for both scenarios. The discussion below focuses on the 1 percent of GDP scenario, but the results under the other scenario are qualitatively similar, especially over the near term (by 2014).

  • By 2019, high external debt ratios persist only in Tajikistan, Tonga, and Afghanistan (where there is no breach in the early projection period).29 A modest breach of a threshold is still recorded in Gambia, and a marginal one in Burkina Faso. However, in St Lucia, and in St. Vincent and the Grenadines, significant public debt vulnerabilities persist.

  • In 2028, higher debt vulnerabilities persist only in Afghanistan, St. Lucia, St. Vincent and the Grenadines, and Tajikistan.

Table 7.

Illustrative Fiscal Adjustment Scenarios 1/

article image

Figures in bold in shaded cells represent a breach of thresholds.

2/ Please refer to Appendix Table, “Debt Burden Ratios in Baseline Scenario.”

Defined as a permanent reduction in public primary spending starting in 2010.

More Favorable Financing Terms

31. LICs benefit from significant levels of concessional financing in their efforts to maintain debt sustainability. Table 8 indicates that concessional financing is projected to account for a very significant share of total external financing for the vulnerable LICs under consideration in the next few years.30 The average concessionality of donor lending has also been very high, with generally more generous average grant elements of financing to countries with more severe debt vulnerabilities, including to those without a Fund-supported program.

Table 8.

Concessionality of External Financing in Vulnerable LICs 1/

article image

As defined in DSAs.

Average grant element of new financing, in percent, calculated with the discount rate in DSAs, which may differ from that used in calculating the concessionality of loans in Fund-supported programs.

32. The simulations suggest that a significant increase in the grant element of new financing would be required to reduce debt vulnerabilities substantially in many countries. Table 9 provides the results of two scenarios in which the average grant element of new loans is higher than in the baseline DSA scenario by 5 and 10 percentage points, respectively. Such a measure takes time to reduce debt vulnerabilities as it only affects new borrowing, and not the existing stock of debt.

Table 9.

Impact of Debt Burden Indicators under Concessionality Scenarios

article image

This is in relation to the above list of countries with breaches of thresholds at the respective dates.

33. The established policies of a number of donors could actually deliver a larger than assumed (in DSAs) average grant element of new financing in some vulnerable LICs. These donors—particularly large multilateral institutions such as IDA and some regional development banks—have policies under which the terms of new financing depend on the debt sustainability situation of the recipient. For instance, IDA provides financing on grant terms to countries at a high risk of debt distress or in debt distress, while low-risk countries get concessional loans. However, for high-risk countries, DSAs normally rely on conservative assumptions with regards to these financing terms, as DSAs are used to establish an eventual need for more favorable terms. For instance, IDA financing is assumed to be in the form of concessional loans, not grants, in the medium and long term. A vulnerable LIC which would continue to remain at a high risk of debt distress in the future would therefore get more concessional terms from these donors than assumed in the latest DSA. For countries where these donors are projected to provide the bulk of new loan resources in the DSA, this could change the debt sustainability outlook much more substantially than suggested in the simulations described in the previous paragraph.

Combined options

34. To address the debt situation of LICs with higher debt vulnerabilities, the above options could be combined. While in the previous sections the impact of various policy options was assessed separately, we explore here the impact of implementing these options together. These scenarios illustrate the impact of efforts by both borrowers and their creditors.

35. A combination of the illustrative policy scenarios—improvements in policy performance, less borrowing, and more favorable borrowing terms—can substantially address the debt vulnerabilities of most vulnerable LICs (Table 10).31

  • By 2014, large breaches of thresholds are limited to São Tomé and Príncipe, Tajikistan, and Tonga. Smaller breaches are observed for Burundi, Djibouti and Grenada.

  • The situation improves still further by 2019, with only small breaches in Afghanistan and Tajikistan.

  • By 2028, only Afghanistan remains in breach of its threshold.

  • Higher public debt vulnerabilities persist in St. Lucia and St. Vincent and the Grenadines over the whole period.

Table 10.

Capacity Improvements, Fiscal Adjustment, and Concessional Borrowing 1/

article image

Numbers in bold in shaded cells indicate a breach of thresholds.

Based on an improvement of the annual CPIA score of 0.5 percent per anum over 2009-14.

A permanent reduction in public primary spending of 1 percent starting in 2010 and an increase in the average grant element of borrowing by 10 percent.

The applicable threshold till 2014 is 100 percent of exports and 150 percent thereafter.

The case of countries where combined options do not seem sufficient

36. Afghanistan will likely require continued provision of external financing mostly in the form of grants in the medium term. The profile of Afghanistan’s DSA is largely driven by the assumption that grant financing, which currently amounts to 50 percent of GDP, would be gradually replaced with concessional loans. The DSA illustrates that too rapid a shift would run the risk of a sustained increase in debt ratios (from the current relatively low levels). Nonetheless, with much of Afghanistan’s external financing needs likely to be met by IDA and the AsDB, these institutions’ policies on financing LICs (see paragraph 31) should prevent such a rapid shift to concessional loans from taking place. Thus, Afghanistan’s debt ratios would eventually not increase as quickly as outlined in the baseline and adjustment scenarios.

37. Stronger fiscal adjustment than simulated above may be required in St. Lucia and St. Vincent and the Grenadines. The baseline scenarios in the most recent DSAs are meant to illustrate that current policies, or policies involving only modest fiscal adjustment, are not sustainable in either country. These DSAs include alternative “active” scenarios which suggest that, given the current fiscal situation, larger—but still realistic—fiscal adjustment would be required to bring about a substantial decline of debt ratios. The recommended adjustment is significantly larger than that simulated above.

Debt Relief

38. Notwithstanding the possibility that higher debt vulnerabilities can be addressed through sustained efforts by debtors and creditors, the need for debt relief at some point in the future cannot be excluded in some cases. It will take time for the policy options discussed above to resolve higher debt vulnerabilities in some countries. Should one or these countries experience difficulties in servicing their debts during this period, for instance in the wake of a large negative shock, debt relief may need to be considered.

39. Debt relief under traditional mechanisms can be particularly effective where bilateral debt, especially to Paris Club creditors, accounts for a substantial part of the country’s outstanding debt.32 Paris Club members have indeed provided such relief consistently in past decades when the need arose. Where the Paris Club accounts for a more limited share of the country’s debt, the effectiveness of traditional mechanisms depends on the willingness of other bilateral and commercial creditors to provide comparable treatment, which the debtor is expected to seek under Paris Club rules.

40. The share of Paris Club creditors in the most vulnerable LICs’ debt is relatively small (Table 11).

  • Official bilateral creditors represent about 33 percent of total debt of which nearly two-thirds is owed to non-Paris Club creditors.

  • Commercial creditors represent 6 percent of total debt.

  • Multilateral creditors as a group account for 61 percent of the debt of vulnerable LICs, of which IDA represents 21 percent, the Asiandevelopment Bank (AsDB) 20 percent and other smaller multilateral institutions 16 percent. The share of the Fund is about 4 percent.

Table 11.

Composition of External Debt (In percent of PV of total debt at end-2009) 1/

article image
Sources: DSAs, GDF, and Fund staff estimates.

Following the DSF methodology using the 4 percent uniform discount rate to calculate the PV of future external debt-service obligations (SM/10/16).

Based on a weighted average, with weights equal to the share of each country’s debt in the total debt to the creditor from all vulnerable LICs.

Source: DSAs, GDF, and Fund staff estimates.

41. However the composition varies significantly across countries.

  • In post-MDRI countries the share of multilateral debt is generally higher than the average. Official bilateral debt is almost entirely due to a few non-Paris Club countries. Debt to Paris Club creditors is generally limited, reflecting very generous debt relief at the completion point and the provision of new financing mostly in the form of grants.

  • A few countries have debts to bilateral creditors (official or commercial) largely exceeding one third (Grenada, Maldives, São Tomé and Príncipe, Tajikistan and Yemen). Some of them (Grenada and Maldives) have a sizeable debt to commercial creditors.

  • The claims of large multilateral creditors are often concentrated on a few countries (e.g., Afghanistan for the AsDB).

42. This structure of debt suggests the following:

  • Relief on all bilateral debt could have a substantial impact on some countries, but only a modest one on others.

  • In all countries, the impact of traditional debt relief would be rather small without comparable treatment from non-Paris Club bilateral and commercial creditors.

IV. Conclusions

43. Recent DSAs indicate that the global crisis has increased LICs’ debt vulnerabilities significantly. The main debt ratios are expected to be negatively affected, as financing requirements have increased while LICs’ payment capacity is expected to be permanently lower.

44. The global crisis, however, is not expected to result in systemic debt difficulties across LICs. Debt ratios are expected to return to a downward trend, and risk rating downgrades have been rare in post-crisis DSAs. Critical assumptions behind this outcome are that the crisis has no permanent impact on long-term growth and that LICs restore the policy space used during the crisis as the recovery firms up.

45. The share of LICs that face higher debt vulnerabilities is significant but has not increased with the crisis. These countries require close monitoring and concerted and sustained action:

  • In about half of these countries, debt vulnerabilities are expected to be reduced substantially through HIPC/MDRI relief (or will require similar treatment).

  • In the other half, debt vulnerabilities could be effectively addressed with concerted efforts from both LICs (enhanced institutions and policies, better fiscal position) and the international community (improved financing terms).

  • These efforts would need to be sustained to produce results.

46. Nevertheless, the need for debt relief in some of these countries at some point in the future cannot be excluded. The effectiveness of traditional debt relief mechanisms would hinge on the participation of all bilateral and commercial creditors, as the share of Paris Club debt in total debt is relatively limited in these countries.

Annex 1. List of DSAS used in the Paper

article image
Source: Debt Sustainability Analyses.

Descriptors for Countries included in Pre- and Post-Crisis Analysis

article image
Source: Debt Sustainability Analyses, Staff estimates

Annex 2. The Medium- and Long-Term Effects of the Crisis on Growth in LICs33

Whether the global financial crisis has a persistent negative growth effect on LICs depends crucially on the nature of the shock. While the origin of the crisis in advanced economies was in the financial sector, most LICs were primarily hit by sharply lower demand for their exports (external demand shock, or ED shock), lower FDI, and for fuel-exporting LICs, a negative terms of trade shock (TOT shock).34

Current WEO projections imply a more rapid recovery of growth in LICs than has been experienced in past global crises. Compared to past global crises, the current crisis is distinguished by the severity of the downturn and the synchronization between LICs and global cyclical growth movement (Imbs, 2010). In past global crises, LICs have tended to recover more slowly than the rest of the world (Figure 1, left panel). However, the current WEO forecasts imply a more rapid V-shaped recovery path out of the recession compared to previous crises, an observation that also applies to other economies. It is notable that unlike in previous crises where terms of trade (TOT) growth moved sharply downward relative to external demand (ED) growth, in the current crisis it is ED that has declined most (Figure 1, right panel).

Figure 1.
Figure 1.

GDP per capita, ED and TOT growth in past and current crises

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Note: The left panel plots the average per capita GDP growth in the world and in LICs while the right panel plots the TOT and ED growth in LICs 5 years before and 5 years after the global crises (centered at zero on the horizontal axis) of 1975, 1982 and 1991, and the current crisis. Also shown in dashed lines are WEO projections until 2013.

Impulse response functions show evidence of large and persistent output losses from external shocks in the medium-term. An impulse response exercise (based on Cerra and Saxena, 2008) can assess the extent to which TOT and ED shocks have historically been associated with permanent output losses. The impact on output is negative and highly persistent under both TOT, and particularly ED shocks (Figure 2). Output losses continue to rise without a sign of a reversal for a number of years after an ED shock, resulting in a cumulative loss of over 6 percent of GDP. The output loss path eventually becomes flat as growth reaches its pre-crisis rate but after lower growth in the medium term and a substantial loss of output.

Figure 2.
Figure 2.

Impulse response of output loss in LICs to TOT and ED shocks

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Note: The left and right panels present impulse responses of output loss, measured as the percentage change from a linear growth trend, to a TOT shock and an ED shock, respectively. The solid line is the mean of output loss, and the dashed line reflects one standard deviation from the mean.

While ED shocks tend to be associated with persistent losses in output levels, they are not linked to long-term downbreaks in growth. A growth breaks methodology (based on Berg, Ostry and Zettelmeyer, 2008) identifies sustained periods of slow growth in LICs and examines whether TOT and ED shocks are correlated with such “cliffs”. TOT growth tends to decrease sharply in the run-up to growth decelerations, which provides suggestive evidence that sharp declines in TOT growth may lead to a sustained period of slow growth (Figure 3, left panel). On the contrary, ED growth shows virtually no co-movement with a growth downbreak (Figure 3, right panel).35 Given evidence that the current crisis has affected LICs primarily through ED rather than TOT (Berg et al. 2010, IMF 2009), these results suggest a low probability that many LICs will suffer from a protracted period of slow growth due to the crisis.

Figure 3.
Figure 3.

TOT vs. ED around periods of growth decelerations in LICs

Citation: Policy Papers 2010, 032; 10.5089/9781498337564.007.A001

Note: The left and right panels plot the behavior of TOT and ED, respectively, in the period leading up to, and following, growth downturns (year 0 on the horizontal axis).