97567 DISCUSSION PAPER African Debt since MFM Global Practice HIPC: No. 2 How Clean is the Slate? March 2015 Dino Merotto Tihomir (Tish) Stucka Mark Roland Thomas MFM DISCUSSION PAPER NO. 2 Abstract The paper finds a moderate evolution in public debt ratios since debt relief among HIPC and MDRI recipient countries in Sub-Saharan Africa, with certain exceptions. For eight countries we find rapid rates of debt accumulation, which could return them to pre-HIPC debt levels in only a few years. Short-term domestic debt has, despite early fears, in general not filled the borrowing space created by debt relief. However, external debt accumulation on commercial terms in some cases presages repayment spikes, which may combine with short-term domestic obligations to amplify refinancing risk and cause abrupt reductions in public spending, with damaging consequences for development. Finally, despite reduced debt, African economies continue to be commodity dependent and prone to shocks. As global interest rates and commodity prices revert to historically more customary levels, these countries should remain prudent: avoid tax-base erosion, prevent large recurrent spending hikes, and invest wisely in growth, by executing projects effectively to enhance infrastructure. These fiscal fundamentals will be as important for debt sustainability as how much is borrowed and on which terms. Corresponding author: (dmerotto@worldbank.org) JEL Classification: E6, H6, O1, O2 Keywords: Debt, poverty, Africa ________________________________________ This series is produced by the Macroeconomics and Fiscal Management (MFM) Global Practice of the World Bank. The papers in this series aim to provide a vehicle for publishing preliminary results on MFM topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For information regarding the MFM Discussion Paper Series, please contact the Editor, Bernard Funck at bfunck@worldbank.org. © 2015 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 All rights reserved. Cover design: Marie-Anne Chambonnier African Debt since HIPC: How Clean is the Slate? Dino Merotto, Tihomir (Tish) Stucka, and Mark Roland Thomas 1. Introduction International debt relief efforts for low-income countries (LICs) gained traction in the late 1990s and built momentum through the 2000s. These initiatives, mainly comprising the Heavily Indebted Poor Countries (HIPC) Initiative, its close cousin the Multilateral Debt Relief Initiative (MDRI), and Paris Club debt rescheduling, have committed over $100 billion dollars in the form of non-payment of current and future debt obligations to over 40 countries (Table 1). About 85 percent of these commitments have been made to African economies. Various assessments have attributed degrees of success to this effort.1 After a slow start, international debt relief responded to civil society demands, and delivered on donor governments’ stated intent, to forgive most debts of the target group of countries. As a consequence, by the definition used by the Bretton Woods financial institutions, measures of debt sustainability among these countries improved dramatically. Beyond these narrow financial measures of impact, the inclusion of policy content in the design of debt relief programs – in an effort to encourage countries to use their newly acquired fiscal space to reduce poverty – seems, according to several authors, broadly to have had its desired effect. This synopsis skims a mountain of country-level analysis and oversimplifies: it should certainly not be read as a lazy declaration of victory or a claim that more could not have been done. This paper focuses on African countries’ public and publicly guaranteed debt since debt relief. It asks how African HIPCs have borrowed post-debt relief, whether this borrowing is any cause for concern, and what the conclusions are for policy. The results, compiling evidence from about 30 countries, paint quite a mixed picture. Overall the news seems to be fairly good: in the five-plus years since debt relief, most countries have borrowed responsibly and show no signs of rapid debt re-accumulation. Jutting from this comforting landscape are some jagged outcrops of concern. A number of countries are borrowing too fast; on current trends each could reach pre-debt relief levels of external debt in only a few years. More broadly, the challenges facing today’s emerging market debt managers are more complex than those that existed a decade or more ago, before debt relief. The range of lenders and debt instruments has increased. In many cases natural resource  The authors are respectively Lead Economist, Senior Economist, and Manager, affiliated with the Macroeconomic and Fiscal Management Global Practice at the World Bank. All statements in this paper are expressions of the views of the authors and do not reflect official views of the World Bank or its Executive Directors. 1 Cohen (2001); World Bank (2001); Easterly (2002); Clements et al (2004); Romero-Barrutieta et al (2011); and Thomas and Giugale (2014), to mention only a few. 1 revenues or the prospects of them have attracted lending but at the same time created public impatience for quick results. Lastly, international capital markets may have brought a new source of discipline to macroeconomic management. This discipline may be more rigorous than that that hitherto provided exclusively by the official sector; it may also be more capricious. Table 1 HIPC and MDRI Debt Relief nominal nominal terms as PV terms as a in US$ bil PV terms terms a share of total share of toal HIPC 65.7 49.4 62 64 MDRI 41.0 28.3 38 36 total 106.8 77.7 100 100 Source: HIPC Progress Report (2013) The next section presents public debt ratios for African HIPCs in recent years. The third section discusses the findings of the debt sustainability analyses that are conducted by World Bank, IMF, and government staff from the countries that have access to concessional (or “soft�) financing windows. To understand this evidence it is also necessary to give a brief outline of how these analyses are constructed. The fourth section provides more detailed analysis of public debt ratios by type of borrower as well as lender (source of finance). The concluding section gives our views on implications for international debt policy. 2. Average Debt Ratios in African HIPCs Country Coverage There are 33 African HIPCs.2 Of these, 30 have reached decision point under the Initiative (Eritrea, Somalia and Sudan have not) and 29 have reached completion point (Chad has not yet). Completion point implies irrevocable debt relief on the part of participating creditors and also leads directly to MDRI cancellation of IMF, World Bank, and African Development Bank (AfDB) debt. In nearly all cases, therefore, it is completion point that has the most discernible impact on stock ratios of debt as well as on flow ratios of debt service. Of the 29 completion point African HIPCs, 26 had reached completion point by 2010, allowing an initial assessment of the impact of debt relief on their subsequent borrowing decisions (this criterion excludes Comoros, Côte d’Ivoire, and Guinea). Finally, inadequate debt data ruled out using Madagascar for our analysis. Our core country coverage is therefore 25 African HIPCs with adequate track record and data since receiving MDRI debt cancellation. The complete set of African HIPCs as of May 2014 is provided in Table 2. 2 This list does not preclude other countries becoming eligible for debt relief that are today not so considered for possibly transient reasons. Instances of countries that could conceivably qualify later include Somalia and Zimbabwe. 2 Table 2 Timing of HIPC and MDRI Debt Relief MDRI 1/ 2006 2007 2009 2010 2011 2012 Benin (2000) Gambia (2000) Burundi (2005) Congo, R. (2006) Togo (2010) CIV (2012) Burkina Faso (2000) STP (2000) CAR (2007) DRC (2003) Comoros (2010) Cameroon (2000) Guinea-Bissau (2000) Guinea (2010) Ethiopia (2001) Liberia (2008) Ghana (2002) Malawi (2000) Mali (2000) Madagascar (2000) Mauritania (2000) Mozambique (2000) Niger (2000) Rwanda (2000) Senegal (2000) Sierra Leone (2002) Tanzania (2000) Uganda (2000) Zambia (2000) 1/ Years in brackets represent HIPC decision point Source: HIPC Progress Report (2013) Debt Evolution For each of these 25 countries, we constructed several debt ratios at three points in time. The first point was just prior to reaching HIPC decision point: in most cases this would constitute a high watermark for external debt and represents our “before� measure of debt. The second point is just after completion point (and therefore factors in MDRI debt relief as well). The difference between debt ratios at these two points is a rough estimate of the overall impact of debt relief on these 25 countries, because the effect of debt relief is of a larger magnitude than any borrowing in the interim. The third point is the most recent available data, which is 2013 (the debt sustainability analyses [DSAs] used to provide the data are conducted annually). Debt ratios may be constructed from several measures. In this paper we present debt-to-GDP ratios, because these are most commonly understood and used as reference points in many discussions of indebtedness. For example, emerging market governments often adopt explicit targets in their medium term economic plans for debt-to-GDP in the range of 30 to 50 percent. Further, debt may be expressed in nominal terms (summing the face value of debt instruments) or in present value (PV) terms (summing the flow of payment obligations discounted to the present or another reference year). Nominal ratios are usually less trouble to calculate. PV calculations better reflect the less burdensome terms of many loans to poor countries and are the ratios targeted by debt relief restructurings under both HIPC and Paris Club deals. Finally, when taking averages across countries, these may be weighted (by, for example, GDP) or 3 unweighted. In weighted averages larger economies will drive the results; in unweighted or simple averages very small economies may bias the analysis. Figure 1 shows weighted debt ratios for our group of 25 economies at the three key points, comparing the trends for nominal and PV measures. The average of the most cited debt ratio, nominal debt-to-GDP, was over 100 percent just prior to decision point. As has been documented elsewhere and was underlined by debt campaigners, this represents an untenable state of affairs for economies with average per capita income (at the time) of less than US$3 a day per person. The PV ratio was less than half this nominal value (41 percent); the difference between these two ratios captures the fact that many of these debts were in the form of concessional finance from development banks and aid agencies. Debt relief – both HIPC (including comparable treatment from many bilateral and in some cases commercial creditors) and MDRI (from the three main multilaterals) – cut these ratios by more than two thirds. In nominal terms the average debt-to-GDP ratio fell from 105 percent to 27 percent. In PV terms the ratio fell from 41 to 13 percent. The debts remaining after debt relief were owed principally to four sources: (1) creditors that refused to participate in HIPC (e.g., some commercial creditors); (2) remaining obligations to participating creditors after the common reduction factor was applied to their claims (e.g., some non-Paris Club bilaterals); (3) debt contracted with the three main multilaterals after December 2004; (4) other debt contracted in the interim period between decision and completion points. It is striking that according to these average ratios borrowing since debt relief has been quite moderate. The GDP-weighted average nominal debt-to-GDP ratio rose from 27 percent to 39 percent in the five or more years to 2013 (most countries in the sample had reached completion point by 2008): average increments of 2 to 3 percentage points a year. In PV terms the story is the same: debt-to-GDP rose from 13 to 18 percent, about one percentage point per year. Of course debt ratios cannot increase indefinitely at a constant rate without eventually causing problems. Yet against this it should be noted that the stated aim of debt relief was not to keep debt ratios forever at their new post-debt relief levels, but rather to give countries the room to borrow again responsibly. The question policy makers have been debating since is what “responsibly� should look like. Figure 1 GDP-Weighted Average Public Debt-to-GDP Ratios for 25 African HIPCs (Nominal and Present Value Measures) 4 120 45 41 Nominal Public Debt to GDP, weighted (LHS) 40 100 PV External PPG Debt to GDP, weighted (RHS) 35 80 30 25 105 60 18 20 40 13 15 10 20 39 27 5 0 0 Decision MDRI 2013 Source: IMF/World Bank Joint DSAs, authors’ analysis Using unweighted rather than GDP-weighted averages does not alter this narrative appreciably (Figure 2). The pre-debt relief average nominal debt-to-GDP ratio was even more extreme, 146 percent. In PV terms this measure was a third as much at 48 percent (i.e., the unweighted average grant element of lending was about two thirds). The higher nominal ratio and the higher grant element both reflect some extreme values among smaller countries that are filtered out by the GDP-weighting. Notable cases are Liberia (debt to GDP of 554 percent), São Tome and Príncipe (388 percent), Mauritania (281 percent) and Guinea Bissau (234 percent). These debt levels were made up of highly concessional loans, so the PV measure is less affected than the nominal by the weighting of averages. In unweighted terms, debt relief brought these ratios down from 146 percent to 31 percent (nominal) and from 48 percent to 15 percent (PV), a similar proportionate decline as the weighted statistics. Borrowing behavior since completion point also looks moderate, as in the weighted statistics, with the nominal ratio rising back up by 10 percentage points (about two p.p. per year) to 41 percent, while the PV ratio rises at about half this rate, 5 percentage points (about one p.p. per year). The analysis so far concerns external debt, the object of the target ratios and reduction factors of the HIPC Initiative and Paris Club rescheduling. An oft-repeated question since debt relief is whether the focus on external debt may have caused a substitution towards domestic borrowing by some countries. The evidence from debt sustainability analyses since debt relief suggests that African HIPCs have not made disproportionate use of domestic borrowing. We draw on a slightly larger sample for this analysis, comprising the 25 countries used above plus Chad (which has reached decision point), recipients of debt relief after 2009 (Comoros, Côte d’Ivoire, and Guinea) and recipients of Paris Club debt relief excluded from the HIPC Initiative (Kenya and Nigeria, which are of interest because of their significant domestic debt markets), making a sample of 31 African debt relief recipients. Figure 2 Unweighted Average Public Debt-to-GDP Ratios for 25 African HIPCs (Nominal and Present Value Measures) 5 160 60 Nominal Public Debt to GDP, weighted (LHS) 140 48 PV External PPG Debt to GDP, weighted (RHS) 50 120 40 100 146 80 30 60 19 20 15 40 10 20 31 41 0 0 Decision MDRI 2013 Source: IMF/World Bank Joint DSAs, authors’ analysis For this sample of countries, external debt ratios and total public debt ratios were extracted from DSAs from the years 2005 and 2013. 2005 is taken as a representative year before the provision of the bulk of debt relief: MDRI does not show up in debt ratio calculations until 2006 at the earliest and Nigeria’s large Paris Club write-off came in 2005. Domestic debt statistics were then constructed as the difference between total public debt and external public debt.3 The weighted average of external public debt to GDP for these 31 countries fell by about half between 2005 and 2013, from 45 percent to 23 percent (Figure 3). Over the same period, the weighted average ratio of domestic public debt to GDP rose only from 12 percent to 15 percent. The data thus lend little credence to the story of substitution from external to domestic sources of financing. Moreover, this small rise in domestic debt is mainly driven by Nigeria. This is not unexpected: larger economies tend to have deeper domestic debt markets (so one expects weighted averages to show a greater tendency towards domestic borrowing) and Nigeria has undergone a well-documented deepening of its domestic debt market recently. Using unweighted statistics, the average ratio of domestic debt to GDP in our sample of 31 countries has remained constant at about 14 percent over the period (in fact rising very slightly from 14.2 percent in 2005 to 14.8 percent in 2013). The moderate return to borrowing documented earlier in this section has thus been predominantly driven by external borrowing. Figure 3 GDP-Weighted Average Debt-to-GDP Ratios for 31 African countries: External versus Domestic Debt Evolution 3 This approach implies differentiating between external and domestic debt according to the currency of issuance rather than by the identity of the holder of the debt. 6 50 45 45 av ext PPG debt-to-GDP av dom debt-to-GDP 40 35 percent of GDP 30 25 20 1516 15 12 10 5 0 2005 2013 pre-MDRI and last actual year Source: IMF/World Bank Joint DSAs, authors’ analysis 7 3. Risks of Debt Distress Debt distress is defined here as either accumulation of arrears, default or formal rescheduling by some subset of creditors (e.g., Paris Club) at the request of the debtor.4 This definition is the one that underpins the debt sustainability framework that has been used by the World Bank and the IMF since 2005 for low- income countries (defined as countries with access to these institutions soft credit windows). Under this framework, countries are rated annually in one of three categories of risk of debt distress: low risk, moderate risk, or high risk/in debt distress. Countries actually in default are by the above definition “in debt distress.� These ratings are based on a 20-year forward simulation of debt ratios under different economic scenarios for the country. A baseline scenario uses the expected path of all the major drivers of debt ratios, such as fiscal deficits, capital account flows, exchange rates, and economic growth. If any of the country’s debt or debt service ratios rise above certain pre-determined indicative thresholds under this baseline, it is deemed to be at high risk.5 Several alternative scenarios then posit departures from the baseline (usually more pessimistic departures): in cases not assessed high risk under the baseline, if under any of these alternative “stress test� scenarios any threshold is breached, the country is rated as at moderate risk. Countries that breach no thresholds under any scenarios are categorized as low risk. The framework is used by several international financial institutions including the World Bank to help govern decisions over lending terms and allocations, with the aim of helping countries avoid future difficulties servicing debt.6 Using this framework as a yardstick, the risks of debt distress among African low-income countries have diminished significantly. For the 37 countries for which we have continuous data, in 2006 half (18) were either in debt distress or at high risk and only five were at low risk (Figure 4). By 2012 this pattern had reversed: debt distress and high risk countries made up the smallest class (11), while more than a third of countries were at low risk. Three things can be said about these trends. First, despite improvement, it is still a concern that more than two-thirds of African countries are at moderate or high risk of debt distress and in particular that one-fifth of countries are either at high risk or actually in debt distress. Second, even in the best of policy environments, the approach can result in countries with low debt today still being rated at moderate risk. This is because the stress test scenarios reflect the country-specific historical volatility of certain economic variables (e.g., exchange rates), which can drive up debt ratios in these scenarios. The number of countries at moderate risk can therefore be viewed with some forbearance. Finally, the apparently 4 Kraay, A. and V. Nehru (2006), “When Is External Debt Sustainable?� World Bank Economic Review, Vol. 20, No. 3, pp. 341-365. 5 The five ratios analyzed are (i) debt to GDP, (ii) debt to government revenue, (iii) debt to exports, (iv) debt service to government revenues, and (v) debt service to exports. The thresholds differ by country to take into account countries’ policy and institutional capacity, using ratings made annually under the World Bank’s Country Policy and Institutional Assessment (CPIA). 6 Here is not the place for a more detailed discussion of the Bank-Find Debt Sustainability Framework. Since 2005 it has been extensively discussed with stakeholders (among whom there are some critics) and has been adjusted along the way to reflect experience and new data. The most up to d ate discussion can be found in ‘Revisiting the Debt Sustainability Framework for Low-Income Countries’, Policy Paper, World Bank and IMF, January 2012, Washington D.C. 8 continual improvement between 2006 and 2012 does not mean this improvement could not begin to be reversed. Certain cases of moderate risk are discussed further below. Figure 4 Debt Distress Risk Ratings for 37 African Low-Income Countries Source: IMF/World Bank Joint DSAs, authors’ analysis 4. Debt Ratios in Greater Detail This section elaborates on country-specific trends to shed light on two questions. First, do the averages described above hide any important underlying variation in borrowing and lending? Examples of such variation might be if resource-rich countries are borrowing faster than non-resource rich, or poorer countries borrowing at a different pace from higher-income countries. The role of new lenders such as China is also examined. Second, which countries are borrowing fastest, and could these face related problems of economic management? Natural resources, income levels, and new lenders New resource discoveries and favorable commodity prices have helped enhance repayment capacity in many African countries, and in many cases improved debt distress risk ratings. This section analyzes debt trajectories in the 14 resource-rich African HIPCs that have reached completion point.7 The definition we use for resource rich is the condition that either 20 percent of exports are natural resources or 20 percent of fiscal revenues are generated from natural resources, over the period 2006-12 (c.f., IMF, 2012). We consider HIPC completion point as “time zero� and compare median public debt-to-GDP ratios between resource rich countries and others. Across both classes, post-MDRI debt ratios remain stable, suggesting that external borrowing has not primarily been driven by resource discoveries or high commodity prices. 7 Cameroon, The Republic of Congo, The Democratic Repub lic of Congo, Côte d’Ivoire, Ghana, Guinea, Liberia, Mali, Mauritania, Niger, Sierra Leone, Tanzania, Mozambique, and Zambia. 9 Similarly, new borrowing has not been appreciably different according to income level. Using a cut-off of per capita GDP in PPP terms of US$1,544, which is the average GDP per capita in purchasing power parity terms for this sample of countries, there is no significant difference between the paths of public debt ratios in poorer versus richer African economies (Figure 6).8 Since debt relief, there has been a well-documented increased lending has come from new sources, offering less concessional terms than traditional official lenders such as IDA or the AfDB. Before debt relief, concessional borrowing was the main source of the rising debt levels in most African countries; whether this dominance will continue remains to be seen as countries make use of more diversified sources of finance. Several African countries have recently issued sovereign bonds or otherwise tapped international markets (Table 3). And emerging economies such as China, Brazil, and India are also playing increasingly active roles on the continent. Ghana was the first African LIC to issue a sovereign bond (in 2007). Half a decade later, eight other countries have followed, issuing a total of US$9 billion. The simple average coupon rate on 10-year bonds was 7.2 percent; on the two 7-year bonds it was 7.5 percent; the two 5-year bonds carried an average coupon of 5.5 percent. In comparison, Greece recently issued a five-year bond at 4.75 percent, despite having defaulted and restructured its debt, and after a four-year exile from international markets. Figure 5 Figure 6 Average Nominal Public Debt to GDP by Average Nominal Public Debt to GDP by Resource Endowment after Completion Point Per Capita income after Completion Point Source: Joint Bank-IMF Debt Sustainability Analyses and authors’ calculations 8 Other distinctions, such as the main language or whether countries are landlocked status, were also investigated, but are insignificant as correlates of debt ratios. 10 Table 3 Sovereign Bond Issuance in Low-Income Africa Pricing Foreign currency rating Face Value Maturity Years to Issuer Coupon Date (Standard and Poor's) US$ (in mil) Date Maturity 2014 Cote d'Ivoire B* 750 2024 5.625 10 2007 Ghana B 750 2017 8.5 10 2013 Ghana B 750 2023 7.875 10 2014 Kenya B+ 1500 2024 6.875 10 2014 Kenya B+ 500 2019 5.875 5 2013 Mozambique B 500 2020 8.5 7 2012 Nigeria BB- 500 2018 5.125 5 2013 Rwanda B 400 2023 6.875 10 2011 Senegal B+ 500 2021 8.75 10 2014 Senegal B+ 500 2024 6.25 10 2013 Tanzania NA 600 2020 6-mth Libor + 600bp 7 2012 Zambia B+ 750 2022 5.375 10 2014 Zambia B+ 1000 2024 8.625 10 Source: Standard and Poor’s; B* refers to Fitch risk rating In addition to borrowing from concessional lenders and issuing sovereign bonds, African economies have also sought finance from new bilateral sources, primarily China. Aggregating debt for ten African LICs9 that have borrowed widely in 2010-13, it remains the case that the largest shares of external public debt are owed to IDA and the AfDB (with 44 and 10 percent, respectively); non-concessional debt now accounts for close to 30 percent of the total. Africa is shifting away from multilaterals. It should be noted that this is occurring even as countries that are borrowing commercially are tending to use their entire IDA and AfDB concessional allocations. Of gross financing needs (GFN) for this sample over the four- year period, only about 30 percent is met by IDA disbursements (Figure 7). Figure 7 Annual IDA Disbursements and Commitments as Share of GFN Source: WB loan kiosk and joint WB-IMF DSAs 9 The sample comprises Cameroon, Ethiopia, Ghana, Kenya, Malawi, Mozambique, Nigeria, Senegal, Tanzania, and Uganda. Loans were converted into US dollars and converted to PV using a discount rate of 5 percent. 11 Roughly one-quarter of external public debt is non-concessional (defined as having a grant element less than 35 percent). Higher-income countries such as Cameroon and Ghana have a larger share of commercial debt than poorer countries such as Burkina Faso or Ethiopia. About half of non-concessional loans are provided by China and commercial banks combined (30 and 20 percent respectively).10 Figure 8 Distribution of Non-Concessional Loans Source: World Bank Medium Term Debt Management Strategies China’s overall lending amounted to US$4.5 billion, with an overall grant element of 30 percent, although the majority of these loans carry variable interest rates, making this concessionality calculation in future a function of the LIBOR interest rate. A rise in LIBOR would reduce the concessionality of Chinese loans. Chines lending windows and financing terms are quite diverse, with creditors including Exim-Bank of China, the Chinese Development Bank, the Government of China, and Chinese companies such as the electricity company ZTE, China International Water and Electric Corporation, Poly Technologies Inc. China, and the telecom company HUAWEI. Episodes of Borrowing Acceleration Post-Debt Relief In 2013, the unweighted mean public debt-to-GDP ratio in all 28 MDRI countries was 40 percent.11 About three quarters of this debt was external. The 18 countries that received MDRI relief over 2006-07 saw their mean external public debt-to-GDP ratio fall from about 100 percent to about 20 percent. The above external debt ratio for 2013 implies that they have added, on average, just less than 10 percentage points on average over the following six years, while domestic debt remained constant as a share of GDP (at about 12 percent). Seven countries received MDRI relief over the later period 2009-11. These countries saw their average external public debt-to-GDP fall from an astronomical 280 percent to 20 percent of GDP, a level at which it then remained until 2013; domestic debt in these countries has actually declined since debt relief. These rather comforting averages hide large variations among countries. To summarize this diversity, African countries can be broadly classified into three groups: (1) those with low levels of public debt after 10 The investment banks with the largest lending shares in this sample of countries are (in order of magnitude) Barclays, Calyon, Citibank, KBC, JP Morgan, and Société Générale. 11 The median was also 40 percent. 12 debt relief followed by only slow debt build-up; (2) those with moderately high levels of public debt after debt relief (owing to the composition of their creditor base) and slow subsequent debt build-up; and (3) those with low levels of public debt after debt relief but a more rapid debt build-up since debt relief. The first group is Benin, Burkina Faso, Burundi, Cameroon, CAR, Comoros, Republic of Congo, Côte d’Ivoire, Democratic Republic of Congo, Guinea, Liberia, Madagascar, Mali, Rwanda, Sierra Leone, and Tanzania. The second group contains three countries where debt ratios have not been brought down sufficiently by debt relief efforts to attain sustainability by most usual definitions: Mauritania, Gambia, and Guinea- Bissau. In 2013 these countries had public debt levels of approximately 90, 70, and 60 percent of GDP, respectively, after debt relief, and without rapid subsequent borrowing. Mauritania, the world’s third largest producer of iron ore, endowed with oil, and with sizable reserves of gold and copper, saw its external public debt fall from 245 percent of GDP in 2001 to 80 percent in 2008. Geopolitical relations with Kuwait have since turned passive debt at HIPC completion point into (potentially) active debt, without any clear immediate prospects for the resolution of this debt. Since then, the country has not managed to translate its resource wealth into a comfortable debt-GDP level, and its risk of debt distress remains high, even after debt relief. The Gambia and Guinea Bissau, in contrast, saw external public debt fall to about 40 and 20 percent of GDP, respectively. However, these two countries have maintained large shares of short term domestic debt, of the order of 30 percent of GDP in both cases. The third group comprises eight countries, all of which received MDRI in 2006. The list of these countries is as follows: Ghana, Malawi, Mozambique, Niger, São Tomé and Príncipe, Senegal, Uganda, and Zambia. São Tomé and Príncipe, as a small island state, is a special case and deserves separate discussion. It has accumulated public debt of 30 percent of GDP in five years, the highest rate of accumulation relative to GDP in the sample. These patterns are likely to continue given its recent decision to sign a credit line with Angola, although of course the speed that these resources are drawn upon – and thus the rate of increase of debt ratios – remains within the discretion of the country’s authorities. While the current public debt stock of 52 percent of GDP is still well below its pre-MDRI level (over 200 percent), debt build up at this pace is unsustainable. São Tomé and Príncipe is already classified as at high risk of debt distress under the joint Bank-Fund debt sustainability Framework. A central challenge for the country is therefore to find ways to finance its investment program without incurring too rapid a pace of debt accumulation in the coming years. The case of São Tomé underlines the particular challenges of macroeconomic management faced by small and relatively isolated island states, particularly in the low-income category. Many such economies have traditionally relied on large non-debt creating flows, for example from previous colonial powers or from migrants’ remittances, to sustain high fiscal and external deficits and the related levels of public spending required to maintain both living standards and integration with the world economy. In the other seven countries, after falling to 26 percent of GDP on average (unweighted mean) from a pre- MDRI average level of 75 percent, average external public debt-to-GDP ratio had rebounded to 46 percent of GDP by 2013 (Figure 9). 13 Figure 9 Debt Build Up in Seven African Low-Income Countries Source: Joint World Bank-IMF DSAs This rate of accumulation of external indebtedness is not sustainable. If this trend were to continue, public debt in these counties would be 86 percent of GDP by 2025 (unweighted mean). These countries could thus be back to pre-MDRI debt stock ratios within a decade. Moreover, these solvency concerns are amplified by possible liquidity issues inherent in the new patterns of borrowing we are seeing. The increasing use of non-concessional financing sources differentiates today’s borrowing from that before debt relief. Debt repayment profiles are therefore more challenging. These concerns are best understood by looking at individual cases. Zambia issued its second Eurobond in early 2014, bringing external public and publicly guaranteed debt to 26 percent of GDP, from only 8 percent in 2006. Beyond this increase of 18 percent of GDP, its portfolio composition is now characterized by a sizable share of commercial debt, and repayment concentrations in 2016, 2022, and 2024. In Niger, the government borrowed in part on commercial terms to finance a large uranium mine, and also issued guarantees for a refinery loan. These transactions raised the stock of debt by one-third, increasing debt vulnerabilities despite potentially large growth dividends. The more detailed technical assistance and capacity building developed, inter alia, by the World Bank and the IMF in recent years, for example on medium term debt management strategy, offers a more detailed understanding of the challenges faced by some of these seven. In general, more attention needs to be given to fiscal costs and risks as new financing becomes more diverse. Not only are commercial debt products costlier and riskier than traditional concessional loans, they are also more varied in structure and, at times, harder to model. Debt management advisory work in Senegal and Ghana, to pick two instances from the above seven, shows that future rollover and re-fixing costs and risks are increasing significantly. In Senegal, risks emanate from the confluence of domestic debt flows with short maturities and external non-concessional debt maturity profiles with maturity bunching due to bullet repayments (Figure 10). As LIC domestic debt markets deepen, yet remain mainly focused on short-term instruments, such concerns are also likely to occur elsewhere. Of course, under scenarios of smooth sailing including the appropriate 14 adjustment of fiscal needs, these challenges can be overcome by an orderly issuance strategy by the debt management office. But under more challenging scenarios, for example shocks to the external environment that might reduce access to finance, rollover risks may materialize and translate into undesirably sudden adjustment to fiscal policy borne of necessity rather than planning. This is never good for development outcomes. Figure 10 Senegal Domestic and External Sovereign Bond Debt Spikes (from 2012 Existing Debt Profile) Source: World Bank medium term debt management strategies The case of Ghana is still more salutary. In Ghana, debt relief combined with the anticipation of oil revenues (oil came on line in 2011) and hard-fought elections in 2009 and 2013 have resulted in a considerable relaxation of fiscal policy, leading to double-digit fiscal and external deficits in 2012 and 2013. Much of the additional spending has come in the form of a reform to civil-service pay, a spending category that is typically hard to rein in once increases have been put in place (and that in addition does nothing to mitigate potential Dutch Disease effects from the dollar inflows associated with oil). As a consequence, public debt to GDP is already touching pre-debt relief levels of about 60 percent. A successful Eurobond placement in 2013 has effectively been used to finance current spending, a strategy that clearly cannot be sustained in the long or even the medium term. With limited access to external borrowing and short-term domestic interest rates well above 20 percent, financing options for the government budget are diminishing. Having been the first country to reach completion point under the enhanced HIPC Initiative and in the first group to receive MDRI, Ghana now faces a significant macroeconomic adjustment to correct its post-debt relief trajectory. 5. Conclusions for International Debt Policy After analyzing public debt developments since debt relief, several conclusions emerge. First, the evolution of public and publicly guaranteed debt ratios among HIPCs is a picture of overall moderation, with a small number of cases that underline the dangers of letting macroeconomic management drift, even if only for a few years. Debt relief has delivered some valuable results, both in terms of the reduction in public debt levels enhancing fiscal space, as well as in terms of the heightened attention to solid macroeconomic management that has in general accompanied this reduction. 15 Second, contrary to popular belief, short term domestic debt has in general not filled the borrowing space created by debt relief. Third, the cases of more rapid debt accumulation highlight new challenges. Managing public debt, including that created through borrowing by SOEs, has become even more of a priority, as borrowing on commercial terms has in some instances led to external debt payment spikes, which can combine with domestic debt obligations to cause refinancing stresses or abrupt adjustments in public spending. Either is bad for development. Fourth, although Africa’s initial external bond issuances after debt relief have so far been a success story, driven by countries’ low debt levels and attractive spreads versus historically low interest rates in high- income countries, as public debt levels creep up, investors may reassess when debt obligations are due to be rolled over. Refinancing in international bond markets may become more challenging. Finally, LIC economies continue to be small and prone to shocks. Price falls of key commodities will continue to challenge economic management and fiscal buffers have been eroded by the necessary response to the global recession that started in 2008. Errors in macroeconomic management, whether unsustainable increases in current spending or poorly planned booms in public investments, can therefore quickly cause systemic macroeconomic problems that take years to correct. 16 References: Clements, Benedict, Bhattacharya, Rina, and Nguyen Toan Quoc. 2004. “External debt, public investment, and growth in low income countries.� in Sanjeev Gupta (ed.), Helping Countries Develop: the Role of Fiscal Policy. International Monetary Fund, Washington, DC. Cohen, Daniel. 2001. “The HIPC initiative: True and false promises.� International Finance, Vol. 4, no. 3, pp. 363-380. Easterly William. 2002. “How did heavily indebted poor countries become heavily indebted? Reviewing two decades of debt relief,� World Development, Vol. 30, No. 10, pp. 1677-1696. Romero-Barrutieta, Alma Lucia, Ales Bulir, and Jose Daniel Rodriguez-Delgado. 2011. “The dynamic implications of debt relief for low-income countries.� IMF Working Paper 11/157. Giugale, Marcelo M. and Mark Thomas. 2014. “African Debt and Debt Relief.� forthcoming in the Oxford Handbook of African Economies. World Bank. 2001. Financial Impact of the HIPC Initiative: First 23 country cases. Washington, DC: World Bank. 17