41143 IDA15 THE ROLE OF IDA IN ENSURING DEBT SUSTAINABILITY: A PROGRESS REPORT International Development Association Resource Mobilization Department (FRM) September 2007 Abbreviations and Acronyms AfDB African Development Bank AfDF African Development Fund AsDB Asian Development Bank CP Completion Point CRS Creditor Reporting System DeMPA Debt Management Performance Assessment Tool DSA Debt Sustainability Analysis DSF Debt Sustainability Framework DRS Debtor Reporting System ECG OECD Working Group on Export Credits and Guarantees FRM Financial Resource Mobilization Department FY Fiscal Year GDP Gross Domestic Product GNI Gross National Income HIPC Heavily Indebted Poor Country IBRD International Bank for Reconstruction and Development IDA International Development Association ISN Interim Strategy Note IMF International Monetary Fund LIC Low Income Country LICUS L o w Income Country Under Stress MDB Multilateral Development Bank MDGs Millennium Development Goals MDRI Multilateral Debt Relief Initiative MVA Modified Volume Approach NCBP Non-Concessional Borrowing Policy NPV N e t Present Value OP Operational Policy PREM Poverty Reduction and Economic Management Department TABLE OF CONTENTS Executive Summary .................................................................................................................... i . I Introduction ................................................................................................................... 1 I1. Progress in Implementing the Debt Sustainability Framework .................................... 3 Features o f Recent DSAs ...................................................................................... 4 I11. Incorporating Debt Distress Risk Ratings into IDA'SGrant Allocation System ..........5 A n Assessment ofthe Grant Level ................................................................. 11 The Impact o f IDA Grants on the PBA system .............................................. 13 I V. Implementing the Non-Concessional Borrowing Policy in Grant-Eligible and Post-MDRI Countries........................................................................................... 16 Update on Creditor Outreach and the DSF ........................................................ 18 Update on Borrower Country Cases and Debtor Reporting ............................ 21 V. Capacity Building in LICs to Manage New Borrowing .............................................. 24 V I. Conclusions and Issues for Discussion ....................................................................... 26 Tables: Table 1 Performance category in FY07-8 based on a single-year CPIA vs a three-year CPIA average: selected countries........................................................ 11 Figures: Figure l a : The number o f traffic lights by source .................................................................... 7 Figure l b : The number o f countries by traffic light ................................................................. 7 Figure 2: Burkina Faso: the NPV o f debt-to-exports ratio. with and without MDRI.............8 Figure 3: Sensitivity o f the NPV o f debt-to-exports ratio post-MDRI................................... 8 Figure 4: Distribution o f grants and credits by period ......................................................... 12 Figure 5: Weighted average CPR by traffic light group in FY08 ........................................ 13 Figure 6: Relationship between IDA CPR and resource transfers ....................................... 15 Boxes Box 1: The Impact o f IDA Grants on Debt Dynamics ..................................................... 14 Box 2: The Bank's Decision-Making Process in a Case o f Non-Concessional Borrowing ............................................................................................................. 17 Box 3: M O U on Cooperation Export-Import Bank o f China and IBRD and IDA ...........20 Annexes: Annex 1: Traffic lights for FY08 IDA allocation: IDA-only countries ................................ 27 Annex 2: Movements in Traffic Lights in IDA14 ................................................................ 29 Annex 3: Grant Eligible and Post-MDRI countries to which the Non-Concessional Borrowing Policy applies ...................................................................................... 32 Annex 4 Principles that would guide exceptions to non-concessional borrowing ceilings .33 EXECUTIVE SUMMARY 1. Following the implementation o f the HIPC Initiative and the MDRI, IDA has taken a number o f concrete steps in order to help countries maintain debt sustainability. These steps include progress in developing, implementing, and broadening understanding o f the low-income country debt sustainability framework (DSF); designing and implementing IDA'Sgrant allocation system linked to DSF risk ratings; developing and implementing IDA'SNon-Concessional Borrowing Policy; and, developing tools and providing technical assistance aimed at strengthening borrowers' debt management capacity. 2. The joint Bank-Fund DSF launched in April 2005 has passed from i t s initial stage o f development to a mature phase o f implementation. The analytical rigor and availability o f forward looking DSAs, the "dynamic'' pillar under the DSF, has steadily improved throughout IDA14. Annual DSAs are now available for the majority o f IDA- only countries, and for many countries they have been completed for the second or third time. These DSAs help identify vulnerabilities and highlight the importance o f both growth and prudent debt management in maintaining l o w debt ratios over the long term. 3. The DSF-based IDA grant allocation system i s functioning smoothly, with a clear process established to translate DSF risk ratings into IDA "traffic lights". Grant eligibility for 50 out o f 59 potentially grant eligible countries i s now based on fonvard- looking DSAs. The assignment o f traffic lights i s subject to a clear internal clearance process, involving country teams and relevant u n i t s in the Bank, which i s working well. The relationship between DSF debt distress risk ratings and IDA traffic lights will continue to be guided by clear rules, while some element o f judgment can be incorporated for those instances where changed country circumstances are not reflected in available DSAs or in borderline cases. 4. The grant share in the I D A 1 4 envelope may reach 22 percent, or about SDR 4.7 billion out o f total financing o f SDR 21.8 billion depending on whether reactivation o f non-accrual countries i s confirmed in FY08. This i s lower than the 30 percent expected at the time o f the I D A 1 4 negotiations. Two factors explain the lower than expected grant share during I D A 1 4 to date: (i) impact o f HIPC and MDRI which the reduced the number o f "red light" countries; and (ii) smaller number o f countries a reactivating from non-accrual status than expected at the time. 5. The Non-Concessional Borrowing Policy (NCBP) approved by the Board in July 2006 i s under implementation, with steady progress being made in creditor coordination. Although this i s a challenging area, IDA i s well placed to lead creditor coordination and introduce borrower disincentives for unwarranted non-concessional borrowing: the two main prongs o f the NCBP. IDA and IMF outreach to other creditors has led to the adoption o f similar grant allocation systems by other MDBs. Progress has also been made in initiating a dialogue with OECD Export Credit Agencies and emerging market bilateral creditors, including a signed MOU for cooperation with China Eximbank. Only one case o f applying borrower disincentives has been discussed thus far - Angola - whose financing terms have been hardened. Lack o f available information and reporting .. - 11 - lags make NCBP implementation particularly challenging, but efforts are being made to address this problem. 6. A renewed emphasis in the Bank on debt management and capacity building among borrowers has become a key element o f IDA'S role in ensuring debt sustainability. This will help improve information flows, adherence to reporting requirements, and the accuracy o f DSAs. The Bank and the IMF have developed and piloted a new diagnostic measure debt management capacity; and (ii) tool to: (i) identify areas in need o f technical assistance. The first round o f implementation o f this tool i s expected in early 2008. In addition the Bank and IMF are working to produce a tool to help countries put together medium-term debt strategies. 7. Debt r elief and increased concessionality o f assistance need to be combined with policies that support private sector-led growth in order to generate sustainable trajectories for debt-burden indicators over time. IDA'Spolicy advice and direct financial assistance in support o f structural reforms and infrastructure investments has direct implications for debt sustainability, since strong growth o f GDP, revenues and exports helps lower debt ratios. While maintaining debt sustainability i s ultimately the responsibility o f debtor countries and all creditors collectively, IDA plays a key role in helping ensure debt sustainability through initiatives that help alleviate countries' debt burden as well as through i t s support to growth-promoting policies. The Role o f IDA in Ensuring Debt Sustainability: A Progress Report . IIntroduction 1. This paper responds t o the request m a d e by IDA Deputies t o assess progress in implementing the policies that IDA has put in place specifically t o help countries maintain debt sustainability. It will focus in particular o n (i) progress in developing the and implementing the l o w income country debt sustainability framework, (ii) IDA's grant ii allocation system which l i n k s IDA financing to the risk ratings arising from the DSF, ( i ) IDA's Non-Concessional Borrowing Policy and the related creditor outreach o n the DSF, and (iv) strengthening borrowers' debt management capacity. 2. O v e r the last 10 years, IDA has played an i m p o r t a n t role in helping countries achieve and maintain debt sustainability. IDA has provided debt r e l i e f and tailors the terms o f its assistance to debt sustainability, while providing technical assistance to borrowers and facilitating coordination among creditors. In addition, IDA's direct financial assistance has aimed at economic growth in client countries, a key factor in maintaining debt sustainability. 3. D e b t relief t h r o u g h the Heavily Indebted P o o r Countries (HIPC) I n i t i a t i v e and the M u l t i l a t e r a l D e b t Relief Initiative (MDRI) has helped lower debt and debt service ratios in m a n y countries, allowing t h e m t o increase poverty-reducing expenditures. For example, the average debt-service-to-exports ratio has declined on average by over 18 percent before the decision point to about 8 percent four years after the decision point.' Lower debt burdens have translated into an increase in poverty- reducing expenditures. For post-Decision Point HIPC countries, such expenditures increased o n average to 9 percent o f GDP in 2006, up from 7 percent in 2000.2 4. D e b t relief needs t o b e combined with policies t h a t support p r i v a t e sector-led g r o w t h in o r d e r t o generate sustainable trajectories f o r debt-burden indicators over time. In fact, "assuring debt sustainability depends not only upon the absolute level o f debt, but also upon the successful implementation o f a comprehensive set o f policies that are expected to enhance economic growth and poverty reduction, o n assuring access to adequate concessional flows from the international community, and on sound debt Furthermore, empirical research4has indicated that movements in debt ratios have been dominated by movements in the denominator (exports and GDP). This further highlights the importance o f growth-promoting macroeconomic and structural policies for achieving debt sustainability. 5. IDA's p l a t f o r m role in supporting the country-based development m o d e l helps low-income countries unleash private sector-led g r o w t h and attain debt IDA and IMF (2007). "Heavily Indebted Poor Countries (HIPC) Initiative and Multilateral Debt Relief Initiative (MDRI) - Status o f Implementation". IDNSecM2007-0540, p. 9. ' 3 Ibid., p. 7. World Bank, "The Challenge o f Maintaining Long-term External Debt Sustainability", April 20,200 1. 4 World Bank (2004). "Is Sovereign Debt Helping Development?", unpublished manuscript, October. ~ustainability.~ IDA's policy advice and financial assistance helps countries improve their macroeconomic performance and strengthen the micro-macro linkages in the development process. IDA's support for investment climate reforms, structural reforms, and infrastructure investments all have direct implications for debt sustainability, since strong growth o f GDP, revenues and exports help lower debt ratios. 6. Besides supporting sustainable development in client countries, I D A has taken a number o f concrete steps in order to help countries address debt sustainability challenges. A key step was the development o f the joint Bank-Fund L o w Income Countries Debt Sustainability Framework (DSF) based o n research by the Bank and the IMF into the determinants o f debt distress risk. The DSF provides a fonvard- looking assessment o f the risk o f debt distress using a baseline scenario, a standard set o f stress tests, and where necessary other alternative scenarios that take into account country-specific vulnerabilities. The framework provides insight into the risks involved in new borrowing and lending, guiding both borrowers and creditors. 7. The IDA14 Replenishmentintroduced a grant allocation framework for I D A based on the DSF. The risk rating determined under the DSF i s translated into a "traffic light" for each IDA-only country. L o w risk ("green light") countries receive regular IDA credits on highly concessional terms. Moderate risk ("yellow light") countries receive a mixture o f IDA credits and grants, while countries at high risk or in debt distress ("red light") receive only grants. This grant system provides a comprehensive rationale for grants based o n debt distress risk, and moves away from a multi-purpose and exogenously determined level o f grants as under IDA13. 8. F o r most low-income countries I D A i s only one o f many financing partners, hence an important focus o f IDA has been on enhanced creditor coordination. Debt sustainability can only be achieved if other creditors take debt sustainability concerns into account in their own lending decisions. As a result a key program o f outreach on the DSF i s ongoing in collaboration with the IMF. 9. I n addition, I D A has stepped up efforts to increase countries' capacity to manage their debts, and i s providing the tools for them to make informed borrowing decisions. The Bank has developed a Debt Management Performance Assessment Tool, piloted in five low-income countries and i s working in collaboration with the IMF to develop debt management capacity in low-income countries, with the f i r s t round o f implementation planned for early 2008. A practitioners' program i s also being considered through which government officials could gain experience working with Bank staff o n debt management issues, creating a two-way flow o f knowledge and learning o n debt management issues in l o w income countries. 10. The paper i s organized a follows: Section I1will focus o n progress in the s implementation o f the low-income country debt sustainability framework (DSF) as the analytical basis for helping borrowers and creditors assess the appropriateness o f their 5 World Bank, "The Role o f IDA in the Global Aid Architecture: Supporting the Country-Based Model", June 2007. -5- borrowing and lending strategies. Section I11will review the implementation o f the IDA14 grant system which uses the DSF risk ratings as a key input for decisions on financing terms. Section I V assesses implementation o f the non-concessional borrowing policy and creditor outreach in the DSF. Section V looks at the role for borrowers, and how IDA assistance with debt management and medium term debt strategies can help inform borrowing decisions. Section V I presents issues for discussion. 1 1 . Progress in Implementing the Debt Sustainability Framework 11. Since its endorsement by the Executive Boards o f the Fund and the Bank in April 2005, the DSF has become a key input to country borrowing strategies and to the lending decisions o f many official creditors. The November 2006 DSF review examined how the framework could help assess the scope for debt accumulation, particularly given the apparent borrowing space created by debt r e l i e f under the MDRI. The review responded to Bank and IMF Boards' request to enhance the rigor and quality o f DSAs and the effectiveness o f the DSF.7 The review emphasized the potential for the DSF to detect emerging vulnerabilities and help prevent the re-emergence o f debt distress, if broadly applied by creditors and borrowers. 12. As o f end-June 2007, the vast majority o f IDA-only countries had been assessed under the joint Bank-Fund DSF for low-income countries. O f 59 potentially grant-eligible IDA-only countries in FY08,50 had had recent joint Bank-Fund DSAs, o f which 42 were conducted in FY07.' O f the remaining nine, a further four were pending and the remaining five still had inadequate data to perform DSAs. In FY08, the vast majority o f IDA-only borrowers will therefore be assessed for the second, third, or even fourth time under the DSF. O f the 50 DSA-based risk ratings mentioned above, 1 (305 percent) were assessed at moderate risk. The highest frequency (21, or 42 percent) was high risk (or in debt distress); low risk ratings made up the remaining 14 (28 percent). 13. A guidance note for the staff o f the Bank and the IMF has been produced based on the November 2006 review and the quality and rigor o f DSAs i s gradually improving. The main improvements entail more systematic analysis o f total public debt (including domestic debt and the debts o f state-owned enterprises); more stringent checks o f the realism o f the macroeconomic assumptions underpinning DSAs; explicit guidance on the appropriate pace o f new borrowing in countries where debt burden ratios are well below the DSF thresholds (as i s the case in MDRI recipients in particular); and 6 World Bank and IMF (2006), "Applying the Debt Sustainability Framework for Low-Income Countries Post Debt Relief', November. In particular the Fund and Bank Boards called on staffs to enhance the rigor and quality o f DSAs and the effectiveness o f the DSF through: (i) strengthening the application o f the DSF b y using i t s built-in precautionary aspects; (ii) designing realistic baseline macroeconomic and growth scenarios; (iii) integrating domestic debt more systematically into the assessment o f debt sustainability; and (iv) introducing additional vulnerability indicators in cases where debt to private external creditors i s significant. The 59 countries excludes those IDA-only countries ineligible for grants in IDA14 as a result o f gap status (Angola, Georgia, Honduras and Sri Lanka), and Nigeria, due to its blend status at the beginning o f the IDA14 period. It also excludes Timor Leste which i s potentially grant eligible, but i t s grant eligibility i s based on its special status rather than on debt distress risk. -4- greater use o f precautionary features o f the DSF, such as retrospective analysis o f departures from the assumptions o f previous DSAs and special scrutiny o f projections o f high growth dividends associated with ambitious borrowing plans. Features o f Recent DSAs 14. Debt sustainability analysis of total public debt, incorporating domestic debt, has been included in all DSAs for which information on domestic debt was available. In most cases, incorporation o f domestic debt did not change the overall message o f the external debt sustainability analysis, either because domestic debt levels were l o w or because fiscal needs were projected to be largely covered by external finance. However, in certain cases (e.g., Bangladesh, Kenya, Moldova, Senegal, and Vietnam), where domestic debt i s a large share o f GDP, or where domestic contingent liabilities were potentially significant, incorporation o f domestic debt gives a clearer picture o f the government's overall debt sustainability challenges.' This should in turn inform the country's medium-term debt management strategy, guiding the appropriate mix between external and domestic debt. 15. Recent DSAs have also placed greater scrutiny on the realism o f macroeconomic assumptions. An example i s the Mongolia DSA, which details the factors that underpin the real GDP growth assumptions in each year. For instance, changes to the GDP profile in Mongolia were explained by developments in the Oyu Tolgoi Mine, with growth highest when the main mine's production was due to start, and declining to more modest rates thereafter. Other cases where particularly detailed accounts o f the assumptions driving growth projections can be found include Ghana, Ethiopia, and Vietnam, with a particular focus o n primary resource sectors described in detail in (for example) Angola, the DRC, Sa0 Tom6 and Principe, Sudan, and Yemen. 16. A number o f DSAs have included tailored country-specific scenarios to assess risk, examples include Mauritania, Ghana, Republic o f Congo, and Guyana. 0 In the Mauritania D S A (November o f 2006), a country-tailored alternative scenario featured lower o i l production than in the baseline. This scenario highlighted Mauritania's key vulnerability, and under this scenario Mauritania's debt service to exports ratio would only stay below the relevant 15 percent threshold if Mauritania could finance the substitution o f o i l revenue with concessional borrowing. lo 0 In Ghana three alternative scenarios were investigated: one looked at lower growth o f exports and GDP than expected from structural reforms and infrastructure investment, one assessed the impact o f higher non-concessional borrowing than in the baseline framework, and one assessed a combination o f these two scenarios. In any o f these scenarios the risk o f debt distress would DSAs are required to explicitly flag situations where public domestic debt may lead to a different risk classification from consideration o f external debt indicators alone. See World Bank and I M F (2006), op.cit., p.23. lo See World Bank, "Mauritania: Joint Bank-Fund Debt Sustainability Analysis", February, 2007. -5- be notably higher than in the baseline. These highlight the main risks for debt sustainability coming from the relatively undiversified exports base, and the need for cautious new borrowing and wise use o f borrowed resources." 0 Inthe Republic o f Congo, a non-adjustment fiscal scenario, which assumes that the authorities will continue with current fiscal policies, shows a strikingly different debt sustainability outlook than the baseline, primarily due to the high non-oil deficit, which would lead to unsustainable levels o f new borrowing. l2 0 In Guyana, three additional stress tests were undertaken. These included only partial realization o f the benefits o f sugar-sector restructuring, lower future gold prices than projected, and weaker growth as a result o f shocks.I3 17. Recent DSAs have also provided comparisonswith previous DSAs. In the case o f Burkina Faso, the more favorable macroeconomic outlook in the 2007 DSA relative to the previous DSA was justified on the basis o f unexpectedly good export performance in the most recent years. The overall moderate risk rating in the DSA was not changed however, as the DSA also highlighted the impacts o f scaling up aid o n debt distress risks.14 These scenarios emphasized the need for increased aid to be mostly o n grant terms to ensure debt sustainability. Other cases that give more detailed attention to the comparison with previous DSAs are Guyana and Vietnam. 18. The DSF framework has thus passed from its initial stages o f development to a more mature phase o f implementation and incrementalimprovement. The availability o f successive forward-looking DSAs for most LICs now provides an important tool for borrowers in designing not only their debt management strategies but also their broader macroeconomic policies. The annual DSAs also provide useful guidance to other creditors, as these DSAs are now publicly available o n the World Bank and IMF websites for countries who have provided consent to the DSA's disclosure. The effectiveness o f the DSF i s ultimately dependent upon its broad use by both borrowers and creditors. 1. 1 1 IncorporatingDebt Distress Risk Ratings into IDA'SGrant Allocation System 19. The grant allocation system in IDA14 i s based on external debt distress risk ratings, and represents the main tool for I D A to help ensure debt sustainability. Under the grant allocation system in IDA, a country's external debt distress risk rating i s translated into a "traffic light" which in turn determines the mix o f credits and grants for 11 See IMF, "Ghana: 2007 Article I V Consultation", June 20,2007. 12 See IMF, "Republic o f Congo: 2007 Article I V Consultation", June, 2007. 13 Guyana has not yet agreed to publish its 2007 DSA. 14 The scaling-up scenarios used conservative estimates taken from analytical work of the potential impact o f scaled up aid on GDP growth. See Burkina Faso DSA for more on this. The DSA's estimates o f the growth impact o f the scaled up aid were derived from the specifications in "Counting Chickens When They Hatch: The Short Term Effect o f Aid on Growth" by Michael Clemens, Steven Radelet, and Rikhil Bhavani (2004). -6- the country. The traffic lights comprise three categories: green corresponds to a low risk o f debt distress; yellow to a moderate risk o f debt distress; and red to a high risk o f debt distress or actually being judged to be in debt distress. A traffic light i s assigned annually to each IDA-only country. The configuration o f credits and grants by traffic light i s then given as follows: a green light results in an allocation o f 100 percent IDA credits; yellow light results in an allocation o f 50 percent IDA credits and 50 percent grants; and red light results in an allocation o f 100 percent grants. 20. The grant allocation system was built on two pillars: a "static" pillar - based on a one-year "snapshot" o f debt indicators relative to the DSF matrix o f policy- dependent external debt burden thresholds; and a "dynamic" pillar - an assessment o f debt distress risk based on a forward-looking DSA that assesses debt trajectories relative to the matrix o f policy-dependent external debt burden thresholds, and the impact o f alternative scenarios and stress tests on these debt trajectories. In FY06-7, debt distress risk ratings were primarily based on the `static' pillar, i.e. a comparison between i t s latest available relevant external debt indicators" and the applicable external debt thresholds.16 This owed to limited availability o f DSAs at an early stage o f DSF implementation. 21. The availability o f DSAs has increased significantly over the IDA14 period, consequently the vast ma'ority o f traffic lights were determined on the basis o f the dynamic pillar in FY08.I? DSAs have increasingly been used as the basis for traffic light ratings throughout IDA14, from 4 in FY06 to 50 in FY08 (see Figure 1.a). At the time o f determining the FY08 traffic lights, DSAs had yet to be completed for only nine countries (mostly HIPC pre-decision point countries) and hence their traffic lights are based on the `static' pillar: namely Bhutan, Democratic Republic o f Congo, Cote d'Ivoire, Kiribati, Maldives, Myanmar, Somalia, Sudan, and Vanuatu. l8 22. The number o f grant-eligible countries (red plus yellow lights) has been stable over the three-year period o f IDA14. Annex 2 documents the changes in traffic lights during IDA14. For FY06,42 countries were eligible for grants: 33 red light countries, nine yellow light c o ~ n t r i e s . ' For FY07, the number o f grant-eligible countries ~ 15 Although there are five debt indicators under the DSF, the indicators which are relevant for the purposes o f determining grant eligibility under the static pillar are: (i) ratio of the net present value the (NPV) of public and publicly guaranteed (PPG) external debt-to- Gross Domestic Product (GDP); (ii) the ratio of NPV o f PPG external debt-to-exports; and (iii) ratio of the service on PPG external debt the to exports. For more on the underlying matrix o f debt indicators see IMF and World Bank (2004a), "Debt Sustainability in Low-Income Countries - Proposal for an Operational Framework and Policy Implications", February. 16 For more on the static pillar see IDA (2004), "Debt Sustainability and Financing Terms in IDA14: Further Considerations on Issues and Options", November. 17 This was endorsed by I D A Deputies at the IDA14 Mid-Term Review meeting in November 2006. See IDA (2006a), "Debt Dynamics and Financing Terms: A Forward-LookingApproach to IDA Grant Eligibility," presented at the IDA14 Mid-Term Review Meeting, November. 18 See Annex 1. For FY08, an unpublished DSA for Eritrea confirms the red light rating obtained with the static pillar. For Kiribati, neither a DSA nor debt data was available and its traffic light was assumed to be green. 19 For KOSOVO, eligibility was based on its status as part of Serbia (previously Serbia and Montenegro) under the United Nations Administration and its eligibility will continue during IDA14 until its status -7- decreased to 39 (30 red light countries, 9 yellow light countries), reflecting the lower incidence o f red light countries, from 33 to 30. In FY08, the number o f grant-eligible countries was 41: 28 red lights and 13 yellow lights (see Figure 1.b). Figure l a : The number of traffic lights by Figure l b : The number o f countries by source traffic light FY06 FY07 FY08 -06 W07 -0s a) N u d e r of traffic lights based on the 'static' pillar ,b) N u d e r of traffic lights based on the 'dynamic' pillar (a) Green (b) Yellow ( c ) Red 23. T h e fact that the number of grant eligible countries has been stable in IDA14 even after debt relief under HIPC and M D R I highlights the importance of increasing economic resilience to help maintain the reduced debt ratios achieved by debt relief. Debt r e l i e f provided under the HIPC Initiative and then MDRI during IDA14 reduced debt ratios for most post-completion point HIPCs, but debt relief "does not automatically translate into credits-only status for i t s recipient^".^' While the static pillar would indicate many more green light countries, the forward-looking DSAs are better predictors o f debt distress. As a result, under the dynamic pillar, o f the 20 post- MDRI countries which could qualify for IDA grants, nine are green light countries in FY08. For several post-MDRI countries, unless improvements in macroeconomic fundamentals are made, debt ratios would converge to pre-MDRI levels in the long t e r m (see Figure 2 showing the impact o f MDRI on the long run trajectory o f the NPV o f debt- to-exports ratio o f Burkina Faso).21 For many countries, baseline indicators remain below their policy-dependent debt thresholds. However, stress tests in the DSF highlight sensitivity to non-concessional external borrowing or exogenous shocks, such as a sharp decline in exports (see Figure 3a and b and Annex 2). 24. I n essence, the grant system based on forward-looking SAs incorporates future prosperity or vulnerabilities into current financing decisions. For instance, in FY06, while the static pillar suggested a red light for Mauritania and a green light for Ethiopia, Rwanda and Zambia, their forward-looking DSAs indicated a l o w risk o f debt distress for Mauritania given i t s strong o i l prospects, a moderate risk for Ethiopia and changes and it i s able to borrow. Timor-Leste was eligible for grants on the basis o f i t s post-conflict status during the IDA14 period (IDA, 2005a). 2o IDA (2006a), 0p.cit. 21 This also reflects debt mechanics: if new borrowing i s kept unchangedthen the weight of existing debt in total debt tends to decrease over time. -8- Zambia, and a high risk for Rwanda given the expected upward trend in the simulated trajectory of their debt ratios, even after reaching the HIPC completion point. The dynamic pillar i s able to assess future risks by looking at the likely trajectory of new borrowing and expected movements in the main economic variables. It also assesses the impact o f stress tests calibrated to the historical economic performance o f GDP growth, exports, and other relevant factors. Switching to the dynamic pillar o f the D S A therefore enables financing terms to reflect debt distress risks over the medium term, and via annual updates, enables IDA financing terms to react to changes in underlying economic conditions. Figure 2. Burkina Faso: NPV o f debt-to-exports ratio, with and without MDRI Staff ' p. 32, Figure 3. Sensitivity o f the NPV o f debt-to-exports ratio post-MDRI (a) Sensitivity to financing terms: Benin 3m ?50 ZW 1% IDO 50 tl Source: IMF (2006), "Appendix 1V: Debt Sustainability Analysis," p. 72, Figure 2. in "Benin: Article I V Consultation, First Review Under the Three-Year Arrangement Under the Poverty Reduction and Growth Facility, and Request for Waiver o f Nonobservance o f Performance Criteria," November 10. -9- (b) Sensitivity to export growth: Ethiopia Note: Most extreme stress test refers to a shock in export growth. Source: IMF and IDA (2007), "The Federal Democratic Republic o f Ethiopia: The Joint IMF/World Bank Debt Sustainability Analysis 2007," p.5. Figure 1. May 22. 25. Like the CPIA, the assignment o f traffic light ratings i s subject to a Bank- wide clearance process, and subsequently traffic lights are published o n IDA'S website. Based on the risk ratings from the most recent DSA, regions are sent a l i s t o f the proposed traffic lights. There may b e exceptional cases where the country team or regional management disagrees with the traffic light. In each case, the determination o f the appropriate traffic light i s approved by the relevant units in the Bank (country team, PRMED, and FRM). This process i s functioning well and has been in place for all traffic light determinations in IDA14. The traffic lights for all IDA-only countries are then published annually in OP 3.10 Annex D in terms o f grant percentage, which i s available o n the Bank website.22 26. There are at least two types o f situations that may call for judgment at the time o f the grant allocation process and lead to differences between published risk ratings and traffic lights. The first situation involves cases where there has been a major change in country circumstances since the most recent DSA was prepared, such as eligibility for MDRI, or major revisions in projected o i l revenues. The second situation represents borderline cases or situations where the traffic light associated with the risk rating needs to be reassessed. In either o f these two circumstances where the country team feels that the risk rating may need to be revisited, this i s done according to the process described above. 27. Actual cases where traffic lights differ from final risk ratings are unusual - only two occurred in FY08 - and will continue to be monitored to see whether each judgment call i s borne out by subsequent developments. In FY07 there were more frequent differences from publishedDSAs primarily because DSAs for Benin, Mali, Mauritania, Tanzania and Uganda needed t o be updated with additional scenarios incorporating the impact o f MDRI. Based on these scenarios, risk ratings and their 22 http://~itere~ource~.worldbank.org/OPSMANUAL/Resources/OP3 lO-AnnexD-RevJul-lO-2007.pdf - 10- associated traffic lights were then approved by the relevant u n i t s in the Bank. Apart from DSAs not reflecting debt relief from MDRI, the differences between final published risk ratings and traffic lights are: In the case of Madagascar, a green light for FY07 was based on preliminary projections as part o f an ongoing joint Bank-Fund DSA that was not yet finalized.23 In the borderline case o f Burkina Faso in 2006, the relevant units in the Bank collectively agreed that the stress test in the DSA that caused a debt ratio to breach i t s threshold was extreme, and alternative scenarios led to a l o w risk and green light rating for the FY07 traffic light exercise. 0 The DSA completed in February 2007 for Siio Tom6 and Principe showed a moderate risk. Since the conclusion o f the DSA, however, the prospects for o i l revenues had declined. A revision o f the macroeconomic assumptions in the DSA suggested a high risk o f debt distress. As a result, relevant u n i t s in the Bank concluded that Silo Tom6 and Principe should be classified as a red light country for the FY08 allocation exercise. Based on the DSA analysis, Ghana's risk o f debt distress i s between "low" and "moderate." The analysis underlying the DSA for Ghana i s based o n the assumption o f a 100 percent credit allocation from IDA. That assumption does not jeopardize Ghana's debt sustainability prospects going forward, given that the probability o f a shock that would produce debt distress i s very low. Therefore, the relevant u n i t s in the Bank jointly agreed that retaining ''green light" status (100 percent o f IDA assistance in credit terms) during FY08 would be an appropriate course o f action from the point o f view o f Ghana's debt sustainability prospects. 28. I t i s too early to assess whether basing the debt thresholds on performance using the three-year moving average o f CPIA has had an impact on the volatility in traffic light ratings. In the mid-term review paper, simulations indicated that this change, introduced for FY08, would reduce volatility. In the first year o f implementation o f this new methodology for determiningthe performance benchmark, country teams were given the flexibility to maintain a 1-year performance rating if the use o f the 3-year average rating would increase ~ o l a t i l i t y .Table 1 shows performance rating movements ~~ in selected countries and indicates that a three-year average helps maintain the same performance categories in FY08 as in F Y 0 7 in some countries. At the same time, it suggests that the three-year average does not completely eliminate volatility in the performance category (see Burkina Faso in Table 1). Annex 1 shows the traffic lights for each country and the basis o f the performance rating in each country in FY08. 23 Once completed, the published 2006 DSA indicated moderate risk. A subsequent DSA, published in July 2007, upgraded the debt distress risk rating to low and confirmed the green-light status for FY08. 24 See World Bank, "Staff Guidance Note on the application o f the Bank-Fund Debt Sustainability Framework for Low-Income Countries", May 3,2007. - 11 - Table 1. Performance category in FY07-8 based on a single-year CPIA vs a three-year CPIA average: selected countries. FY07 FY08 Country based on 2005 CPIA based on 2006 CPIA Based on 2004-6 average Burkina Faso Strong Medium Medium Maldives Strong Medium Strong Mauritania Weak Medium Weak Nicaragua Medium Strong Medium Vietnam Medium Strong Medium Yemen, Republic o f Medium Weak Medium Source: World Bank staff calculations. 29. The move towards the dynamic pillar in the IDA14 grant allocation system i s thus progressing as expected and no changes are proposed. This forward-looking traffic light system based o n the risk o f debt distress i s proactive and enables IDA to tailor i t s financing terms in anticipation o f future risks. At the same time regular updates o f DSAs enable IDA'Sfinancing terms to adjust in response to actual movements in economic variables and the pace o f borrowing. Risk ratings are receiving greater scrutiny in cases where DSAs indicate a low-moderate or a high-moderate rating to avoid overly conservative or mechanistic risk ratings. The flexibility employed when translating DSF risk ratings into traffic lights will continue to be based on the due process described above and carefully monitored. An Assessment o f the Grant Level 30. Grant financing in IDA14 may reach up to SDR 4.7 billion, o r about 22 percent o f the total SDR 24.8 billion IDA14 envelope (see Figure 4). The level o f grants has been steady throughout IDA14, but i s lower than the 30 percent projected at the time o f the I D A 1 4 replenishment negotiation^.^^ In FY06 commitments in the form o f grants amounted to about SDRl.3 billion out o f a total volume o f credits and grants o f SDR 6.5 billion (US$9.5 billion). In FY07 eligible IDA countries committed grants o f SDR 1.5 billion out o f a total IDA commitment o f SDR 7.9 billion (US$ 11.8 billion). Grant usage was slightly lower than projected in FY07, while the total commitment was higher than the total allocation (SDR 7.7 billion). In FY08 grants are expected to be 19 percent o f total IDA financing, and 27 percent if reactivation takes place for countries in arrears. *' World Bank, "Additions to IDA Resources: Fourteenth Replenishment. Working Together to Achieve the Millennium Development goals", March 10,2005. - 12- Figure 4. Distribution o f grants and credits by period (SDR billion) Grants based on debt sustainability FY06 FY07 FY08 3 1. Two key factors explain the lower grant share to date than estimated at the time o f IDA14. First, although HIPC and MDRI did not automatically lead to green light status for all countries, there are fewer red light countries (of the 20 post-completion point HIPCs which could qualify for IDA grants, one i s s t i l l rated as red light and 10 are rated as yellow light).26 Second, the reactivation o f countries in arrears - for which grants amounting to about SDR 1 billion were estimated to be allocated - has not yet materialized, lowering the grant share to date. Irrespective o f the grant share, however, additional grant flows have been delivered through HIPC and MDRI - the equivalent o f budget support in the form o f grants.27 32. As outlined in the IDA14 Deputies report, there are two transitional cases which are exceptions to the debt-distress-based grant eligibility criterion - Kosovo and Timor-Leste. For IDA14, Kosovo was eligible for grants based on i t s status as part o f Serbia (and previously Serbia and Montenegro) under United Nations Administration. In IDA15, as long as KOSOVO'S remains as it was in IDA14, it i s proposed that status KOSOVO'S eligibility for grants would continue. Timor-Leste was grant-eligible in IDA 14 to avoid a "sudden shift o f status vis-a-vis other IDA13 grant-recipient countries under the post- conflict criterion".28 Management proposes that Timor-Leste continues to have access to the grants equivalent o f i t s P B A allocation @e. 60 percent), given i t s continued post-conflict status and constraints on i t s ability to contract external debt. 33. The impact o f IDA'S grant allocation system on debt sustainability i s contingent on a number o f factors, including whether borrowers and other creditors will incorporate debt sustainability considerations into their own financing decisions. The expectation i s that where IDA i s a key creditor, the debt r e l i e f already 26 There are 22 post-completionpoint HIPCs, but Bolivia and Honduras are not eligible for IDA grants as Bolivia i s a blend country and Honduras i s a gap country. 27 See World Bank, "Assessing Implementation o f the IDA14 Grant Framework". Once MDRI r e l i e f i s counted, the grant share in IDA14 i s estimated at about 24 percent. 28 See IDA (2005), "Additions to IDA Resources: Fourteenth Replenishment"., pp.27. - 13 - delivered through HIPC and MDRI as well as grants where necessary going forward, will help countries maintain sustainable debt trajectories, and gradually allow a reduction in the need for grants or debt r e l i e f in the future. Where IDA i s not a key creditor, there will be risks in maintaining debt sustainability unless borrowers and other creditors take debt sustainability considerations into account in their borrowing and lending decisions. The impact o f IDA grants on debt dynamics i s further elaborated upon in B o x 1. The Impact o f I D A Grants on the PBA System 34. The discount on grant volumes embedded in the grant allocation system (modified volume approach, o r MVA) aims to maintain IDA's performance incentive. Under the PBA system, the main determinant for a country's IDA allocation amount i s the country's IDA performance rating (CPR).29 An improvement in performance, holding all else constant, increases the allocation amount and therefore the resource transfer amount. Any modifications, including the grant allocation system, to the P B A system must ensure that the positive relationship between performance and resource transfer remains intact. Given the negative correlation between risk o f debt distress and performance (see Figure 5), grants based o n debt sustainability can result in higher resource transfers to low-performing countries, thereby weakening the performance incentive in the PBA. This weakening i s moderated by applying the 20 percent discount to grant volumes. 1/2/3/ Figure 5. Weighted average CPR by traffic light group in FY08 I 3.76 I Red Yellow Green 11Weights are group-specific ones that are inversely proportional to the variance o f an observation in the traffic light group. 21 The number o f observations: Red=25 (excluding countries with non-accrual status and with CPR unavailable, namely Liberia, Myanmar and Somalia); Yellow=l3; Green=l8. 3/ See IDA (2006a) "Debt Dynamics and Financing Terms: A Forward-Looking Approach to IDA Grant Eligibility," for further analyses on this. 29 See IDA (2007), "IDA's Performance-Based Allocation System: Simplification o f the Formula and Other Outstanding Issues", September. Figure 1, pp. 2-3 for an explanation o f t h e CPR. -14- Box 1: The Impact o f IDA Grants on Debt Dynamics A country's debt dynamics depend on the existing debt burden, the concessionality o f new financing and i t s growth performance. The impact o f the existing debt burden on debt dynamics i s through the difference between growth rates (GDP or exports) and interest rates. The debt dynamics, a l l else constant, are stable when growth rates are the same as interest rates. A positive new financing requirement, holding all else constant, raises debt ratios. The grant element in new financing determines, all else constant, the extent to which the debt dynamics increase as a result o f new financing: i.e. new financing with 100 percent concessionality keeps the debt dynamics unchanged, while new financing with zero percent concessionality will fully impact the debt dynamics. The impact o f IDA grants on debt dynamics, in turn, depends on the share o f IDA in total new financing. As shown in Figure B.l.a, when IDA accounts for more than 80 percent o f total external financing, IDA14 grants clearly reduce the speed o f new debt accumulation (the dotted line in the figure). Continued IDA grant financing (IDA14 and 15) results in even further reducing the speed. On the other hand, in an economy with a small IDA exposure in the total external financing, the impact o f IDA grants on the debt dynamics i s minimal (Figure B.1.b). (a) Rwanda: NPV of debt-to-exuorts ratio 150 - ,1 " _, ~~~ "l"__ -(I) Baseline with IDA regula i -- (11) Baseline post-IDA14 gra 100 - -- (iilBaseline. post-IDA14 grants ----(iii)Baseline: post-IDA14 & 15 grants I 04 I , , , , I , I , , , , I I , I ! 5 , ( 1 + .@ Q" ,.o" e ,d." e`." ` ,d." 9 ` ,dL' % ++ i o ~ eIDA rcprescnts, on average 11 percent of external debt in NPV terms Even in an economy with high IDA exposure, an Figure B.2. Impact o f an adverse export shock on adverse economic shock can erode the positive impact o f IDA grants on debt dynamics. As shown in Figure B.2, an extreme negative shock 350 , the NPV o f debt-to-exports ratio (percent) ........ .................. i on exports lowers exports earnings and requires more financing needs, and therefore shifts the debt-to-exports ratio upwards, beyond 200 percent, despite IDA14 and 15 grants. Nonetheless, it i s notable that the impact o f the adverse shock on debt dynamics i s smaller in the presence o f grants. These suggest that a ' `' - -Most extreme stress test pre-IDA14 grants 1 " Most extreme stress test post-IDA14 & 15 grants country's debt sustainability calls not only for 0 50 -- (111) Baseline post-IDA14 & 15 grants grant financing but also for better economic / 1 , / , 1 1 1 1 1 1 , , / 1 1 ) , , I j management and improved resilience so as to moderate the impact o f a given shock on the economy. I Source: Staff simulation, and IDA (2006a), "Debt Dynamics and FinancingTerms: A Forward-LookingApproach to IDA Grant Eligibility," pp. 15-20, prepared for the IDA14 Mid-Term Review Meeting, Nov. 20-21. - 15- 35. O v e r a l l for IDA14 (Figure 6.d), the relationship between performance and resource transfers remains strong. Figure 6 shows the regression results o f the real resource transfer with and without grants against IDA'Scountry performance rating. Figure 6.a (baseline) shows that, as expected, without any grant in IDA allocations (i.e. credits only), IDA performance i s positively associated with resource transfers (as measured by the grant element o f IDA assistance volumes) per capita. Even after the introduction o f the I D A 1 4 grant allocation system and the associated discount o n such grants in FY08, the relationship s t i l l holds (Figure 6. b. ~ o s ~ - M V A ) . ~ ' association The becomes somewhat weaker when taking into account MDRI netting out and the reallocation o f resources netted out under MDRI (another addition to the P B A system, Figure 6.c), but remains strong overall for I D A 1 4 (Figure 6.d). 1/2/3/4/ Figure 6. Relationship between I D A CPR and resource transfers (d) IDA14 final allocation I/ IDA CPR coefficient is: (a) 2.94 (b) 2.48 (c) 2.28 (d) 5.56 (all are statistically significant at the 1 percent level). 2/ The number o f observations is 53 countries, excluding countries with population less than 1 million, those in non-accrual status and those with capped allocations. 3/ The variable FY08 IDA CPR i s expressed in logarithms, taking into account the quadratic function o f CPR in the PBA allocation formula. 41 The following grant element i s assumed: 60 percent for IDA credit; 80 percent for the mix o f IDA credit and grants; 100 percent for IDA grants. 30 See IDA (2006a), "Debt Dynamics and Financing Terms: A Forward-Looking Approach to IDA Grant Eligibility," pp. 15-20, prepared for the IDA14 Mid-Term Review Meeting, Nov. 20-21 for the relationship in FY06 and FY07. - 16- 36. Another link between debt-related issues and IDA's resource allocation system refers to the impact o f MDRI's netting out mechanism on new IDA allocations to post-completionpoint HIPCs. The netting-out mechanism was put in place to preserve the equity between MDRI beneficiaries and non-MDRI countries. However, given the upward trajectory o f debt service r e l i e f over the medium-term, many MDRI recipients would be receiving large cuts to their IDA allocations perhaps even to zero as a result o f the netting out mechanism. While MDRI netting out may not reduce the overall assistance package which consists o f new IDA plus debt service relief, it does reduce new IDA commitments thus limiting IDA's ability to play a platform role in the country. This may create an incentive for affected countries to resort to other forms o f less favorable financing i f alternative concessional resources are constrained. IV. Implementingthe Non-ConcessionalBorrowing Policy in Grant-Eligible and Post-MDFU Countries 37. In July 2006, the Executive Board of IDA approved a framework for IDA's response to non-concessionalborrowing risks in grant-eligible and post-MDRI c~untries.~~was pointed out in that paper, debt stock ratios post-MDRI in most As recipient countries will be lower than debt ratios in many middle-income countries. Debt r e l i e f and IDA grants increase the apparent borrowing space and financing options available to recipients that can be used to accelerate progress in meeting the MDGs. At the same time, this presents a challenge for borrowers to manage that borrowing space effectively to increase growth and meet poverty reduction targets while maintaining debt sustainability. 38. The primary objective o f IDA's Non-ConcessionalBorrowing Policy (NCBP) i s to address the risk that non-concessionalloans to grant-eligible and post-MDRI countries may lead to a rapid re-accumulationo f debt and thus undermine borrowers' debt sustainability prospects. Low-income countries (LICs) continue to require significant investments in order to make progress towards MDGs, and much more can be done without jeopardizing debt sustainability if debt flows are o n concessional terms. I t i s important to note that although debt r e l i e f led to l o w debt ratios post-MDRI, other broad economic circumstances are largely unchanged. In fact simulations outlined in the November 2006 DSF review paper showed that some countries could return to pre- debt r e l i e f levels o f indebtedness in 6 to 10 years unless lenders respond to higher debt distress risks with higher concessionality. 39. The focus on non-concessionalborrowing stems from the greater risks that such borrowing puts on debt sustainability. Although large volumes o f concessional financing could also increase debt distress risks, for a given borrowing path, non- concessional borrowing yields a smaller net resource f l o w and worse debt dynamics than concessional borrowing. Non-concessional lending increases debt ratios more rapidly, and in general repayments are required quickly, even though most investments require a long gestation period to yield returns. When the expected growth impact o f an 31 See IDA (2006b). "IDA Countries and Non-Concessional Debt: Dealing with the `Free Rider' Problem in IDA14 Grant Recipient and Post-MDRI Countries." IDAR2006-0 137, July. - 17- investment is not realized, non-concessional financing poses a greater risk to debt sustainability. As a result concessional finance remains the most appropriate source o f financing for LICs with large financing needs. It should be noted, however, that any borrowing, even o n concessional terms, poses risks if the expected growth dividend does not materialize. 40. T h e NCBP i s n o t a blanket restriction on non-concessional borrowing, and acknowledges that u n d e r certain circumstances non-concessional loans can appropriately b e part o f a financing m i x that helps promote economic growth. While concessional financing continues to be the most appropriate form o f financing for most LICs, there may be cases where non-concessional borrowing may warrant an exception to the policy (i.e. no disincentive mechanisms would apply in such a case). The NCBP paper describes the considerations that feed into decisions o n whether a particular instance o f non-concessional borrowing could be considered an exception to the policy (See Annex 4). This assessment i s o n the basis o f both country-specific policies and investment plans and loan-specific criteria. The Bank's process for addressing new cases o f non-concessional borrowing and determining the appropriate response i s summarized in Box 2 (below). Box 2. T h e Bank's Decision-Making Process in a Case o f Non-Concessional Borrowing Data i s received by FRM from annual reports to DEC, as w e l l as via alternative sources. Follow-ups are undertaken (via the country team, and w i t h other creditors) to verify the accuracy o f the data, and obtain as much information as possible on the projects for which non-concessional borrowing was contracted. Each year, at the time o f the IDA allocation exercise, representatives from the Regions, CFP, PREM, D E C and OPCS meet to review countries' non-concessional borrowing, if anything significant has been raised prior to the allocation exercise. For cases coming up in between the yearly IDA allocation exercise, the same group convenes as needed to discuss the response, once sufficient information i s received about the nature o f the loan. Based on principles outlined in B o x 3 o f the NCBP paper, this group makes recommendations on whether a non-concessional loan would be considered a breach o f the NCBP. The group also recommends the appropriate disincentive mechanism in response to a breach (either a cut in allocation volumes, or a hardening o f the financing terms). Based o n these factors the group makes a written recommendation to the Bank's Operations Committee for a decision on the appropriate disincentive measure, along with detailed justification for the decision. The Board is informed o f the decision to apply a disincentive mechanism ahead o f the next IDA sroiect in the affected country. 41. E v e n in countries w h e r e IDA's policies and decisions may have little i m p a c t on the decisions by authorities, other facets o f IDA's r o l e may nonetheless help shape borrowing decisions. Capacity building in debt management, providing a tool for developing medium term debt strategies, and providing advice o n managing o i l revenues are other ways to increase the likelihood that borrowing i s responsible. IDA i s also well placed to lead, together with the IMF, dialogues with other groups o f creditors. -18- 42. This section will review implementation o f the two key pillars o f the NCB policy: (a) increased incorporation o f the DSF among creditors and borrowers in their lending and borrowing decisions (including increased creditor collaboration, strengthened information flows, strengthened debt reporting and debt management capacity); and (b) discouraging non-concessional borrowing through borrower disincentives. Given that borrowers ultimately decide on which loans to contract, improved debt management and the development o f medium-term debt strategies are critical to the NCBP and are discussed in detail in section V. Update on Creditor Outreach and the DSF 43. For most low-income countries, I D A i s only one o f many financing partners, and hence creditor outreach i s vital to maintaining debt sustainability. Discussions on the L I C DSF have been held with nearly all major multilateral and bilateral creditors to low-income countries. The World Bank and the IMF have undertaken extensive outreach activities involving the Export Credit Group o f the OECD, the Berne Union (of export credit insurers), the OPEC Fund, and other Arab multilateral creditors and some emerging market donors. N o t only have these exchanges contributed to a common understandingo f the principles o f collectively sustainable lending to low-income countries, but information flows have also improved. 44. T o increase accessibility o f the DSF, a number o f steps have been taken by staffs o f the Bank and Fund to make it more user-friendly, and make individual DSAs readily available. Based on feedback received during creditor and borrower consultations, a dedicated webpage on the Bank and on the Fund's website has been established so that DSAs can be easily accessed, where countries agree to their publication. 32 To increase understanding and broaden acceptance o f the DSF, this page provides easy l i n k s to each new DSA as well as l i n k s to the various DSF policy documents and the DSF user's guide. In addition, webpages have been set up to provide guidance to creditors and borrowers on concessionality, including a concessionality calculator and relevant information on the Bank's non-concessional borrowing policy and the Fund's concessionality policy. 33 As the NCBP applies to grant-eligible and IDA-only post-MDRI countries, this l i s t changes slightly from year to year (see Annex 4 for the l i s t in FY08, which i s also posted on the external website). 45. Discussions are underway to agree on a more active role for the major regional banks34in the DSA process. These discussions are in response to the willingness expressed by these institutions to harmonize their lending practices broadly along the lines suggested by the risk assessments contained in DSAs carried out by Bank and IMF staff. This process i s close to completion, which will lead to concrete guidelines signed by the leaders o f the respective banks describing processes for the staff o f these institutions to have input into the DSA process. 32 See http://www.worldbank.orpJdebt and http://imf.org/dsa 33 This website can be accessed via the World Bank external website by searching for "non- concessional", and via the IMF external website (http://imf.org/concessionality). 34 The African Development Bank, Asian Development Bank, Inter-American Development Bank, European Bank for Reconstruction and Development, and European Investment Bank. - 19- 46. As a result o f outreach o n harmonizing lending policies with the DSF, an increasing number o f MDBs are incorporatingelements o f the DSF into their own financing terms, and some o f them are contemplating adoption o f their own policies on non-concessionalborrowing. The African Development Fund has a grant allocation system that i s almost identical to that of IDA, with a similar traffic light system and a modified volume approach.35 Another creditor, the International Fund for Agricultural Development has also incorporated the DSF into i t s decisions on grant eligibility. Similarly, the Asian Development Fund discussed a proposal at i t s Board in late July for a grant allocation system that mirrors that o f IDA14. However, available data suggest that some MDBs may s t i l l be lending on non-concessional t e r m s to grant-eligible and post-MDRI countries, and outreach with these creditors in particular i s being strengthened. 47. Progress made in collaboration among multilateral creditors on the DSF could potentially have an important impact on the debt ratios in most LICs, as multilateral creditors are bound to continue to account for the bulk o f new borrowing. Although interest in L I C financing in a post-MDRI context i s growing among other creditors, notably bilateral creditors, multilateral creditors will be the key drivers in the debt ratios in most LICs. The November 2006 review pointed out that full disbursement o f MDB allocations on concessional credit terms combined with conservative growth assumptions would not lead to a rapid re-accumulation o f debt in most cases. In cases where MDB lending would lead to risks o f debt distress, the DSF would detect this early on and MDB assistance would be provided on more concessional terms. 48. Given the new landscape in many LICs post-MDRI, extending creditor collaboration to bilateral creditors, in particular non-traditionalcreditors, becomes vital to maintaining debt sustainability. While debt distress in most LICs would not increase rapidly with traditional multilateral creditors, especially given their increased harmonization with the IDA14 grant allocation system, significant scaling up and non- concessional borrowing could exacerbate debt distress risks. The 2006 DSF review pointed out that debt sustainability would be jeopardized "relatively rapidly under scenarios o f significant increases in new borrowing by LICs from other creditors in addition to the main MDBs, particularly if borrowing i s contracted at or near market interest rates.7y36 49. A dialogue i s ongoing with the OECD Working Group on Export Credits and Guarantees (ECG) on the importance o f incorporating debt sustainability considerations into lending decisions. World Bank and IMF colleagues attended regular ECG meetings and workshops in 2006 and 2007. Information sharing has improved as a result o f this dialogue, and at the April 2007 meeting the OECD E C G agreed to reinstate a streamlined version o f reporting o f commitment data o n lending to 35 Small differences may arise as a result o f different assessments o f performance, or as a result o f the different timetable - with IDA'Sgrant eligibility determined at the start o f the fiscal year (July), while the AfDB's allocations and eligibility start at the beginning o f the calendar year. 36 World Bank and IMF, "Applying the Debt Sustainability Framework for Low-Income Countries Post Debt Relief', November 2006. - 20 - IDA-only countries, and to share (on a confidential basis) this information with the Bank and the IMF to help supplement debtor information in the DSAs. As a result o f improved outreach on the DSF and debt distress risks, the ECG also agreed to extend the statement o f principles o n unproductive expenditures from the group o f HIPCs to all IDA-only low- income countries. There are also ongoing discussions o n h o w to promote sound and sustainable lending practices, in particular in relation with the DSF. 50. Particularly important, given the recent increase in the volume o f new lending from non-OECD bilateral creditors, have been the discussions held with the major external financing institutions o f these countries. Delegations have held constructive dialogues with China and India in particular: mutual understanding has increased considerably. This dialogue w i l l continue to be an important component o f creditor outreach by the Bank and the Fund.37 Among other areas o f cooperation, an M O U was signed in M a y 2007 between the Chinese Exim Bank and the World Bank to "enhance their ongoing cooperation within their respective authority (including staff secondments, knowledge sharing and exchange on various aspects o f development assistance, such as fiduciary and financial management, procurement, and environmental and social impact analyses)". The scope o f the M O U i s as follows: strengthened communication and knowledge exchange, participation in donor coordination mechanisms and frameworks, co-financing o f projects, and staff exchange (see B o x 3). A similar M O U i s being discussed between the Bank and the China Development Bank. Box 3: M O U on Cooperation between the Export-Import Bank o f China and the World Bank An M O U with the Export-Import Bank o f China was signed with the World Bank o n M a y 21,2007. The MOU aims to improve the development impacts o f the respective operations o f both organizations through cooperation. The M O U specifies (i) principles on which the parties will enhance their ongoing the cooperation (including staff secondments, knowledge sharing and exchange o n various aspects o f development assistance such as fiduciary management, procurement and environmental and social impact analysis) and (ii) certain key terms and conditions t o be utilized when co-financing. Among the principles for financing sustainable development on which the cooperation will be based are: Promote sustainable economic and social development; 0 Ensure that assistance contributes to measurable development results and impacts; 0 Take measures (consistent with each party's respective guidelines, policies and procedures) to identify, minimize and mitigate adverse environmental and social consequences o f development initiatives; and 0 Collaborate o n ensuring sustainable development financing through appropriate levels o f concessionality. The scope o f the cooperation: 0 Strengthen communication and knowledge exchange, including country analysis, financing plans and specific projects in partner countries. 0 Facilitate participation in donor coordination mechanisms and frameworks. Identify possible projects suitable for cofinancing, and work together on these. Encourage exchange o f experience and staff training. 37 Discussions with non-OECD bilateral creditors on the DSF have also been held within the OECD E C G forum and in the Berne Union. -21 - 5 1. The Bank will continue to explore increased opportunities for collaboration with creditors, including the OECD, emerging market bilateral creditors and commercial creditors. The China E x i m Bank MOU with the World Bank i s a positive step in that direction. Further areas that are being discussed include the possibility o f focusing on collaboration with emerging market creditors in individual African country cases. Discussions have also been held with representatives from investment banks to share information o n the DSF, to understand h o w investment banks active in emerging markets assess risks in IDA countries, and to discuss ways to enhance the dialogue. The Bank will continue sharing data and technical workshops to familiarize creditors with the technical details o f debt sustainability analyses. This will also afford opportunities to the Bank to learn more about the policies o f other creditors, including their risk assessments and other factors that determine their lending decisions. The Bank and IMF will also continue t o provide advice to OECD country ECAs in their goal o f developing sustainable lending guidelines that could be tabled for approval at their fall E C G meeting. Update on Borrower Country Cases and Debtor Reporting 52. Implementation o f the second prong o f the NCBP has preceded as set out in the July 2006 NCBP paper. However, it i s too early to assess the effectiveness o f the policy given the single case to date. While the f i r s t prong o f the NCBP involved strengthened creditor collaboration, and strengthened debt reporting and debt management capacity, the second prong involves discouraging unwarranted non- concessional borrowing through disincentive mechanisms. After Board approval o f the overall NCBP framework, staff circulated an internal guidance note that detailed the various aspects o f the policy and the need for country teams to be involved in i t s implementation. An increased focus on adherence to OP 14.10 was also complemented by increased consultations with other creditors, and improving access to creditor-based debt data to complement the DRS. 53. The first case o f non-concessional borrowing to be assessed under the new policy was that o f Angola. Angola contracted substantial amounts o f non-concessional borrowing over the last few years. I t i s estimated that this amounted to $15.5 billion (about 35 percent o f i t s 2006 GDP) since 2004. While most o f this borrowing took place before the approval o f the non-concessional borrowing policy, several large loans were contracted after the policy was in place, and formed the basis o f IDA'Sassessment. 54. I n the case o f Angola the implementation o f the policy led to the decision to harden the terms o f Angola's borrowing. This decision was based o n the country- specific and the loan specific factors described in the N C B P paper (reproduced in Annex 4). In certain circumstances some non-concessional borrowing may be warranted if the financing were to help build a key piece o f infrastructure, particularly in a strong- performing country. The loan-specific assessment took into account the costs o f the borrowing and the lack o f information on the projects to be financed from the loans and their estimated returns. The country-specific assessment o f the policy environment showed that Angola's weak policy and institutional frameworks do not provide a strong platform for such volumes o f borrowing o n commercial terms. - 22 - 55. The decision to switch to hard-term lending in Angola balances the tradeoffs between debt sustainability and progress towards the MDGs. Providing IDA grants to a country which consistently borrows large volumes on non-concessional t e r m s was f e l t to be a poor use o f IDA concessionality. At the same time, cutting the volumes o f IDA financing in this post-conflict country was f e l t to be counter-productive as this would further reduce IDA'Salready limited engagement needed for technical assistance and capacity building. In terms o f the impact o n debt sustainability, IDA lending on any t e r m s only marginally affects the trajectory o f the NPV o f debt-to-GDP or the NPV o f debt-to exports ratio, given the small share o f IDA in overall financing flows. 56. The Angola case illustrates some limitations of the NCBP, in particular when IDA financing i s small relative to other sources and when information i s inadequate. Where IDA financing i s dwarfed by other sources o f financing, the possibility o f a disincentive effect i s small. The case also points out the difficulties with weak debt reporting. Although there may be informal sources o f information or media reports o n non-concessional borrowing, a decision by IDA requires additional time to receive confirmation o f the size and nature o f the borrowing from the borrower. 57. The Angola case also points out the importance of looking beyond the grant allocation system and the NCBP in helping resource-rich countries achieve debt sustainability. Here, attention to helping countries with debt management, medium t e r m debt strategies, and resource management policies will be important complements to the policy. The Angola Interim Strategy Note (ISN) outlines the Bank's continued engagement with the government in policy dialogue and nonlending technical assistance in the petroleum sector. For many countries technical assistance from the Bank can help over the long term to build up their institutional and technical capacity to manage o i l and mineral resources, and the related social and environmental impacts. Technical assistance can also help improve transparency, promote private sector investment and generally improve the contribution o f the mineral sector to the country's economic development and poverty reduction. The multi-donor Extractive Industries Transparency Initiative trust fund administered by the Bank could be an additional source o f grant financing to help countries improve the transparency o f activities in this sector. 58. Unless future cases also involve large and clearly disproportionate non- concessional borrowing, application o f this prong o f the non-concessionalborrowing policy may be difficult. In particular where borrowing i s being contracted by low-risk countries with strong policies and institutions, and where a lack o f concessional financing i s leading to non-concessional borrowing, it will be difficult to determine the appropriate disincentive. In such cases a more in-depth assessment o f the potential returns to the investment are needed, requiring more detailed information on the country's investment plans. At the same time, based o n existing literature, empirical evidence o n the link between debt-financed investments and growth i s ambigu~us.~' 38 See World Bank and IMF, "Applying the Debt Sustainability Framework for Low-Income Countries Post Debt Relief', November 2006, Appendix 3. - 23 - 59. A number o f small non-concessional loans have been identified by the DRS thus far for the period July to December 2007 for countries subject to the NCBP. While a number o f cases o f non-concessional borrowing also came on the radar screen in FY06, these took place before the approval o f the non-concessional borrowing policy. In FY07 non-concessional borrowing by Angola was confirmed from a number o f emerging market credit01-s.~~ number o f commercial loans were also reported by another A country, but it was determined that these formed part o f an overall package that is concessional. T w o other countries reported multilateral loans that did not meet the Bank's definition o f concessionality, although the amounts are small.40 Staff i s following up with the authorities and the creditors concerned on the nature o f these loans. Staff will also continue to follow-up with the relevant creditors during regular consultations. 60. As a result o f the NCBP and increased efforts to encourage adherence to OP14.10, most countries by the end of July had provided their annual reports to the DRS. A special Bank working group has been leading the efforts to increase adherence to reporting requirements under OP14.10 and increase the quality o f these reports.41 Only 16 IDA-only countries had not yet provided their annual reports under OP14.10 by the end o f July, 2007. This adherence level was about two months ahead o f reporting performance from previous years. Countries that were not compliant with the annual reporting requirement were informed that they are required to submit their end-2006 annual report or provide an acceptable action plan for such reporting before any new IDA operation can be presented to the Board. Subsequently five additional countries submitted acceptable reports, including one country which reported for the first time in almost a decade. Reports with significant gaps or quality issues were received from another 3 countries, and staff i s following up on these. 61. However, more progress i s needed on quarterly reporting to have timely information. Very few countries have provided quarterly reporting for the first or second quarter o f calendar year 2007, and more work needs to be done to improve adherence to quarterly reporting. Here the increased Bank emphasis o n debt management should go a long way to improving capacity to report o n a quarterly basis. As discussed earlier in the paper, parallel efforts are also underway for improved information flows from other creditors on the terms and volumes o f new lending to low-income countries. 62. The advanced reporting requirement introduced by the NCBP has led to discussions with some borrowing countries. This requirement asked that countries report to the Bank their plans for non-concessional borrowing ahead o f contracting the loans, giving the Bank a chance to present the country with alternative financing 39 In Angola, debtor reporting was a key focus o f the IDA N C B P response. A full debt report was submitted by Angola in August 2007, although a few remaining issues are being followed-up with the authorities. 40 The definition o f concessionality i s set out in IDA (2006b), 0p.cit. and an accompanying template published on the IDA website provides a tool to calculate the concessionality o f any given loan. 4' The first measures included increased scrutiny for countries ahead o f Board presentation, and special reminder letters were sent out at the beginning o f the year to all I D A - o n l y countries. Where reports are incomplete or are internally inconsistent, a further dialogue with the country authorities takes place to correct the inconsistencies. - 24 - scenarios and implications o f such borrowing. The supplemental letter on Financial and Economic Data that i s included in the signing package for a l l IDA Credit and Grant Agreements was revised to include this new requirement for grant-eligible and post- MDRI countries. While no specific report was submitted as a result o f the advance reporting requirement, several countries discussed their desire to increase access to financing with the Bank country team, and indicated that without additional concessional financing that they would consider commercial financing for important infrastructure investments. A dialogue with these countries has been taking place in the context o f their Bank and IMF programs o n the best means t o finance the investments, including through broader assessments o f fiscal space. 63. I n countries that have expressed interest in contracting new loans, the Bank can help by providing advice on debt management, help develop medium term debt strategies, and outline best borrowing practices. Bank technical assistance can also provide guidance to ensure favorable loan terms and facilitate the best use o f such finance in order to reduce the risks o f costly financing for non-viable projects. For instance, the Bank can focus on ways to mobilize additional private capital to complement ongoing and planned public investments. It can also outline the menu o f financing options from the Bank group, including guarantees from IDA and MIGA, investment financing (debt, equity and quasi-equity) from IFC, and IBRD enclave lending and enclave guarantees to provide additional support to a country's infrastructure development. 64. In sum, the t w o main prongs o f the NCBP are being implemented, with steady progress being made in creditor coordination around the DSF, and with borrower disincentivesapplied to one case. IDA and IMF outreach to other creditors has led to the adoption o f similar grant allocation systems by MDBs. An ongoing dialogue on the importance o f taking debt sustainability into consideration in lending decisions has taken place with OECD Export Credit Agencies and emerging market bilateral creditors. With 6 months o f information since the approval o f the NCBP, only one case o f borrower disincentives has been discussed by the Board thus far - Angola - whose financing terms have been hardened. Very recently two new cases o f non- concessional borrowing have been confirmed and brought to the attention o f staff, including a Eurobond issue o f US$750 million by Ghana. In addition, media has been reporting on large scale borrowing in another grant-eligible country that has yet to be confirmed. Staff i s in the process o f verifying information regarding these cases, and as new cases are brought forward, the process outlined in B o x 2 will be followed. V. Capacity Building in LICs to Manage New Borrowing 65. Despite the efforts o f various technical assistance providers on public debt management, much remains to be done. The recent Bank-Fund paper on strengthening debt management practices points out that in LICs "challenges tend to be more acute - - 25 - capacity, institutional arrangements, governance, all need considerable ~trengthening."~~ That paper points out that a number o f international providers o f technical assistance provide debt management related capacity building. However, each specializes in a particular area, such as providing debt management software, or improving analytical capability, or training staff in debt management units. There remains a need for a systematic approach to deal with the entire debt management process, including case-by- case assessment o f gaps and weaknesses. 66. Efforts towards capacity building for improving debt sustainability in LICs have been ramped up along three dimensions: 0 First, debt sustainability analysis training workshops have been delivered to groups o f countries covering most IDA-only borrowers. Recent DSF workshops have been given in Mexico City, Accra, Dakar, Maputo, and Windhoek, reaching about 40 LICs. 0 Second, a Debt Management Performance Assessment T o o l (DeMPA) was launched in 2006 and has been piloted in five LICs (Malawi, Gambia, Guyana, Nicaragua, and Albania). The tool assesses `sound practice' for effective and efficient debt management based on what has worked well in different country situations. This comprehensive and standardized tool has been developed jointly by P R E M and the Bank's Treasury, incorporating broad consultation with the IMF and all existing service providers in this area. The tool builds on PEFA indicators for public financial management, and covers core central government debt management functions relating to: governance and strategy development, coordination with macroeconomic policies, borrowing and related financing activities, cash flow forecasting and cash balance management, operational risk management, and debt records and reporting. The measurement o f 14 performance indicators allows the Debt Management Performance Assessment tool to pinpoint areas in need o f priority attention. DeMPA will enable a common assessment among donors, and provides a methodology to monitor debt management performance over time. 0 Third, work has begun towards strengthening joint technical assistance by the Bank and the IMF in low-income countries. In the coming year, medium-term debt management strategies will be developed with Bank-Fund technical assistance in an initial group o f LICs, based on analytical tools and methodological approaches currently under development. These strategies should help countries towards better decision making over debt composition, as well as the level o f borrowing. 67. The Bank i s working to develop and apply these tools more broadly across LICs. The Bank plans to apply DeMPA in 60 Low-Income Countries over the next three 42 See World Bank, "Strengthening Debt Management Practices - Lessons from Country Experiences and issues Going Forward", April, 2007. -26- years. D e M P A will be disseminated externally for further comment and refined to provide web-based materials that will be the basis for staff and client training programs. The debt management toolkit will initially be applied in 4-6 countries in 2008 before rolling out more broadly as part o f country assistance strategies. 68. This j o i n t w o r k o f IDA and the IMF will complement ongoing outreach t o creditors and help reinforce sustainable lending practices. Ultimately the most effective means o f attaining debt sustainability in LICs i s strengthened debt management capacity: the ability to analyze debt strategy, fiscal policy, and broader economic policy in light o f their long-term effects on debt sustainability and vulnerability to economic shocks, and to manage the existing and new debt stock to control debt service and reduce fiscal and foreign exchange risks coming from the public debt portfolio. Building this capacity will be a long-run endeavor, one that i s part o f the core mandate o f IDA but that can only be delivered in partnership with other institutions. VI. Conclusions and Issues f o r Discussion 69. D e b t relief and increased concessionality o f assistance need to b e combined with policies that support private sector-led g r o w t h in o r d e r to generate sustainable trajectories f o r debt-burden indicators over time. IDA's policy advice and financial assistance helps countries improve their macroeconomic performance and strengthen the micro-macro linkages in the development process. IDA's support for investment climate reforms and infrastructure investments all have direct implications for debt sustainability, since strong growth o f GDP, revenues and exports help lower debt ratios. While maintaining debt sustainability i s ultimately the responsibility o f debtor countries and all creditors collectively, IDA plays a key role in helping ensure debt sustainability through initiatives that alleviate countries' debt burden as well as through i t s support to growth- promoting policies. 70. Deputies may wish to consider the following issues for discussion: 0 D o Deputies agree that the DSF risk ratings based on forward-looking DSAs provide an effective basis for the IDA grant allocation system? 0 D o Deputies agree that the I D A 1 4 grant allocation system i s functioning reasonably well and that this system should continue to be used as the basis for determining the mix o f credits and grants in IDA15? 0 D o Deputies agree that Timor-Leste should continue to have access to the grant equivalent o f i t s P B A allocation (i.e. 60 percent), given i t s continued post-conflict status, and constraints on i t s ability to contract external debt? 0 D o Deputies agree that based o n i t s status as part o f Serbia under United Nations Administration, Kosovo' s eligibility for grants would continue in I D A 1 5 as long as i t s status remains the same? - 27 - Annex 1. Traffic lights for FYO8 IDA allocations: IDA-only countries (based on joint Bank-Fund L I C DSAs available as o f end-June, 2007) Debt stock indicators (2005, percent) 4/ FY08 traffic NPV o f debt-to- NPV of debt- Country light GDP to-exports Bhutan Red I1 79 357 Burkina Faso Yellow Cape Verde Green Ghana Green Maldives Green 11 33 43 Nicaragua Yellow Samoa Green Tanzania Green Uganda Green Vietnam Green Medium performance (3.75xCPIA<3.25) 3/ 40 I so Bangladesh Green Benin Yellow Cameroon Green Ethiopia Yellow Guyana Yellow Kenya Green Kyrgyz Republic Red Lesotho Yellow Madagascar Green Malawi Yellow Mali Green Mongolia Yellow Mozambique Green Nepal Red Niger Yellow Rwanda Red Senegal Green Yemen, Republic o f Yellow Zambia Green Afghanistan Red Burundi Red Cambodia Yellow Central African Republic Red Chad Red Comoros Red Congo, Democratic Republic o f Red 1/ 107 389 Congo, Republic o f Red Cote d'Ivoire Red 11 61 137 Djibouti Red Eritrea Red 11 8/ 42 675 Gambia, The Red Guinea Red Gumea-Bissau Red Haiti Red Kiribati Green 61 Lao People'sDemocratic Republic Red Liberia Red Mauritania Yellow Moldova Green Myanmar Red 1/2/ N/A 128 Sao Tome and Principe Red 71 Sierra Leone Yellow Solomon Islands Red Somalia Red 1/21 NIA 1091 51 Sudan Red 11 65 473 Tajikistan Red Togo Red Tonga Red Vanuatu Green 11 15 35 -28- 1/ Due to the unavailability o f a joint Bank-Fund DSA at the time o f the FY08 allocation exercise, the traffic light i s based on the historic debt stock indicators (the first pillar o f the DSF) and performance category is based on 2004-6 CPIA average. 2/ 2006 CPIA ratings are unavailable and a weak performer i s assumed. 3/ Based on performance category used in DSAs, unless otherwise specified. 4/ The debt service-to-exports indicator i s not presented as the two debt stock indicators already exhibit a red light for the countries. N P V o f debt refers to public and publicly guaranteed external debt plus I M F credits in consistency with the DSF. Exports refer to 2003-5 year average. The source o f N P V o f debt data i s the World Bank Global Development Finance 2007; that o f GDP i s the World Development Indicators 2007; and exports the World Economic Outlook 2007. 5/ The ratio i s as o f end July 2006 ("HPC - issues related to the sunset clause", August 2006). 6/ Neither DSA nor debt data available. "Green" light i s assumed. 7/ Based on an expected new DSA rating o f high debt distress risk (to be conducted in September). 8/ An unpublished DSA for Eritrea also confirms the red light rating. N/A not available. I < P E i V I 3 (r 0 rn I .. ..X...... .......... . . ........ . . I. .. ..___ W E A G " B z d" d B W E i V r- E i V I 4 m Lo I Q V m E P E Lo E 2 e 9 "2 p5 . - c! v) a2 . II a a cd h 0 I .I E M E *E E L 8 8 I cd E 0I . L v) v) P a2 2 I c\1 m s fo I za2 5 3 I I 'c 0 5 g 0 CI . % I L Y E a 0 0 !CI 8 pc rn E cd I a2 P .I . MI I W CI E i .. rc) x Q) E c c I I - 33 - Annex 4 Principles that would guide exceptions to non-concessional borrowing ceilings43 The concessionality benchmark proposed for the purposes o f identifying breaches o f the Non- concessional Borrowing Policy (NCBP) has been a proven benchmark in PRGF programs, and has served as a useful tool in that context to provide the borrower some ``leverage'' with the creditor in obtaining the best possible financing for a potential investment. PRGF programs clearly define ceilings on allowable non-concessional borrowing in countries (which are often zero). In the context o f the PRGF, these limits can be overridden in one o f three ways: (i) agreement ex ante within finding program criteria by defining a subceiling to accommodate a specific non-concessional loan, (ii) alternative financing or co-financing that would make the investment concessional or (iii) making a case that a waiver be granted for the performance criteria. Similar to considerations that feed into decisions on non-concessional borrowing limits in the PRGF, a number o f country-specific and loan-specific factors would be taken into account in the N C B P to assess whether an exception to the zero-ceiling using the proposed benchmark i s warranted. Although many proposed loans may have merit o n specific economic or financial terms, the country environment in which they occur will strongly influence actual outcomes. There should be a favorable assessment at both the country-specific level and the loan-specific level to warrant an exception. Country-specific: Overall borrowing plans o f the country. A modest level o f overall borrowing by the country o n the basis o f the D S A to accommodate a particular investment may warrant consideration. For such a consideration, clear reporting o f overall borrowing plans i s needed, and enhanced creditor coordination through the DSF would facilitate this possibility. 0 Impact o f borrowing on the macroeconomic framework. Whether or not the borrowing would have a deleterious effect o n the macroeconomic framework would influence the consideration o f an exception. Impact on the risk o f debt distress. The current risk classification and whether or not the loan i s likely to lead to a higher risk o f debt distress will be a key consideration. Given their lower-risk o f debt distress, and generally better performance, more flexibility i s envisaged for "green light" countries. In addition, "yellow light" countries could benefit from somewhat greater (although s t i l l exceptional) flexibility than "red light" ones. Strength o f policies and institutions, especially public expenditure management and debt management. As the fiscal space Board paper makes clear, policies and institutions in particular those governing the efficiency o f public investment are critical.` Without these, even high return projects may fail to meet objectives. Loan-specific: 0 Development content and potential impact o f the loan, Le., investment will unlock a proven bottleneck to development as determined by analytical work such as a PER. 0 Estimated economic, financial and social returns to investment o f the project, weighted by the probability that the project w i l l succeed. 0 Lender equity stake in the project. N o additional costs associated with the loan, Le., collateralization, hidden costs. 0 N o other sources o f more concessional financing are available. 0 Concessionality o f the overall financing package for a particular investment. 11 See World Bank (2006). "Fiscal Policy for Growth and Development: A n Interim Report", DC2006- 0003, April 6. 43 Based o n IDA (2006b), Op. cit. B o x 3.