IDA’S NON-CONCESSIONAL BORROWING POLICY: 2019 REVIEW Development Finance Corporate IDA and IBRD (DFCII) September 24, 2019 ACRONYMS AND ABBREVIATIONS Fiscal year (FY) = July 1 to June 30 AfDB African Development Bank EMATUM Empresa Moçambicana de Atum ASA Advisory Services and Analytics ENHI Empresa Nacional de Hidrocarbonetos de Moçambique AsDB Asian Development Bank FCS Fragile and Conflict-Affected Situations CEO Chief Executive Officer FCV Fragility, Conflict and Violence CFP Concessional Finance and Global G&I Governance and Institutions Partnerships CFPVP Concessional Finance and Global GAF Grant Allocation Framework Partnerships Vice Presidency CPIA Country Policy and Institutional GDP Gross Domestic Product Assessment CRO Chief Risk Officer GVC Global Value Chains DAC Development Assistance HIPC Heavily Indebted Poor Country Committee DEC Development Economics IADB Inter-American Development Bank DECVP Development Economics Vice IDA International Development Presidency Association DeMPA Debt Management Performance IFAD International Fund for Assessment Agricultural Development DFID Department for International IFIs International Financial Development Institutions DFiVP Development Finance Vice IMF International Monetary Fund Presidency DLP Debt Limits Policy JET Jobs and Economic Transformation DMF Debt Management Facility LIC DSF Debt Sustainability Framework for Low-Income Countries DMS Debt Management Strategy LICs Low-Income Countries DPF Development Policy Financing LNG Liquified Natural Gas DRS Debtor Reporting System MD Managing Director DSA Debt Sustainability Analysis MDBs Multilateral Development Banks DSEP Debt Sustainability Enhancement MDRI Multilateral Debt Relief Initiative Program DSF Debt Sustainability Framework MFD Maximizing Finance for Development MOU Memorandum of Understanding MoF Ministry of Finance MPA Multi-Pronged Approach PPG Public and Publicly Guaranteed MTI Macroeconomics, Trade and PREM Poverty Reduction and Economic Investment Management NCB Non-Concessional Borrowing OC Operation Committee NCBP Non-Concessional Borrowing PRMVP Poverty Reduction and Economic Policy Management OECD Organization for Economic Co- SCD Systematic Country Diagnostics operation and Development OPCS Operations Policy and Country SDFP Sustainable Development Finance Services Policy OPSVP Operations Policy and Country SDGs Sustainable Development Goals Services Vice Presidency PBA Performance-Based Allocation SOEs State-Owned Enterprises PEFA Public Expenditure and Financial TA Technical Assistance Accountability POC Program for Creditor Outreach TDB Eastern and Southern Africa Trade and Development Bank TABLE OF CONTENTS EXECUTIVE SUMMARY ..................................................................................................................i I. INTRODUCTION ...................................................................................................................1 II. KEY FEATURES AND EVOLUTION OF THE NCBP ........................................................ 3 III. REVIEW OF NCBP IMPLEMENTATION ............................................................................7 A. Update on Country Cases since the 2015 Review ............................................................. 8 B. Implementation Assessment ............................................................................................ 12 C. Effectiveness of the NCBP .............................................................................................. 18 D. Creditor Coordination and Outreach ................................................................................21 E. Summary of Key Findings ............................................................................................... 25 IV. CONCLUSION AND NEXT STEPS ....................................................................................26 LIST OF ANNEXES Annex 1. Additional Tables ...............................................................................................................29 Annex 2. NCBP Implementation – Survey of the World Bank (WB) Country Teams ..................... 31 Annex 3. Outreach on Debt – Multilateral Development Bank’s (MDB) Survey on Sustainable Lending Practices ................................................................................................................34 Annex 4. Public Debt Dynamics and Debt Vulnerabilities ............................................................... 37 Annex 5. Summary of Proposals Made in the 2015 Review ............................................................. 39 Annex 6. Principles that Guide Exceptions to Non-Concessional Borrowing Ceilings ....................41 LIST OF BOXES, FIGURES AND TABLES Boxes Box 1. A Recap of Background Papers on Debt Vulnerabilities in IDA Countries ............................ 7 Figures Figure 1. Evolution of the NCBP .........................................................................................................4 Figure 2. NCBP Country Cases, Responses and Debt Risk, 2007-2019 ...........................................17 Figure 3. Share of Concessional Loans in Total External Debt (%) ..................................................20 Figure 4. Factors that have constrained the impact of NCBP: Country Team ..................................20 Figure 5. Factors that Influence MDB NCBPs ..................................................................................24 Tables Table 1. List of IDA Countries with NCBP Responses by Year (as of August 2019) ...................... 15 Table 2. List of Recent Outreach Events ........................................................................................... 22 i EXECUTIVE SUMMARY i. The IDA Non-Concessional Borrowing Policy (NCBP) was introduced in 2006.1 The policy is part of IDA’s toolkit to help countries improve debt sustainability. It aimed at addressing situations in post-Multilateral Debt Relief Initiative (MDRI) and grant-eligible IDA-only countries in which debt relief or IDA grants could potentially: i) cross-subsidize creditors that offer non- concessional loans to recipient countries, or ii) create incentives for these countries to overborrow, thereby eroding the gains to debt sustainability obtained from debt relief and grants. The NCBP acknowledges that non-concessional borrowing (NCB), under certain circumstances, can be a useful component of a financing mix that helps finance development needs of low-income countries. ii. From its onset, the NCBP framework included periodic implementation reports to the Board of Executive Directors. Reviews and updates were completed in September 2007, June 2008, April 2010, and October 2015.2 As the policy evolved, countries’ development paths became more heterogenous and the development finance landscape became more complex. In this regard, adjustments to the policy were proposed in these respective reviews, mainly to enhance flexibility, consistency, transparency, and to align the policy to internal and external changes, including to the Debt Limits Policy (DLP) of the International Monetary Fund (IMF). iii. This Review is undertaken in the context of the WB-IMF Multipronged Approach (MPA) to address debt vulnerabilities and has been taking place in parallel with the IMF review of its DLP. Preliminary findings of the review were presented at the IDA19 Replenishment meetings in April and June 2019. At these meetings, Participants also discussed and voiced their support for a proposal to transition from the NCBP to a new Sustainable Development Finance Policy (SDFP), given the renewed challenges on debt sustainability in IDA countries. Therefore, this will be the last review of the NCBP, and it will primarily aim at drawing lessons to inform the SDFP. Preparation and implementation of the SDFP would be included as an undertaking in the IDA19 Replenishment Agreement, and Management is committed to bringing a full-fledged proposal to the Board during the third quarter of FY20, in coordination with IMF’s completion of the review of its DLP. iv. The Review takes place against the backdrop of an evolving development finance landscape which has become relatively more complex, and that poses challenges for IDA countries. In the context of limited supply of concessional finance and large development finance needs, the role of non-traditional bilateral and private lenders, including non-Paris Club creditors, offering predominantly non-concessional funding through riskier, more sophisticated, and expensive debt instruments, has been growing. Given low debt management capacity in many IDA countries, this may have contributed to rising debt vulnerabilities. 1 IDA.2006. “IDA countries and non-concessional debt: dealing with the ‘free rider’ problem in IDA14 grant-recipient and post-MDRI countries”. IDA/R2006-0137/1, June. 2 The 2015 Review was presented to the WB Executive Board on October 29. ii v. This Review was informed by a comprehensive internal and external consultation process, involving consultations with the IMF, other Multilateral Development Banks (MDBs), World Bank (WB) country teams, and other stakeholders. It was also informed by an analysis of: i) the implementation experience obtained from the full inventory of country cases reviewed under the NCBP; ii) country-level data since 2007; and iii) findings from an extensive survey of country teams and of MDBs. The Review has also benefited from the analysis of debt developments that were carried out over the last year. vi. The Review assessed the experience with the NCBP across IDA countries, with a focus on the period following the 2015 Review, including an overall analysis of the experience since the policy’s introduction in 2006. While noting the shortcomings of assessing the effectiveness of the NCBP, including the lack of “counterfactuals” and the multiple factors influencing borrowing decisions, several key lessons were gleaned through this review. vii. The NCBP was implemented as intended and has provided a platform for dialogue with IDA countries on debt sustainability, but there is room for improvement. The NCBP has provided a flexible framework to assess the impact of and rationale for NCB based on country- and project-specific factors. The NCBP and IDA’s responses have complemented and/or fostered IDA country dialogue and programs, along with influencing countries to seek alternative sources of concessional financing. However, the Review found that, overall, the NCBP had a positive but limited impact in helping countries take steps towards reducing debt vulnerabilities, and that there were areas in the implementation of the policy which need to be enhanced, including country and debt coverage, further streamlining of the internal decision-making process, as well as transparency and communication of NCBP decisions. viii. This Review’s finding regarding the limited impact of the NCBP is consistent with the findings of the previous reviews. Further adjustments to the policy implementation were made in the 2008, 2010 and 2015 Reviews in response to these findings. Given the limited impact of these adjustments, Management’s proposal is to move to a more comprehensive and ambitious SDFP next fiscal year. The SDFP would aim to introduce an incentive-based system with the appropriate performance and policy actions necessary to help client countries address the causes of debt vulnerabilities. ix. The narrow scope of the policy in terms of country coverage and focus on non- concessional external debt may have limited the ability of the policy to influence prudent lending decisions more broadly. NCBP only applies to grant-eligible and non-gap MDRI recipient countries. IDA’s role in promoting debt sustainability could thus be further enhanced by expanding the coverage of the policy to all IDA countries. In addition, NCBP implementation also focuses on external NCB. The rationale for focusing on external NCB, at the time of the policy’s introduction, stemmed from the greater risks that such borrowing put on debt sustainability. Yet, broader factors beyond non-concessional external debt are driving external debt vulnerability in the evolving landscape, and the share of domestic debt is rising. A broader view of debt and a broader application of the policy would enhance its impact and help tighten its alignment with the current IMF’s DLP, which encompasses both external and domestic public debt. x. Progress has been made in streamlining internal processes to enable faster decision- making in responding to cases of non-compliance or requests for waivers, but more needs to iii be done in terms of outreach and agility. A survey of country teams indicated that the process of reviewing NCBP cases takes longer than expected and that there has been a lack of brief and concise information on the NCBP that is easily accessible, which is important especially in countries with high staff turnover rates. The survey pointed to the importance of heightened outreach activities, including a communications strategy, that could help streamline the decision- making process. xi. Regarding enhancing transparency, while changes in IDA terms or allocation volumes resulting from NCBP non-compliance are publicly available, consultations and surveys suggest that the information is not easily accessible to most stakeholders. In addition, the current disclosure framework lacks the context required to understand the rationale behind the NCBP Committee’s decisions on waivers and responses. While the WB Board is informed about NCBP cases and responses after a Management decision is made, these decisions are not publicly available and are only disseminated, with a considerable lag, in the periodic NCBP Reviews and updates. Given the value of clear signaling to both creditors and borrowers, the surveys identify a need for revisiting the dissemination framework and stressed the importance of continued outreach and communication. xii. The NCBP alignment with the IMF’s DLP has been a critical aspect of the policy and should be sustained, hence the 2019 Review of the NCBP was conducted in parallel with the IMF’s review of its DLP. The joint WB-IMF shared framework to establish debt limits facilitates continuity of the policy advice to countries that move in and out of IMF arrangements. Strong coordination, including in setting ceilings, will continue to be important especially given the recommended increase in country coverage in the forthcoming SDFP. This includes setting ceilings for countries that are not grant eligible, and that would hence be outside of the scope of the current NCBP. xiii. The Review confirmed that the NCBP has a positive impact on creditors’ lending decisions, but this can be enhanced including through greater outreach and joint principles. Some creditors rely heavily on the NCBP and the IMF’s DLP, especially in cases where zero ceilings are in place, and others, such as the African Development Bank (AfDB), have adopted polices consistent with the NCBP. A survey of MDBs suggests that more than half are familiar with the NCBP and factor it into their lending decisions. IDA has had extensive dialogue with official creditors to better complement policies on NCB, and this continues to date. IDA and the AfDB have agreed to identify a set of high-level principles for promoting sustainable lending. Outreach through IDA’s “Lending to LICs” mailbox3 could be expanded, including to increase engagement with non-Paris Club bilateral and private creditors which remains sporadic. 3 The World Bank and the International Monetary Fund have established special email service accounts (LendingToLICS@worldbank.org and LendingToLICS@imf.org) in response to requests from creditors. The purpose includes: a) to help creditors in making lending decisions that take debt sustainability and concessionality requirements into account, and b) to enable creditors to ask for clarification on the DSF, aspects of Bank or IMF concessionality policies, and the application of these policies in specific countries where creditors may be contemplating doing business. iv xiv. The lessons emerging from the Review have underpinned the proposal to transition to a new, more comprehensive and ambitious SDFP that is being developed in the context of the IDA19 Replenishment. The name reflects the aims of incentivizing sustainable borrowing and encouraging coordination among all creditors and broadening the focus beyond external NCB. xv. Implementation of the forthcoming SDFP would start on July 1, 2020. The Board Paper on implementation of the SDFP will be delivered during the third quarter of FY20 in coordination with the review of the IMF’s DLP. The SDFP would, inter alia: i) broaden coverage to apply to all IDA countries; ii) focus on incentives rather than disincentives; iii) enhance creditor outreach through a broader scope and more information sharing, especially among the MDBs; and iv) take a broader view of debt which will tighten alignment with the IMF’s DLP. The new policy would continue to be rules-based and relatively more flexible, to maintain consistency with country context and capacity – especially for Fragile and Conflict-Affected Situations (FCS) and Small States. However, the primary role of ensuring debt sustainability rests with country borrowers and would need cooperation and coordination by all stakeholders, including non-Paris Club and other non-traditional creditors, as well as the IMF. 1 I. INTRODUCTION 1. The IDA Non-Concessional Borrowing Policy (NCBP) was introduced in 2006. The policy is part of IDA’s toolkit to help countries improve debt sustainability. It aimed at addressing situations in post-Multilateral Debt Relief Initiative (MDRI) and grant-eligible IDA-only countries in which IDA’s debt relief and/or grants could potentially: (i) cross-subsidize creditors that offer non-concessional loans to recipient countries, or (ii) create incentives for these countries to overborrow, thereby eroding the gains to debt sustainability obtained from debt relief. The policy is pursued through a two-pronged strategy involving: i) enhancing creditor coordination; and ii) encouraging appropriate borrowing behavior through borrower disincentives to discourage external non-concessional borrowing (NCB) by grant eligible and post-MDRI countries. Under the policy, NCB is reviewed either based on loan-by-loan or on borrowing ceilings. In cases of breaches, IDA can either issue waivers or respond by reducing allocated IDA volumes, hardening terms, or a combination of both. 2. Periodic implementation updates and reports on the NCBP framework were presented to the Board of Executive Directors on several occasions. Reviews4 and updates were completed in September 2007, June 2008, April 2010, and October 2015. The policy evolved as countries development paths became heterogenous, and the development finance landscape became more complex. In this regard, the adjustments to the policy were proposed in these respective reviews, mainly to enhance flexibility, transparency and to align it to internal and external changes, including to the Debt Limits Policy (DLP) of the International Monetary Fund (IMF). 3. The 2015 Review helped to ensure continued alignment of the NCBP with the evolving DLP and to improve transparency around policy implementation. It included a number of enhancements to the NCBP’s implementation arrangements such as: (i) streamlined assessment of debt management capacity, (ii) introduction of debt ceilings on total public and publicly guaranteed (PPG) debt (concessional and non-concessional) in present value terms for countries at low or moderate risk of debt distress with adequate capacity, and (iii) enhanced transparency and reporting including publicly available information on the IDA website. The summary of proposals made in the 2015 Review is presented in Annex 5. 4. This Review is undertaken in the context of the World Bank-International Monetary Fund (WB-IMF) Multipronged Approach (MPA) to address debt vulnerabilities. The IMF is undertaking a parallel review of its DLP. Preliminary findings of the review were presented at the IDA19 Replenishment meetings in April and June 2019. At these meetings, Participants also discussed, and voiced their support for, a proposal to transition from the NCBP to a new Sustainable Development Finance Policy (SDFP), given the renewed challenges on debt sustainability. Therefore, this will be the last review of the NCBP, and it will primarily aim at drawing lessons to inform the SDFP. Preparation and implementation of the SDFP would be 4 IDA. 2007. “The role of IDA in ensuring debt sustainability: a progress report”. IDA/SecM2007-0590, September; IDA 2008. “IDA’s Non-Concessional Borrowing Policy: Review and Update”. IDA/SecM2008 -0473, June; IDA 2010. “IDA’s Non-Concessional Borrowing Policy: Progress Update.” IDA/SecM2010 -0240, April; IDA 2015. “IDA’s Non-Concessional Borrowing Policy: Review and Update”. IDA/SecM2015 -0161, October. 2 included as an undertaking in the IDA19 Replenishment Agreement, and Management is committed to bringing a full-fledged proposal to the Board for its approval in the third quarter of 2020. 5. The Review takes stock of the experience since the introduction of the policy in 2006, with added focus on the period following the previous Review in 2015. In particular, the Review assesses the implementation of the NCBP and examines: i) if the NCBP was implemented as intended; ii) the extent to which the NCBP has been effective in meeting its objectives; and iii) if adjustments are needed including in response to the changing development finance landscape. The Review was informed by a comprehensive analysis of: i) country-level data since 2007; ii) record of implementation experience obtained from the full inventory of country cases reviewed under the NCBP; iii) findings from an extensive consultation efforts and a survey of country teams and of Multilateral Development Banks (MDBs) undertaken as part of the review; and iv) a comprehensive analysis of debt dynamics in IDA countries (See Box 1). 6. The Review takes place against the backdrop of an evolving development finance landscape5 with growing importance of non-traditional private lenders. In this more differentiated and complex finance architecture, the challenge for IDA countries is to tap into the positive global trends in private finance. The World Bank Group (WBG) has adopted the Maximizing Finance for Development (MFD) approach to help countries maximize their development resources by systematically leveraging all sources of finance, expertise, and solutions to support developing countries’ sustainable growth. 7. The new landscape presents not only financing opportunities, but also debt sustainability challenges for IDA countries. In the context of limited supply of concessional finance, private lenders, especially non-traditional private lenders, including the Non-Paris club creditors, offer riskier, more sophisticated and expensive debt instruments. Some of the debts are collateralized, which may have the advantage of reducing borrowing costs, but might also reduce budget flexibility by earmarking revenues. Collateralization may also make the process of debt restructuring, when necessary, relatively more complex. Moreover, most IDA countries lack appropriate capacity to manage public debt in this new landscape. While acknowledging that there are multitude of factors that determine debt developments, such as commodity price shocks, exchange rates fluctuations, weak economic growth and as well as weak fiscal controls, one of the consequences of the changing landscapes has been rising debt levels6,7 and debt vulnerabilities in IDA countries, which could hamper access to sustainable sources of development finance required to progress toward the achievement of the Sustainable Development Goals (SDGs) and other elements of the 2030 Agenda.8 5 For further discussion on trends in development finance see: IDA. 2019. “Debt Vulnerabilities in IDA Countries: Policy Options for IDA19”. IDA/SecM2019-0084, March 28. 6 See Annex 4 for detailed exposition of debt dynamics and developments . 7 IDA. 2019. “Addressing Debt Vulnerabilities in IDA Countries: Policy Options for IDA19”. Paper for IDA19 Replenishment, June. 8 The 2030 Agenda for Growth, Resilience and Opportunity sets the overarching theme for IDA18. It covers the 17 Sustainable Development Goals (SDGs), the COP21 climate agreement, the Addis Ababa Action Agenda, and the Sendai Framework for disaster risk management. See: IDA. 2017. “I DA18: Towards 2030, Investing in Growth, Resilience and Opportunity”, IDA Executive Directors Report, January. 3 8. The remainder of this paper is organized as follows. Section II provides a brief account of the features and evolution of the NCBP. Section III provides the review of NCBP implementation. In addition to an update of country cases reviewed by the NCBP Committee since 2015, it assesses the NCBP implementation to determine if the policy has been broadly implemented as intended and the extent to which it has been effective in meeting its objectives. An assessment of progress in creditor coordination and a summary of the key findings in support of transitioning to a new and more comprehensive and ambitious policy – the SDFP – are also discussed in this section. Section IV concludes and provides the next steps. II. KEY FEATURES AND EVOLUTION OF THE NCBP 9. The NCBP was approved by the World Bank (WB) Board in June 2006. The aim of the policy was to help grant-eligible and MDRI recipient countries take steps toward managing debt sustainability risks posed by external NCB. While the provision of grants and debt relief strengthened the debt sustainability prospects of IDA countries and increased resources for achieving the Millennium Development Goals (MDGs), it also increased the potential risk of “free riding”.9 Indeed, IDA and its shareholders aimed to lower the risk of debt distress in low-income countries by providing financial assistance on appropriate concessional terms; in contrast, other creditors could gain from non-concessional lending following large-scale debt relief or in conjunction with grants provided by IDA. IDA grants and debt relief could also introduce incentives for countries to overborrow from other creditors, which would force IDA to continue to increase the grant share of its assistance and/or defeat the original purpose of the MDRI. As such, from its onset, the NCBP had two prongs: (i) encouraging appropriate borrowing behavior by creating disincentives for external NCB; and (ii) enhancing communication, information flows and creditor coordination. 10. All MDRI and IDA grant-recipient countries became subject to the NCBP. The policy focused on external non-concessional PPG central government debt, consistent with the fiscal coverage under the Debt Sustainability Framework (DSF). Concessionality was established in terms of a minimum grant element10 requirement set at 35 percent. 11. To implement the NCBP, IDA used two instruments to respond to cases of NCB that were deemed to be inconsistent with the policy (based on country-specific and loan-specific factors): reducing the volume of IDA allocation, and/or a hardening term on which IDA allocations would be provided. The policy recognized the trade-offs emerging from the use of these instruments at the country level: volume reductions would decrease resources necessary to reach the MDGs, while hardening of terms could exacerbate existing debt sustainability problems. To minimize adverse impacts, instruments were applied in a way that took account of a country’s overall debt sustainability and access to financial markets. Volume reductions would primarily be used in countries where debt sustainability was a major concern and hardening of terms would 9 The term “free riding” was used in the original 2006 paper (see footnote 1) to refer to situations in which IDA’s debt relief or grants could potentially cross-subsidize lenders that offer non- concessional loans to recipient countries. 10 The grant element is the difference between the face value of a loan and its present value (PV), expressed as a percentage of the face value of the loan. The PV of a loan is the discounted value of the future debt service payments using the unified discount rate set currently at 5 percent. 4 primarily be used in countries with stronger prospects for debt sustainability and greater degree of market access. 12. The NCBP recognizes the existence of trade-offs between the goals of debt sustainability and financing development needs. In this regard, NCBP was not a blanket restriction on NCB and acknowledged that under certain circumstances, NCB can appropriately be part of a financing mix that helps promote economic growth. There may be cases where NCB may warrant an exception to the policy. Countries can ask for a waiver or exception to the policy on a loan-by-loan or a loan package basis, or they can request a debt ceiling in advance. Waivers or exceptions to the policy were guided by country-specific and loan-specific considerations (See Annex 6). 13. A key building block of the updated NCBP was the establishment of debt limits for countries subject to the policy. Although not part of the original NCBP, debt limits became an important feature of the policy and were applied using a differentiated approach based on the country’s debt vulnerabilities and capacity. Countries assessed at high risk/in debt distress under the joint WB-IMF DSF, had a zero NCB ceiling. However, these countries could request ex-ante or ex-post waivers for loans with high developmental impact. Countries at low or moderate risk of debt distress had the option to request a non-zero debt ceiling or request exceptions on a loan-by- loan basis. 14. Coordination with the IMF was, from the onset, an important component of the NCBP. Under the agreed framework, IDA countries subject to the NCBP under an IMF program were monitored under the IMF’s DLP. For countries without IMF programs, NCB issues were addressed under the NCBP. 15. The NCBP has evolved over time, with adjustments proposed in various reviews and updates. While initially the focus was on “free-riding”, this later evolved to emphasize the need to protect debt sustainability and exercise IDA’s fiduciary responsibility. Various adjustments to the policy were proposed, in different reviews, mainly to enhance flexibility, transparency and to align it to internal and external changes, including to the IMF’s DLP. The evolution of the NCBP is summarized in Figure 1. Figure 1. Evolution of the NCBP 5 16. In September 2007, and in the context of IDA15, a progress report on the implementation of the NCBP was presented to the Board. The report concluded that steady progress was made in creditor coordination. It noted that IDA and IMF outreach to other creditors led to the adoption of similar grant allocation systems by other MDBs. Progress had also been made in initiating dialogue with the Organization for Economic Cooperation and Development (OECD) Export Credit Agencies and emerging market bilateral creditors, including a signed Memorandum of Understanding (MOU) for cooperation with China’s EXIM Bank. Most IDA beneficiaries adhered to the policy, and borrower disincentives were discussed for only one country, Angola whose financing terms were hardened. Data availability was identified as a key challenge for the NCBP implementation. Among the actions to address this issue, IDA and the IMF developed and piloted a diagnostic tool to: i) measure debt management capacity; and ii) identify areas in need of technical assistance (TA). In addition, the WB and the IMF initiated work to develop a tool to help countries put together medium-term debt strategies. The 2007 report underscored that debt relief and increased concessionality of 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. 17. The June 2008 Review and Update concluded that the implementation of the NCBP was making good progress. IDA and the IMF had accelerated efforts to enhance borrowers’ debt management capacity and the development of medium-term debt management strategies. The new Debt Management Performance Assessment (DeMPA) tool for assessing debt management capacity was established and applied in 17 countries. This tool along with the DSF and the Medium-Term Debt Management Strategy (under development at the time) were expected to inform borrowers of the risks and trade-offs in any financing contemplated, whether concessional or non-concessional. A number of cases of NCB were discussed by management and appropriate IDA responses were determined according to the methodology set out in the policy framework. The creditor outreach had some success. The African Development Bank (AfDB), Asian Development Bank (AsDB), and the International Fund for Agricultural Development (IFAD) had adopted grant allocation systems similar to that of IDA. Other creditors such as the Inter-American Development Bank (IADB) took the risk of debt distress into account when determining the level of concessionality to offer to countries. The AfDB also introduced a new policy on non- concessional debt accumulation under ADF11, similar to the NCBP. The OECD Export Credit Group agreed to a set of principles and guidelines on sustainable lending that take the concessionality requirements of IDA and the IMF into account. Dedicated websites were established by the WB and the IMF with information on individual Debt Sustainability Analyses (DSA) as well as WB and IMF’s concessionality policies. In addition, email boxes (“Lending to LICs” mailbox) were established to respond to questions from creditors regarding, amongst others, the DSF and concessionality policies. 18. The April 2010 Progress Update, in addition to summarizing progress in the implementation of the policy, proposed adjustments informed by lessons learned. These included: i) limitations in the ability of the NCBP to affect NCB decisions, particularly relevant when IDA financing is small relative to other financing sources; ii) upstream dialogue on NCB cases was crucial as it may lead to better financing terms. Early discussion also helped reduce the time needed for assessing non-concessional borrowing cases; iii) the NCBP’s flexibility permitted a range of IDA responses in the cases considered but could be enhanced; and iv) further outreach 6 efforts were needed to enhance awareness of the rules and elements warranting the consideration of exceptions to the policy. 19. The report retained the key features of the policy and introduced adjustments to the implementation arrangements to strengthen its flexibility, enhancing country ownership of debt management and streamlining the NCBP process. These adjustments relate to three areas: i) setting of debt limits, where a differentiated approach, based on the countries’ macroeconomic and public financial management capacity and debt vulnerability, would be applied in setting such limits. The enhanced flexibility under this approach would be applied only at the authorities’ request; ii) streamlining the WB’s decision process, in which the decision level, within IDA Management, to respond to NCB cases will be determined following a risk-based approach; and iii) enhancing the communication of NCBP decisions to creditors and borrowers. 20. The 2015 Review and Update provided detailed information about NCBP cases reviewed by IDA Management since 2010 and took stock of the lessons learned. It retained the key features of the policy and introduced further adjustments to the NCBP to harmonize it with the updated IMF DLP and augment transparency. Adjustments made included: i) a joint IDA-IMF streamlined capacity assessment, which narrows the focus from a wide range of debt and public financial management indicators to the authorities’ ability to record and monitor external PPG debt in a timely manner; ii) to enhance further the flexibility of the policy, countries at low or moderate risk of debt distress that have adequate capacity may request ceilings on total external PPG debt in present value terms; iii) in addition to regular Board notes on IDA NCBP measures as well as detailed descriptions of all NCBP cases in Board updates, IDA measures would be reported as part of OP3.10 Annex D, comprising a table with aggregated loan information starting on July 1, 2016. Annex D is a public document. In addition, a real-time table with agreed ceilings and IDA decisions would be established on the IDA’s external website. 21. At inception of the NCBP, recommendations on the response to breaches of the policy were made by representatives from across the Bank and were discussed and agreed upon at the Operation Committee (OC). The representatives at that time included the Regions, Concessional Finance and Global Partnerships (CFP), Poverty Reduction and Economic Management (PREM), Operations Policy and Country Services (OPCS) and Development Economics (DEC), who met annually or as needed to review countries’ NCB. The OC was chaired by the relevant Managing Director (MD) and consisted of the Bank’s operational Vice Presidents. The Board was subsequently informed about management decisions through a Board note, and text that was included in relevant strategy papers or budget support operations. In 2010, given the significant transaction costs of the OC arrangements and the lengthy decision-making process involved, the OC step was dropped, and the representatives (who became known as the NCBP Committee), proposed recommendations to the CFPVP, and subsequently to the relevant MD for approval. The MD responsible for the Region concerned could vet the NCBP Committee’s decision, and if there were concerns about the impact of the recommended response the MD would convene a meeting at vice-president level (CFPVP, OPSVP, DECVP, PRMVP, and Regional VPs) including other networks when appropriate. The Board continued to be informed of NCBP decisions. The current governance arrangements have adapted to changes in the WB and the high- profile nature of some of the cases. The NCBP Committee is now chaired by the Director of IDA Mobilization and IBRD Corporate Finance (DFCII), and includes representatives from Macroeconomics, Trade and Investment (MTI), OPCS, Legal (LEG), Chief Risk Officer (CRO) 7 and DEC, and often more senior level staff including practice managers, and directors are in attendance at these meetings. DFi serves as the Secretariat of the Committee. The NCBP Committee continues to make recommendations based on country- and project-specific considerations to DFiVP for approval. The DFiVP provides the Chief Executive Officer (CEO) with recommendations for final agreement, after which the terms and volume of IDA allocations are adjusted as agreed. The Board is subsequently informed of Management’s decision. III. REVIEW OF NCBP IMPLEMENTATION 22. This section provides a review of country cases since the 2015 Review and takes stock of implementation experiences since the policy was introduced in 2006, with a focus on the revisions introduced in the 2015 Update. In particular, the Review assesses the implementation of the NCBP mainly to determine: i) if it was implemented as intended; ii) the extent to which the NCBP has been effective in meeting its objectives; and iii) adjustments needed including in response to the changing development finance landscape. Box 1. A Recap of Background Papers on Debt Vulnerabilities in IDA Countries World Bank staff prepared three papers over the past 12 months that analyze public debt dynamics and developments in IDA countries and set out potential policy options to address debt vulnerabilities. These papers have served as critical inputs to this Review and to the development of the forthcoming Sustainable Development Finance Policy (SDFP). The paper1 on debt vulnerabilities that was presented at the 2018 International Monetary Fund-World Bank (IMF-WB) Annual meetings in Bali, Indonesia, aimed at promoting informed, inclusive discussion on debt vulnerability issues in IDA countries and possible IDA policy options to be considered during the IDA18 Mid-Term Review. The paper provided: i) an overview of the growing indebtedness in IDA countries; ii) a description of key pillars of the joint WB-IMF multi-pronged approach (MPA) to help countries address debt vulnerabilities; and iii) an outline of the sustainable lending framework IDA has been implementing to help countries better manage and mitigate debt risks. A follow-up paper2 was prepared for the IDA19 Replenishment meeting in April 2019. The primary focus of that paper was on the review of IDA’s allocation and financial policies, i.e., those related to the Performance-Based Allocation (PBA) system, the Grant Allocation Framework (GAF), and the Non-Concessional Borrowing Policy (NCBP) as well as on options for change in these areas. The paper also outlines the key components of the broader agenda for helping IDA countries address debt vulnerabilities, including an increased focus on debt and policy drivers of debt vulnerabilities in country programs (through improving debt transparency, strengthening debt management capacity and enhancing domestic resource mobilization). The preliminary findings of the NCBP review, which was in progress in tandem with the IMF review of its debt limit policy, were also presented in this paper. A third paper that took account of the advice and guidance provided by Participants at the April 2019 IDA Replenishment meeting was presented for discussion at the June 2019 IDA Replenishment meeting. This paper lays out concrete policy options to adapt and expand elements of IDA’s allocation and financial policies to support financial sustainability and achieve country develo pment goals while minimizing risks of debt distress. The paper made three main proposals: i) continued implementation of the WB-IMF MPA for Addressing Emerging Debt Vulnerabilities; ii) introduction of new policy commitments under the IDA19 Special Themes of Jobs and Economic Transformation (JET) and Governance and Institutions (G&I) that promote sustainable economic growth and improved debt management policies; and iii) replacing IDA’s current NCBP with a broader SDFP . In addition to the updated NCBP review, the basic structure of the forthcoming SDFP was presented in that paper and following Participants’ inputs, additional information, including the principles underlying the SDFP, would be provided as an Annex in the 2019 Deputies Report. 1/ IDA. 2018. “Debt Vulnerabilities in IDA Countries”. IDA/SecM2018-0213, October 4. 2/ IDA. 2019. “Debt Vulnerabilities in IDA Countries: Policy Options for IDA19”. IDA/SecM2019-0084, March 28. 3/ IDA. 2019. “Addressing Debt Vulnerabilities in IDA Countries: Policy Options for IDA19”. IDA/SecM2019-076, July 26. 8 23. The Review was informed by the record of experience and country cases, surveys of country teams and analytical work. Extensive consultations were undertaken with various groups of stakeholders: i) WB country teams regarding their experiences in implementing the NCBP and views on potential adjustments needed to enhance the effectiveness of the framework; ii) IDA borrowers on the NCBP experience in assisting them address increased debt vulnerabilities; iii) MDBs on identifying key common principles and coordinating actions to enhance sustainable lending; and IMF together with other IDA stakeholders. The Review was also informed by three papers that were undertaken to analyze public debt dynamics and vulnerabilities in IDA countries. These are summarized in Box 1. A. Update on Country Cases since the 2015 Review 24. The number of countries covered by the NCBP declined from 43 in FY15 to 40 in FY19. Three countries, Côte d’Ivoire, Lao PDR and Zambia, exited the NCBP because they reached Gap status. In FY15, 68 percent of NCBP countries were at moderate and low risk of debt distress (49 and 19 percent, respectively). From about a third of countries at high risk or in debt distress in FY15, the share jumped to half in FY19. Only 10 percent of NCBP countries are now at low risk of debt distress. Of the 40 countries currently subject to the NCBP, 15 countries have an IMF program as of May 2019, and of these, six had a non-zero ceiling while one, Senegal, had no debt limit. 25. The number of country cases discussed in the context of the NCBP has increased since the 2015 NCBP Review, with 17 cases in 6 countries (Table 1). Several countries, such as Ethiopia and the Maldives, have been reviewed by the NCBP Committee multiple times since 2015. Country and loan-specific circumstances indicated that NCB was part of an adequate financing plan in three cases pertaining to three countries: Ethiopia, Tajikistan, and Comoros. Most waivers were granted ex-post, and for projects such as electricity, water, and transport infrastructure. IDA financing terms were hardened for three countries: Maldives, Ethiopia, and Mozambique. Mozambique’s IDA allocation was reduced by 10 percent in FY18 and again in FY19. Volume reductions were deemed more appropriate than the alternative of hardening terms given the country’s high risk of debt distress and limited debt management capacity. The grant portion was partially converted to regular IDA credits for Ethiopia in FY19, and for Maldives in FY18 and FY19. 26. Individual NCB cases reviewed by IDA since 2015 are as follows: • The Maldives (2016-2019): Between June 2015 and September 2016, the Government of the Maldives contracted eight loans and issued a bond for US$ 1.0 billion (about 29 percent of the country’s 2016 GDP) on non-concessional terms. The loans had a combined grant element of close to zero percent. The Maldives is an upper-middle income small island state, highly vulnerable to natural disaster and climate change. To address these risks, the Government has undertaken large debt-financed investments aiming to relocate the population from the vulnerable islands and atolls to larger islands in Greater Malé and to improve service delivery while also addressing constraints to growth in the tourism, such as the expansion of the airport runaway. Maldives signed and guaranteed further non- concessional loans in the amount of US$1.0 billion (25 percent of 2016 GDP) in FY18 to support its large effort to consolidate its population around the capital city, a strategy that 9 could reduce the cost of service delivery while supporting the jobs and climate change agenda. Following the breach of the policy by contracting large non-concessional borrowing IDA decided to harden the country’s terms from 100 percent grants, given the country’s high risk of debt distress, to a 50 percent grant and 50 percent credit basis in FY18 and FY19. Based on country and loan specific factors, management decided to maintain the hardening of the terms in FY20. The Government of the Maldives has taken steps to increase transparency and accountability and made debt statistics publicly available. The Ministry of Finance (MoF) has published its fiscal and debt strategy for 2019-21, which includes a Medium-Term Fiscal Strategy and Medium-Term Debt Management Strategy. The MoF also began publishing its semi-annual public debt bulletin, which provides the disbursed and outstanding debt of the government including guaranteed and on-lent loans, external and domestic debt breakdowns, and summary of debt statistics including quarterly fiscal developments, quarterly reports for State-Owned Enterprises (SOEs). Updated records on sovereign guaranteed debt are also available in the MoF website. • Ethiopia (2015-2019): Ethiopia has been under IDA NCBP review annually since 2013. In FY13, IDA established NCB ceilings of US$1 billion for FY13 and, in principle for FY14 and FY15. However, the Government of Ethiopia breached this ceiling by contracting NCB amounting to US$5.8 billion (12 percent of GDP) in FY13 and US$2.2 billion (4 percent of GDP) in FY14 to finance a broad range of infrastructure investment programs. The combined grant elements of these loans were 18 and 12 percent in FY13 and FY14, respectively. In parallel, the risk of debt distress shifted from low to moderate in FY16. IDA responded to the policy breach by converting the grant portion of the Performance-Based Allocation (PBA) into regular IDA credits. Without the response, Ethiopia would have received 50 percent of its allocations on grant terms and the other 50 percent on credit terms given its moderate risk of debt distress. In addition, IDA applied a five percent volume reduction to Ethiopia’s FY15 IDA allocation. The Government complied with the NCB ceiling in FY15, having issued a US$1.0 billion 10-year Eurobond consistent with the NCBP ceiling. However, due to debt sustainability considerations, the NCB ceiling was reduced to US$750 million for FY16 (1.1 percent of GDP) and further down to US$400 million for FY17 and FY18 (0.5 percent and 0.3 percent of GDP, respectively). The reported NCB contracted in FY16 and FY17 was substantively within the NCBP ceilings. However, in FY17, the Government breached the NCBP ceiling by a wide margin. There were also delays in the reporting of an additional NCB contracted in FY17 amounting to US$1.14 billion. Meanwhile, Ethiopia’s risk of debt distress was downgraded from moderate to high, and the NCB ceiling was reduced to zero. In response to the breach of FY18 NCBP ceiling and reporting delays, IDA hardened Ethiopia’s terms by converting 50 percent of the grant portion of its IDA allocation into credit terms in FY19. For FY20, while commending the authorities for not contracting any NCB in FY19, Management maintained the responses in place in FY19 given the history of multiple breaches and short record of adherence to the NCBP ceiling. Ethiopia has rolled out a robust set of reforms likely to improve transparency and debt management. The MoF issued the Public Debt Management and Guarantee Issuance 10 Directive in January 2017 which considerably strengthened monitoring of NCB by SOEs. Moreover, restructuring in October 2018 placed the former Ministry of Public Enterprises under the MoF, which is expected to further enhance its monitoring of SOE activities, including borrowing. The Government has also committed to refrain from taking on new NCB and has expanded the scope of its existing debt-reporting arrangements, supported by the WB’s recent Development Policy Financing (DPF) operation. • Mozambique (2016-2019): Between 2009 and 2014, Mozambique contracted, with no disclosure, NCB of US$1.3 billion by issuing guarantees to state controlled companies and through direct borrowing from lenders. The debt comprised two guarantees for loans contracted by commercial companies with state equity participation amounting to US$1.16 billion, as well as direct loans of US$133 million from bilateral creditors. This debt was about 10 percent of GDP and was not previously disclosed to the WB and the IMF. This NCB by Mozambique breached the US$1.5 billion ceiling under the IMF program in place since 2013 and shifted Mozambique’s risk of debt distress from moderate to high. IDA’s response was to convert the grants that Mozambique would have received for FY17 as a high-risk country into regular IDA credits and to apply a 20 percent volume reduction. In March 2017, Mozambique contracted a further US$138 million in NCB to finance the migration from analog to digital broadcasting. Considering this borrowing and given the earlier undisclosed NCB, IDA applied a 10 percent volume reduction to Mozambique’s PBA in FY18, while maintaining 100 percent grant terms consistent with the country’s debt risk rating. In September 2018, IDA reviewed Mozambique again, noting that the country did not contract any new known NCB in the review period. However, the review also noted that limited progress had been made with respect to resolving debt defaults and restoring debt sustainability. The review also noted that the external audit of the previously undisclosed debt fell short of expectations due to significant information gaps. In view of these developments, IDA decided to retain the same response approved for FY18 (10 percent volume reduction to FY19 PBA, with IDA financing on 100 percent grant terms) in FY19. Considering good progress in pursuing accountability regarding the non- disclosed debts, enhanced reporting and steps to close debt related legal and regulatory loopholes, as well as no new NCB, IDA Management decided to lift Mozambique’s volume discount starting in FY20. Mozambique continued to work closely with technical assistance (TA) from the WB and the IMF to strengthen debt management, to reform the governance of SOEs, and to improve public investment management in response to the hidden debt crisis. Regulations adopted to strengthen the management of public debt and guarantees in December 2017. The Government approved a new SOE law in 2018 and supporting regulations in 2019. A new public investment management system has been introduced. The WB and the IMF supported the Government’s actions to implement fiscal measures, increase transparency in financial markets and modernize the foreign exchange market. The Government made some progress in restructuring the MOZAM 2023 bond which was issued to finance the EMATUM (Empresa Moçambicana de Atum) Tuna Company in 2013 by reaching an agreement in principle with bondholders. The Government has pursued transparency and accountability regarding the hidden debt cases, including through the launch of high-profile legal proceedings in Mozambique and in foreign countries. The Government also took 11 steps to legally challenge and obtain compensation for the SOE (Proindicus and Mozambique Asset Management) guarantees. In May 2019, Mozambique requested an ex-ante waiver for US$4.6 billion (36.4 percent of GDP) for planned borrowing and issuance of a sovereign guarantee to the state-owned hydrocarbons company, Empresa Nacional de Hidrocarbonetos de Moçambique (ENH) to participate in the Liquified Natural Gas (LNG) projects investment. While the LNG production in Mozambique is expected to be transformational for economic growth and fiscal revenues, enabling greater investment in human capital and physical infrastructure and enhanced prospects for future economic development, the participation of ENH was believed to help attract international investors to catalyze the development of gas sector in the country. After careful deliberation, Management approved to grant the waiver request conditional to fulfill ex-ante and ex-post conditionalities. These include publishing amount and terms of guarantee, publishing ENH financial statement, continued disclosure of ENH annual audited financial statement and debt contracted for all LNG projects, regular reporting of sovereign guarantee, and inclusion of this information in the annual State’s General Account and shared to parliament. • Tajikistan (2017-2019): In September 2017, Tajikistan issued a Eurobond worth US$500 million (7.2 percent of 2016 GDP) maturing in 10 years and priced at 7.125 percent. The loan was contracted to finance the construction of the Rogun hydropower plant. The issuance of the Eurobond led Tajikistan’s risk of debt distress to shift from moderate to high in November 2017. After considering several project and country specific factors (including an external shock that depressed commodity prices and remittance inflows, weak policy and institutional performance, banking sector issues, and the project’s prospect for export earnings) and lack of concessional financing alternatives, IDA Management granted a waiver in FY18. The WB has provided TA to the MoF in Tajikistan through a trust fund from the Debt Management Facility (DMF). These activities included a follow-up DeMPA, the formulation of a Reform Plan (FY17), and the Debt Management Strategy (DMS FY18). The WB and the IMF jointly provided TA on developing a Medium-Term Debt Strategy (MTDS) followed by government approval of the MTDS for 2018-20 in December 2018, taking into account policy recommendations of the joint WB-IMF TA. The MoF also requested support from the WB and other contributors to upgrade its staff capacity and skills in public debt management. The FY19 annual NCBP review confirmed that Tajikistan was in compliance with the NCBP Committee recommendation as no new NCB was contracted. • Comoros (2018-2019): In July 2018, Comoros signed a two-tranche non-concessional loan agreement in the amount of Euro 40 million (equivalent to about 8 percent of GDP) with the Eastern and Southern Africa Trade and Development Bank (TDB) to finance the rebuilding of the El-Maarouf Hospital. The disbursement of the first tranche of the loan (25 million euros) had a grant element of zero percent. IDA management didn’t grant a loan-specific waiver as the loan was deemed costly and excessive, which may exacerbate the country’s risk of debt distress. In addition, there were gaps in the information on the financial viability and financing plan of the project that was provided. Instead, 12 Management recommended that the Government of Comoros should consider requesting a modest non-zero NCB ceiling. In May 2019, the Government of Comoros requested a non-zero NCB ceiling of Euro 25 million. The ceiling was approved along with recommended specific policy actions including improving debt management and transparency, limiting the size of external arrears and engaging in policy dialogue with the WB on the prioritization of investments within the framework of competing needs and limited resources. B. Implementation Assessment 27. This subsection considers various issues concerning NCBP implementation with the primary objective of assessing whether the NCBP was generally implemented as intended. It considers experiences since the policy’s introduction, focusing on the period since the 2015 Review. Overall the assessment found that the NCBP was broadly implemented as intended with the joint IMF-WB Debt Sustainability Framework for Low-Income Countries (LIC DSF) as a key input and has provided a flexible framework to assess the impact of and rationale for NCB based on country and project-specific factors. The NCBP and IDA’s responses have complemented or fostered IDA country dialogue and programs, along with influencing countries to seek alternative sources of concessional financing. 28. Implementation experience shows that the establishment of debt limits/ceilings was informed by the joint IMF-WB’s LIC DSF, as intended. As the policy intended, countries at high risk of debt distress have a zero NCB ceiling but can borrow non-concessionally under exceptional circumstances on a loan-by-loan basis (a waiver is required before borrowing). Those at moderate and low risk of debt distress have the option to request NCB ceilings in addition to exceptions on a loan-by-loan basis. The nature of these borrowing ceilings also depends on the level of capacity, as discussed in detail in the 2015 Review.11 Experience shows that countries’ risk of debt distress as assessed under the LIC DSF has informed NCBP implementation, including the setting of debt ceilings. 29. Requests for setting ex-ante non-zero debt ceilings for countries at low and moderate risk of debt distress have been limited to two countries. Ceilings can be helpful as these encourage careful planning of borrowing decisions and provide comfort up front that borrowings would not breach the NCBP. Since 2013, nominal ceilings on external NCB were established for Ethiopia which was then assessed to be at moderate risk of debt distress. Ethiopia’s NCB ceiling was subsequently reduced to zero after the country moved to high risk of debt distress in 2017. In addition, in 2019, a nominal ceiling on external NCB was also established for the Comoros. 30. The limited number of requests for ceilings could be due to a combination of factors. The survey and consultations with country teams revealed that the level of awareness on the NCBP is relatively limited in some countries. For example, they cited that there is no access to clear and concise material which explained how NCBP works and what the reporting obligations are. They also were of the view that submitting debt data to the Debtor Reporting System (DRS) fulfil the NCBP debt reporting obligations. The limited request for ceilings may also be because some non- zero ceiling eligible countries had their ceiling set under the IMF’s DLP for those with IMF 11 IDA. 2015. “IDA’s Non-Concessional Borrowing Policy: Review and Update”. IDA/SecM2015 -161, October. 13 programs. Of the 25 NCBP waivers recorded since 2007, 10 were for countries with IMF programs that included conditionality under the DLP. The NCBP Committee responded to all cases that were brought before it, although an analysis of DRS data12 showed that there were cases of countries undertaking NCB without prior ceiling or loan-by-loan exception requests – and these are being addressed.13 31. The responses to the breaches under the NCBP followed the recommendations of the policy as initially envisaged in the 2006 policy paper considering country circumstances. The 2006 Board Paper proposed that in most cases volume reduction would be more suitable when debt sustainability is fragile; while hardening terms would be more suitable when debt risk is low, although a flexible approach was intended from the policy inception. An analysis of the inventory of country cases in Table 1 shows that most responses (12 of 15) have involved hardening of terms, including in cases where the country in question was at moderate (e.g., Ethiopia FY16-17) or high risk of debt distress (e.g., Maldives FY18-19). The decisions were justified in terms of the flexibility provided from the policy inception, and the importance of maintaining the volume of IDA financing in most cases, notably countries with elevated debt risks. In some cases, this may have been out of a concern that affected countries might further turn to non-concessional sources to compensate for reduced IDA allocations. Consultations with country teams also indicate the country authorities’ preference to maintain IDA volumes. 32. Waivers were requested and granted to several countries, reflecting the policy’s acknowledgement that there are situations where NCB can be appropriate. Of the 51 country cases presented in Table 1, twenty-five (49 percent) resulted in waivers and 11 were granted in line with IMF waivers under the DLP. When the IMF grants waivers under the DLP, the NCBP Committee typically followed suit to maintain consistency with the IMF. The decisions to grant waivers are considered carefully by the NCBP Committee. They are made in terms of the policy which makes provision, even for countries at high level of debt distress, for borrowing non- concessionally under exceptional circumstances. One example of such exceptional circumstance could be in the financing of large initial investments in projects with potentially high risk-adjusted rates of return and in this regard, various country- and loan-specific factors are considered (See Annex 6 for details). The high percentage of waivers granted is reflective of these considerations, and it is important to note that there have been fewer waivers granted since the 2015 Review. 33. Regarding coordination with the IMF, where countries are not under an IMF program, IDA takes the lead in establishing debt limits on external NCB, in consultation with the IMF country teams. For IDA-eligible countries under an IMF program, the IMF takes the lead in setting debt limits following discussions with the WB country teams. Borrowing limits under the NCBP in such circumstances have been aligned in principle with the borrowing plans as well as any applicable debt limits envisaged under the IMF program. The case for continued coordination remains strong, particularly given the recommendation to extend the coverage of the new policy to all IDA countries. 12 The challenge in the analysis of DRS data was that it used a 25 percent grant element as a concessionality benchmark whereas NCBP used 35 percent. However, this has changed as DRS now uses 35 percent as a concessionality benchmark. 13 Through country teams the authorities of countries with unreported NCBP cases were contacted and where appropriate, cases were submitted to the NCBP Committee ex-post. 14 34. The NCBP 2015 Review highlighted enhancements to the NCBP’s implementation arrangements including: (i) streamlined capacity assessment, (ii) adjustments to internal processes, and (iii) enhanced transparency and reporting. The capacity assessment exercise, which is conducted once a year, shows that as of February 2018, the quality of debt monitoring in 32 out of 70 assessed countries was weak while 38 had sufficient quality. The situation was reversed by February 2019 with 38 countries with weak quality of debt monitoring and a decline to 32 countries with sufficient debt monitoring quality. The capacity assessment focuses on a country’s ability to adequately capture and monitor the contracting and disbursement of all new public external loans and it is based on the Country Policy and Institutional Assessment (CPIA), published DeMPA and Public Expenditure and Financial Accountability (PEFA) indicators. While detailed changes in each indicator are not available, movements on capacity measurements reflect changes in these indicators. 35. The NCBP governance arrangements remain adequate although there is room for further streamlining the decision-making process and considering how this would be adapted to the new SDFP. Progress has been made in streamlining internal processes to enable faster decision-making in responding to cases of non-compliance or requests for waivers. However, even the recent review processes have taken longer than expected given the various steps needed before final clearance, and in some cases given the complexity of cases involved. A survey of country teams also indicated the gaps in communication between the NCBP Committee, country teams and authorities which further slows down the NCBP review process. While there is no recommended period within which cases need to be finalized, enhancing authorities’ awareness of the NCBP including through effective communication; and introducing a timeframe for finalizing cases, may help to further streamline the process and improve the policy’s effectiveness. 36. There have been some improvements in the NCBP’s transparency since the 2015 Review. While changes in IDA terms or allocation volumes resulting from NCBP non-compliance are publicly available14, consultation with country teams reveals that the information is not easily accessible to some country authorities and some creditors. Enhanced public disclosure of NCBP decisions would improve transparency and amplify the signaling effect to borrowers and creditors, thereby potentially making the policy more effective. Enhanced disclosure could also include publication of an annual report on country cases reviewed by the NCBP Committee and the status of debt reporting under the NCBP as part of the broader push to improve debt transparency under the IMF-WB MPA. The specific terms and conditions of disclosure would need to be carefully designed to protect applicable confidentiality requirements. 37. The current disclosure framework does not provide sufficient context for IDA countries and others to understand the rationale behind the NCBP Committee’s decisions on waivers and responses. The WB Board of Executive Directors is informed of NCBP responses immediately after a Management decision is made, but this information is not publicly available. Summaries of country cases reviewed by the NCBP Committee are publicly disseminated only in the periodic NCBP reviews and updates, which are available with a considerable lag. Consultations with country teams also revealed that brief and concise information on NCBP is not easily 14 Please see Bank Directive “Financial Terms and Conditions of Bank Financing ”. 15 accessible by the authorities, which is important given the high staff turnover rates in member countries. Table 1. List of IDA Countries with NCBP Responses by Year (as of August 2019) IDA FY of Hardening Volume Both (V&H) Waiver provided No Response/ cycle Decision** reduction Previous response discontinued IDA18 2019 Ethiopia Mozambique Maldives Tajikistan Comoros1 2018 Maldives Mozambique Tajikistan, Benin* Comoros*** Ethiopia IDA17 2017 Maldives Mozambique Ethiopia 2016 Ethiopia Mozambique 2015 Ethiopia IDA16 2014 Lao PDR Ethiopia Chad, Zambia Madagascar*, Sao-Tome & Principe*, Togo 2013 Burundi*, Comoros*, Chad, Ethiopia, Guinea, Kyrgyz* 2012 ‘ Zambia, Cameroon IDA15 2011 Burundi*, Ghana (2), Ethiopia Côte d’Ivoire* 2010 Ghana Chad Mauritania 2009 Congo Rep, Ghana Cameroon, DRC, Rwanda, Senegal IDA14 2008 Rwanda* 2007 Ghana Angola Mauritania*, Mali* Total 11 3 2 25 10 1Management didn’t approve the waiver request and subsequently a Euro 25 million ceiling was established. * Waivers granted to maintain consistency with IMF program. ** Fiscal year of decision of IDA response with application to the following year’s IDA allocation. *** Country request for waiver did not qualify and NCBP Committee suggested that country request a ceiling. 38. Further, the results from the survey of IDA country teams suggest that they often face difficulties in obtaining project-specific quantitative information, such as expected rates of return especially in projects where the WB is not involved. Thus, there exists a need to enhance transparency further, by expanding the information available from borrowers, including information on collateralizations and contract terms. 39. While the consultations suggest that the NCBP provided an external anchor that may have helped countries contain NCB, country teams also raised some implementation challenges that the new policy should address including: 16 ▪ The NCBP’s lack of a clear exit strategy - no clear guidelines regarding the conditions required to lift the responses. In the past two fiscal years or so, however, there was an improved articulation of conditions required for responses to be lifted. In this regard, the NCBP Committee began setting agreed actions with the countries which set expectations on exit strategies. This has in part informed the performance and policy actions envisaged in the new SDFP. ▪ While the new LIC-DSF guidelines provide enhanced guidance on efforts to include SOEs debt, some country authorities raised concerns regarding the challenges they face in collecting SOEs debt. ▪ Due to differences in fiscal years, the timing of the NCBP review poses challenges to some countries– critical data needed for the review is often not yet finalized at the time of the review. ▪ Limited awareness about the NCBP process is sometimes due to the lack of clarity on how NCBP works and the basis of NCBP Committee recommendations. They suggested continued process of communication especially given the frequent staff turnover. This would be critical as the forthcoming SDFP would also apply to these countries. ▪ In certain countries, political pressures complicate discussions on debt sustainability. ▪ The coverage of non-concessional debt needs to expand to include medium and long-term foreign deposits in central banks. Even though this was part of the LIC-DSF, it was not considered under the NCBP until last year. ▪ Expanding the NCBP scope to include Gap countries - some country teams with Gap countries in their portfolio indicated that applicability of the NCBP to their countries would have made their dialogue with the Gap countries authorities much easier. 40. The analysis of the inventory of country cases, excluding those covered by the IMF DLP, provided additional insights on implementation of the NCBP. These include: ▪ The NCBP reviewed only cases that were brought to the attention of the NCBP Committee, to either request a waiver, recommend a response in the case of non-compliance and non- granting of a waiver, or to request a debt ceiling. The new policy will review all countries and more systematically. ▪ The NCBP implementation dealt with NCB mostly on an ex-post basis, while a forward- looking engagement with countries was needed. Such engagement will be done under the SDFP. ▪ There was a need to better integrate NCBP considerations into country programs, for instance through DPFs and Advisory Services and Analytics (ASA). Under SDFP the link will be strengthened considerably by focusing on policy actions emanating from the country programs. ▪ The risk of debt distress has been increasing steadily since 2013, implying that growing debt vulnerabilities have transcended beyond post-MDRI and grant-eligible IDA countries. This has been confirmed in several debt papers (See Box 1) which suggest that IDA’s role 17 in helping countries deal with debt vulnerabilities may need to be broadened to include all IDA countries. ▪ An average of fewer than four country cases were reviewed per annum under the NCBP over the 12-year period (Figure 2). At the same time the share of NCBP countries at higher risk of debt distress increased substantially. This raises the question of whether the low number of reviews was due to countries not reporting their NCB to the NCBP Committee. The Review did find some evidence of countries undertaking NCB without requesting a waiver or ceiling. In FY19, potential gaps in reporting NCB were systematically analyzed. Countries have been reporting NCB to the WB’s DRS15 and information is also reported in the context of the DSAs. The intention under the NCBP was to encourage countries to report up-front planned NCB and either seek a waiver or ask for a ceiling (i.e., the approach was supposed to be ex-ante). However, in a number of cases, these loans were not brought to the attention of NCBP Committee. Ten such country cases have been identified over FY15-FY17, and of these, eight countries are at low or moderate risk of debt distress and two are at high risk. Figure 2. NCBP Country Cases, Responses and Debt Risk, 2007-2019 a. NCBP Cases and Responses b. NCBP Responses & Risk of Debt Distress (in percent) 18 6.0 100 11 16 90 24 5.0 11 14 80 70 12 4.0 4 26 60 10 3.0 50 100 28 8 40 6 2.0 30 4 53 1.0 20 36 2 10 0 0.0 0 2007-09 2010-12 2013-15 2016-19 Low Moderate High waiver hardening Volume cut Harden+Volume Others Avg cases (RHA) waiver hardening Volume cut Harden+Volume Others Source: World Bank staff calculations. ‘Others’ include ceilings, no response and updates ▪ Consultations with country teams suggest that some reporting issues result from countries transitioning to or from IMF programs. In such a case, for instance, the country may still be under the impression that the non-zero ceiling under such an IMF program is still in place; in other cases, countries do not have NCB ceilings, and reporting obligations to the NCBP Committee are not clear. 15 IDA’s general conditions (General Conditions for IDA Financing, IDA, July 2017) oblige all client countries to report long - term external debt on a loan-level basis through the Debtor Reporting System (DRS). Specifically, countries must report quarterly and annual data providing information on new debt (quarterly data is to be reported within one month after the end of each quarter). 18 C. Effectiveness of the NCBP 41. Public debt levels in IDA countries have risen substantially over the past five years. Having fallen steadily for many years on the back of strong growth and debt relief, starting from 2013, median public debt in IDA countries increased by 14 percentage points of GDP to reach about 50 percent of GDP in 2018 (see Figure A4.1, Annex 4). The increase in public debt levels was broad-based across IDA-eligible countries (see Figure A4.2, Annex 4), but slightly larger among commodity exporters. The increase in debt levels for countries affected by Fragility, Conflict and Violence (FCV) was slightly lower than for countries not affected by conflict (nine percentage points, compared to 13 percentage points for non-FCV countries), while Small States experienced a modest increase. 42. A decomposition of public debt dynamics across IDA countries points to several potential reasons for increased debt vulnerabilities. These include: i) weaknesses in fiscal policy frameworks (including, weak domestic resource mobilization and efficiency of public expenditures); ii) changing composition of debt towards more expensive and riskier sources of financing; and iii) weak debt transparency. In addition, several countries were affected by internal conflict (Yemen and Burundi experienced increases in the public debt-to-GDP ratio) or by shocks (Ebola epidemics in the case of Liberia and Sierra Leone – where the public debt-to-GDP ratio increased by about 20 percentage points between 2013 and 2017). 43. Greater reliance on new sources of financing compounded with weaknesses in debt management, led to increased debt service and refinancing risks. To respond to their enormous financing needs in a context of scarce concessional resources, IDA countries have increased their reliance on financing from non-traditional sources. Since 2010, 18 IDA countries have issued a growing amount of international bonds, thanks to the low interest environment and the search for higher yields from international investors. Commodity dependent countries, including some Heavily Indebted Poor Countries (HIPCs) and Fragile and Conflict-Affected Situations (FCS), constitute the bulk of issuers. At the same time, non-Paris Club creditors have become a more important source of financing over the past decade, especially in commodity-dependent Sub- Saharan African countries. In 2017, non-Paris Club debt accounted for about one-fifth of the total PPG external debt in IDA countries. 44. Despite increases in public debt and a shift towards riskier debt portfolios, debt management capacity in IDA countries remains weak. Areas of particular concern include: weaknesses in debt management governance; weaknesses in public financial management and regulatory frameworks for domestic borrowing, loan guarantees, on-lending and derivatives; lack of operational risk management; and insufficient staff capacity in debt management offices to adequately assess fiscal and debt risks and deal with a diverse and fragmented landscape of investors and emerging creditors. In addition, and despite significant improvements in debt data, IDA countries public debt statistics suffer from limited debt data coverage and debt transparency. This is especially the case for debts to SOEs, contingent liabilities related to Public Private Partnerships (PPPs), and collateralized debt. At the same time, the public sector is one of the most heterogenous categories in terms of variety of definitions – individual countries differ in regard to the degree of centralization or federalization, and the corresponding budgetary and regulatory arrangements. 19 45. These changing circumstances complicate assessments of the effectiveness of the NCBP. Efforts to undertake quantitative analysis have important caveats. Methodologies in these types of analysis suffer from well-known shortcomings, including their inability to establish causality between variables and the fact that rigorous counterfactuals cannot be constructed given the lack of a control group. Moreover, debt sustainability is a complex function of many endogenous and exogenous factors such as exchange rates movements and natural disasters, which makes it difficult to isolate the impact of the NCBP on debt sustainability. 46. With these caveats in mind, the effectiveness of the NCBP was assessed by analyzing its impact on debt sustainability and on the level of NCB. The analysis of the impact of the NCBP on debt sustainability focused on the risk of external debt distress derived from the joint WB-IMF LIC DSF. It compares the risk of external debt distress for the 40 countries currently subject to the NCBP with a sample of 28 IDA countries covered by the LIC-DSF over the 2006- 2018 period and to which NCBP does not apply. The share of NCBP countries at high risk of debt distress or in distress fell from 2006 to 2014 and then rose sharply, from 25 percent in 2015 to 50 percent by 2018. This outcome tracked very closely the evolution for the 28 non-NCBP IDA countries for which LIC-DSF debt risk ratings are available. This analysis shows that the 40 NCBP countries, as a group, do not appear to have performed better on aggregate than the sample of IDA countries not subject to NCBP - which would suggest that NCBP may have had limited impact on containing debt vulnerabilities in NCBP countries. However, this should be interpreted with caution. The counterfactual of comparing NCBP countries with other IDA countries not covered by the Policy may not present the full picture, as NCBP countries tend to have weaker institutional capacity. 47. To analyze the impact of NCBP on the level of NCB, this Review compares the change in the share of concessional debt in the total stock of external debt for the 40 current NCBP countries with the change in the share for other IDA countries not subject to NCBP, over the 2007-2018 period. Data on the share of concessional debt is available through the WB’s DRS. The change in shares as depicted in Figure 3 shows that, although the dynamics between NCBP countries and non-NCBP countries were broadly similar between 2007 and 2014, those under the NCBP saw less of a decline than non-NCBP countries from 2015 – the shares remain relatively flat for NCBP countries whereas non-NCBP countries continue to decline, albeit modestly. This suggest that NCBP may have had an intended positive impact on NCB notwithstanding a declining concessional financing environment. This is consistent with the results from the surveys and consultations which indicate that NCBP played a positive role in promoting debt sustainability in some countries. 20 Figure 3. Share of Concessional Loans in Total External Debt (%) 65 60 55 50 45 40 35 30 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018p Post-HIPC NCBP Non-NCBP 2018p represents preliminary data for 2018. Source: Debtor Reporting System 48. Despite the shortcomings in the analytical methodology, on aggregate, the analysis suggests that NCBP may have had a positive but limited ability in incentivizing borrowers towards concessional borrowing. This finding is consistent with those made in the 2010 and 2015 NCBP Reviews. This limited effect appears to be the result of a complex and interlocking series of factors. It could be an indication that NCBP’s implementation, which focused mostly on external NCB, is no longer consistent with the changing landscape where broader factors, such as exchange rates fluctuations and low levels of revenue mobilization, are driving external debt vulnerability and domestic debt is rising. In addition, it could also be the case that given these broader factors driving external debt vulnerabilities, the available responses (volume reduction and changing of terms) are no longer sufficient to meaningfully counteract the general increase in vulnerabilities. This would point towards the need for: i) broader coverage of public debt, including all domestic public debt, guarantees and other contingent liabilities, and ii) additional instruments/ performance and policy actions that would address the causes of debt vulnerabilities. Figure 4. Factors that have constrained the impact of NCBP: Country Team Survey Results 70 60 50 40 30 20 10 0 Limited information on Lack of comprehensive Administrative Other debt terms borrowing plans procedures Source: World Bank estimates 2019 21 49. The country team surveys and consultations also revealed a number of factors that may have constrained the impact of the NCBP on IDA countries borrowing behavior. Among these factors, country teams highlighted limited information on the terms of NCB and lack of comprehensive borrowing plans by country authorities (Figure 4). Other factors cited include limited government understanding of the NCBP, political influences on borrowing decisions, and timing mismatches between NCBP monitoring and debt data availability. These factors may help explain why the NCBP’s utility in dialogues at country level does not appear to have translated into discernable impact on aggregate debt sustainability indictors in the quantitative analysis. These factors need to be addressed in the new policy including through continuous dialogue with client countries on how they can leverage other potential financial resources from the WB (and other MDBs) such as: i) leveraging IDA resources through syndicated financing16 in the context of PPPs; and ii) utilizing WBG products such as commodity hedges to manage the volatility of price fluctuations in key commodities (i.e., agricultural crops, oil, etc.) and disaster risk financing and insurance. 50. Results from the survey of country teams suggest that the NCBP provided an external anchor that helped drive down NCB in certain countries, despite certain challenges which they cited. In particular, the results suggest that: i) the NCBP and NCBP responses have complemented or fostered IDA country dialogue and programs which are the WB’s primary instruments to help countries address debt vulnerabilities; ii) the NCBP has influenced countries to seek alternative sources of concessional financing (Figure A2.3) and helped some countries avoid instances of NCB and/or improve the terms of financing packages; iii) the NCBP has helped promote better debt strategy, management and reporting practices in certain country contexts; and iv) NCBP was most effective when there was no IMF program or WB DPF in place. While most country teams surveyed noted that the NCBP has played a role in debt sustainability through country dialogue, less than half noted that it played a role through the disincentive effect of the NCBP responses (Figure A2.3). This suggests that a majority of country teams surveyed believe that responses had limited impact on borrowing decisions. 51. The findings from analytical work, as well as results from consultations and surveys reinforce the findings of the 2015 NCBP Review on the limits of the policy’s ability to affect borrowing decisions thereby influencing the direction of debt vulnerabilities. This suggest that updates necessary to improve the effectiveness of the NCBP, in the context of changing development landscape and the need to broaden the scope of the NCBP, warrant a transition to a broader SDFP. The new policy would represent a significant step towards addressing the vulnerabilities in IDA countries and would consider all inputs from various stakeholders, including Deputies, country teams, borrowers and other MDBs. D. Creditor Coordination and Outreach 52. An assessment of the Outreach Program implementation shows that, overall, the NCBP’s impact on creditor coordination has been effective but limited in scope. Multilateral and traditional bilateral creditors which were at the core of the global coalition for the HIPC 16 For example, in an infrastructure project, IDA can support the necessary initial government investment in the construction of a project in a syndicated financing whereby commercial lenders can take on some of the upfront financing and the financing of the operational phase in a PPP structure. 22 Initiative and MDRI are responsive and engaged with the WB on the need for sustainable lending practices. However, outreach with non-traditional bilateral and private creditors that have continued non-concessional lending to IDA countries including those at high risk of debt distress, need to be strengthened. For bilateral creditors, expanding membership in the Paris Club could be an important step in strengthening engagement on the NCBP or its successor SDFP. Table 2. List of Recent Outreach Events Engagement Events Specific Objectives Participants Timeline Type I. Building a 1. NCBP Discussed how effective NCBP has NCBP Committee February 13, 2019 common Committee been in achieving its overall members understanding objectives, strengthening the link on debt issues to the actual impact of NCB on in LICs and debt risk, and sought ideas on how exchange to streamline the internal review information process. 2. Seminar with Discussed IDA’s ongoing review IMF, AfDB, ADB, February 28, 2019 MDBs of policy options to address debt IFAD, IDB, EIB, vulnerabilities DFCII presented EC, CABEI MDBs survey result. 3. Consultation Consulted on effectiveness of Country economic February 27, 2019 with Country NCBP and its responses, other team in Southern Teams approaches to address debt Africa, Ethiopia, vulnerabilities. DFCII presented Sudan and South the CTs survey result. Sudan Country economic April 9, 2019 team in the Pacific Islands 4. Online Surveys • MDBs on Lending Practices • February 2019 • World Bank Country Teams on • February 2019 NCBP Implementation • March 2019 • Countries subject to NCBP on NCBP implementation (joint efforts with IMF on NCBP/DLP countries) II. 5. Consultation Presented the proposed IDA policy Côte d'Ivoire May 16-17, 2019 Consultations with IDA toolkit and consulted with Minister of Economy and feedback Borrowers and Cameroon, Côte d'Ivoire, Senegal, and Finance; Debt on the AfDB Uganda, and Zambia on how to Directors from proposed effectively adapt IDA policy Cameroon, Côte changes to the framework to better service client d’Ivoire, Senegal, IDA Policy countries, jointly hosted with Uganda, and Zambia Toolkit AfDB in Abidjan17 6. Virtual Distributed a one-pager on the MTI practice May 21-26 2019 consultation with forthcoming SDFP for comments managers MTI Practice Managers 7. Consultation Presented the proposed IDA policy IMF, AfDB, ADB, July 10, 2019 with MDBs toolkit and distributed the June IFAD, IDB paper on Debt 8. Consultation Presented SDFP for comments and Country economic July 18, 2019 with Country suggestions (distributed the June team in the Pacific Teams Debt Paper in advance) Islands Country economic September 25, 2019 team in AFR Source: World Bank estimates 2019 17 For additional details, please see “Box 2: Abidjan High-Level Consultation on Addressing Debt Vulnerabilities”, Addressing Debt Vulnerabilities in IDA Countries: Options for IDA19, June 2019. 23 53. As part of IDA’s enhanced creditor outreach, several rounds of engagements with a broad range of stakeholders were undertaken to deepen dialogue and coordination among MDBs on sustainable lending practices. One of the outcomes have been increased transparency with most MDBs now providing information on country-level commitments, with many regularly updating information posted in their websites. Table 2 shows examples of various recent outreach activities carried out or planned in 2019, which represent ongoing engagements as part of the NCBP implementation but also were part of the NCBP review process. 54. Engagement with non-Paris Club, bilateral and private creditors remains sporadic. The IMF and the WB have been delivering workshops for non-Paris Club bilateral creditors on debt sustainability analysis and lending frameworks and a more diversified group of agencies have been reaching out to IDA through its “Lending to LICs” mailbox. Starting from 2015, thirty inquiries on average per year have been received/answered by IDA staff. Most of the inquiries related to concessionality requirements in countries of interest, including the minimum grant element requirement, calculation of the grant element for specific loans, NCB ceilings, and the status of specific entities within countries (e.g., whether it is a SOE and if its debt is considered external PPG). The inquiries also often include information regarding intentions to provide financing along with loan details. Some inquiries referred to general DLP/NCBP related matters, such as the grant element calculator formula and updates on the review of the policies. Over the years, agencies such as European Investment Bank (EIB), AfDB, Department for International Development (DFID), various Export Agencies, and more recently China Export and Credit Insurance Corporation, have reached out through the “Lending to LICs” mailbox. Countries of most interest were Benin, Burkina Faso, Ethiopia, Ghana, Guinea, Maldives, Mali, Mozambique, Senegal, Tanzania, Zambia, Comoros, Solomon Islands, and Uganda—with Ethiopia and Mozambique leading the list in terms of number of inquiries. 55. The results from consultations and survey of 18 MDBs18 revealed that: • The NCBP had a positive impact on creditors’ lending decisions but this can be enhanced. More than half of the MDBs surveyed are familiar with the NCBP and factor it into their lending decisions. Fifty percent of the respondents have a formula-based country resource allocation system, all of whom factor debt sustainability into the determination of volumes and/or terms. All the responding institutions that consider debt sustainability in their lending decisions use the joint IMF-WB LIC DSF. Half of them complement the LIC DSF with their own analysis. • Alignment with the IMF’s DLP is critical and should be sustained. Joint WB-IMF shared framework to establish debt limits ensures continuity of the policy advice to countries that move in and out of IMF arrangements. Going forward, continued coordination in setting ceilings is critical - in particular, for countries at moderate risk of 18 Ten completed responses were received from a survey administered to 18 MDBs: African Development Bank (AfDB), Arab Fund for Economic and Social Development (AFESD/FADES), Arab Monetary Fund (AMF), Caribbean Development Bank (CDB), Central American Bank for Economic Integration (CABEI), Corporación Andina de Fomento (CAF), European Commission (EC), European Investment Bank (EIB), Inter-American Development Bank (IaDB), International Fund for Agricultural Development (IFAD), Arab Bank for Economic Development in Africa (BADEA), Islamic Development Bank, Kuwait Fund for Arab Economic Development (KF), Nordic Development Fund, Saudi Fund for Development (SF), Banque Ouest Africaine de Développement (BOAD) - West African Development Bank. 24 debt distress – potentially including the setting of ceilings for countries that are not grant eligible. • Thirty percent of responding institutions have a policy on NCB broadly similar to IDA’s and most are undertaking or plan to undertake reviews of their policies. Their policies consider risk of debt distress rating most often, followed by public financial management capacity (Figure 5). Reponses to breaches to their NCB policy are broadly similar to IDA’s. The engagements with other stakeholders have brought different perspectives and experiences to help sharpen IDA and ADF policies and tools to better help address debt vulnerabilities in client countries. Figure 5. Factors that Influence MDB NCBPs 100 80 60 40 20 0 Risk of debt Macro-fiscal Government Country program Public financial Other distress rating situation investment dialogue management programs capacity Source: World Bank estimates 2019 56. IDA is working towards a set of general principles to promote information sharing and coordination among MDBs with respect to the implementation of resource allocation frameworks and selected debt and financing policies. These principles are intended to be a platform open to all MDBs and IFIs and consultations are ongoing with the IMF, AfDB, IFAD, and others. The principles under discussion include consideration of debt sustainability in resource allocation decisions (volumes and/or terms); engaging in dialogue on policies to reduce debt vulnerabilities, possibly in the form of country platforms; exchange of information on policies; and supporting efforts on financing solutions that enhance borrower country’s resilience. 57. The effectiveness of the NCBP as part of IDA’s sustainable lending toolkit hinges on stronger collective action among borrowers, creditors and international partners. The IMF- WB MPA provides an opportunity to strengthen the International Financial Institutions’ (IFIs) inclusive global platforms and partnerships with borrowing countries; multilateral, bilateral and private creditors; and international partners and stakeholders to promote sound economic policies, prudent debt management and sustainable lending practices. The Addis Ababa Action Agenda calls for debtors and creditors to work together to prevent and resolve unsustainable debt situations. Against this background, creditors should aim for good practice in lending, drawing on principles 25 for sustainable lending such as the guidelines being championed by the Group of Twenty (G-20)19 and the OECD20. Furthermore, the Export Credit Group has developed its own Sustainable Lending Guidelines, which draws parallels to the IMF’s DLP and IDA’s NCBP and helps determine Export Credit Agency (ECA) lending terms to countries subject to the NCBP. E. Summary of Key Findings 58. The following is a summary of the key findings from the NCBP review that will help shape the successor SDFP: • Implementation: The NCBP was implemented as intended and this has been confirmed by previous reviews. It has provided a flexible framework to assess countries’ NCB cases based on country and project-specific factors. The results from the surveys suggest that the NCBP and NCBP responses have complemented or fostered IDA country dialogue and programs which are the WB’s primary instruments to help countries address debt vulnerabilities. The results also indicated that NCBP has influenced certain countries to seek alternative sources of concessional financing and helped some countries avoid instances of NCB and/or improve the terms of financing packages in some cases. The Review, however, also found some limits to the policy which need to be addressed as detailed below. • Coverage: Growing debt vulnerabilities have transcended beyond post-MDRI and grant- eligible IDA-only countries into other IDA countries. However, NCBP still applies only to grant-eligible IDA-only and MDRI recipient countries (currently 40 countries). Moreover, all IDA clients receive concessional financing, and thus, the NCBP objective to help limit the risk of cross-subsidies and moral hazard applies to all IDA countries, whether they benefited from MDRI in the past or not. Therefore, there is a strong rationale to expand the scope of the new policy to include all IDA countries21 which would help enhance IDA’s role in helping promote debt sustainability. Furthermore, the implementation of the NCBP focused mostly on external NCB. However, the financing landscape has been changing with the share of domestic debt is rising22. The new policy should also take a broader view of debt, which will tighten alignment with the current IMF DLP which encompasses both external and domestic public debt. • Borrower incentives: This Review found that the impact of NCBP on borrowers’ incentives have been positive but limited. This is consistent with the findings of the 2008, 2010 and 2015 Reviews and from its inception, the policy acknowledged this limitation. As detailed in Section II, the 2008, 2010, and 2015 reports recommended further 19 See: G20. 2017. “G20 Operational Guidelines for Sustainable Financing”. 20 See: OECD. 2016. “Principles and Guidelines to Promote Sustainable Lending Practices in the Provision of Official Export Credits to Lower-Income Countries”. 21 As the policy’s country coverage is expanded, the design will need to mitigate the inherent trade -offs with other IDA objectives (e.g., increase resources to FCS countries) or the risk of penalizing countries facing external shocks such as natural disasters—concerns that have been identified and emphasized by Participants to the IDA19 Replenishment Meetings. 22 Domestic debt has received more prominence under the revised LIC DSF which has been implemented since July 2018. The revised LIC DSF now provides both external and total public debt distress ratings. 26 adjustments to the NCBP implementation, including alignment to DLP and enhancing transparency, and flexibility. While the policy was able to influence the size of IDA allocations to countries that were subject to the responses, these adjustments were not able to significantly affect the broader borrowers’ incentives. Given the limited impact of these adjustments, Management’s proposal is now to move to a more comprehensive and ambitious SDFP next fiscal year. The SDFP would aim to introduce an incentive-based system with the appropriate performance and policy actions necessary to help client countries address the causes of debt vulnerabilities. • Impact on creditor coordination: The NCBP’s impact on creditor coordination, overall, has been effective with room for further improvements. Consideration should be given as to how to align the forthcoming SDFP with the IMF’s DLP as revised under the ongoing review and enhance coordination during implementation. Coordination with MDBs could be further enhanced to better complement policies to support debt sustainability and work towards a set of general principles for sustainable lending. Engagement with non-Paris Club bilateral and private creditors needs to be reinforced. Greater simplicity, clarity, and transparency, as well as enhanced predictability and consistency of decisions, under the new policy could improve the signaling effect. Outreach to private and public creditors through the IDA’s “Lending to LICs” mailbox could be expanded including through enhanced dissemination of the mailbox and monitoring of its impact. • Communication and transparency: Results from consultations and surveys pointed towards a need to enhance transparency and communication. The new policy should ensure continued effective communication and transparency, including communicating the correct reporting obligations and encouraging systematic and timely reporting of NCB both of which are critical to the effectiveness of policy. As officials change from time to time in country authorities and offices, it is important that communication be a continuous and enhanced process backed up with appropriate TA, if necessary. IV. CONCLUSION AND NEXT STEPS 59. This 2019 Review assessed the experience with the NCBP across IDA countries following the 2015 Review, including an overall analysis of the experience since the policy’s introduction in 2006. It was informed by a comprehensive internal and external consultation process, involving consultations with the IMF, other MDBs, WB country teams, and other stakeholders. The Review concluded that NCBP was implemented as intended and that it played a positive but limited role in promoting debt sustainability in IDA countries, and thus there is room for improvement. Given the change in the development finance landscape and the results of these reviews, the Review recommends transitioning to a forthcoming SDFP. 60. Preliminary findings of the Review were presented at the IDA19 Replenishment meetings in April and June 2019. At these meetings, participants voiced their support for a proposal to transition from the NCBP to the forthcoming SDFP given the renewed challenges on debt sustainability in IDA countries. The name reflects the objectives of the forthcoming policy, which are to enhance incentives for countries to move towards sustainable financing and promote coordination among all creditors. 27 61. The forthcoming SDFP would support IDA countries increase their capabilities to address debt related vulnerabilities, using concrete actions supported by lending, diagnostics and TA. It would i) be broader and apply to all IDA countries systematically; ii) support IDA countries increase their capabilities to address debt related vulnerabilities, using concrete actions supported by lending, diagnostics and TA; and iii) enhance creditor outreach with broader scope and more information sharing, especially among the MDBs. The forthcoming SDFP will have two pillars: • The Debt Sustainability Enhancement Program (DSEP) would enhance incentives for countries to move toward sustainable financing. Incentives will take the form of a share of the country’s allocation, for countries at moderate and high risk of debt distress, that will be set aside and released upon satisfactory implementation of the agreed policy actions. NCBP SDFP Objective and Broad objective. Narrow Underpinned by MPA coverage country coverage (post- MDRI and IDA grant only Clearer objectives recipients) Broader country coverage - applies to all IDA countries. Borrower incentives Volume cuts and hardening The Debt Sustainability of terms Enhancement Program to enhance incentives for countries to Often driven by loan-by-loan move toward sustainable financing. consideration Outreach and Outreach was effective, but The Program of Creditor transparency its scope limited. Outreach, building on IDA’s global platform and convening role In some cases, reports about to promote debt transparency, non-concessional borrowing outreach and creditor coordination were not reported. on sustainable lending practices. • The Program for Creditor Outreach (PCO) will build on IDA’s global platform and convening role. The objective of the program is to facilitate information sharing, dialogue and coordination, including coordination among MDBs, to help mitigate debt related risks. 62. The DSEP will enable better tailoring to the country circumstances, including circumstances faced by Small States and FCS. The performance and policy actions would aim to strengthen: (i) fiscal sustainability, (ii) debt management, and/or (iii) debt reporting and coverage. 63. The PCO implementation will promote stronger collective action among borrowers, creditors and international development partners. Given the new development finance landscape, IDA’s efforts to help countries increase capabilities required to address their debt - 28 related vulnerabilities would need cooperation and coordination by all involved including non- Paris Club and other non-traditional creditors. The SDFP, through its PCO, would seek to strengthen coordinated actions by various actors to promote sound economic policies, prudent debt management and sustainable lending practices. 64. The forthcoming SDFP, expected to be submitted for Board approval during the third quarter of FY20. 29 Annex 1. Additional Tables Table A1. 1: NCBP Implementation Update since 2015 (As of August-2019) No. Country DSA risk DSA Institution Public Eligibility Current NCB Type of date Leading debt/GDP to NCBP ceiling NCBP NCBP WEO, ceiling Ceiling monitoring 2018 NCBP countries led by IMF Support Program 1 Afghanistan High Nov-18 IMF 7.1 Zero 2 Benin Moderate Jul-19 IMF 54.6 Non-zero CFAF 797 bil PV 3 Burkina Faso Moderate Jan-19 IMF 43.0 Non-zero CFAF 370 bil PV 4 Chad High Mar-19 IMF 46.6 Zero 5 Gambia, The High, in distress Apr-19 IMF 83.2 Zero 6 Guinea Moderate Dec-18 IMF 38.7 Non-zero $650 mil Nominal 7 Madagascar Low Jul-19 IMF 39.7 Non-zero $900 mil PV 8 Malawi Moderate Nov-18 IMF 61.3 Zero 9 Mauritania High May-19 IMF 83.9 Non-zero $307 mil Nominal 10 Niger Moderate Jul-19 IMF 55.1 Zero CFAF 225 bil PV 11 Senegal Low Jan-19 IMF 64.4 No limit 12 Sierra Leone High Dec-18 IMF 71.3 Zero 13 Togo Moderate Jul-19 IMF 74.6 Zero NCBP countries led by World Bank Monitoring 1) Active IDA NCBP countries 14 Comoros Moderate Jun-18 WB 31.2 Non-zero Euro 25 mil Nominal 15 Ethiopia High Nov-18 WB 61.1 Zero 16 Maldives* High Oct-17 WB 61.5 Zero 17 Mozambique High, in distress Apr-19 WB 100.4 Zero 18 Tajikistan High Aug-17 WB 47.9 Zero 2) Others under WB NCBP 19 Burundi High Mar-15 WB 58.4 Zero 20 Central African Republic High Jul-19 IMF 48.5 Zero 21 DRC Moderate Oct-17 WB 15.7 Zero 22 Guinea Bissau Moderate May-18 IMF 56.1 Zero 23 Haiti High Nov-16 WB 33.0 Zero 24 Kiribati* High Dec-18 WB 21.9 Zero 25 Kyrgyz Republic Moderate Jul-19 WB 56.0 Non-zero Loan by loan 26 Liberia Moderate Jun-19 WB 40.5 Non-zero Loan by loan 27 Mali Moderate May-18 WB 36.6 Non-zero Loan by loan 28 Marshall Islands* High Aug-18 WB 25.2 Zero 29 Micronesia* High Sep-17 WB 20.2 Zero 30 Samoa* High Mar-19 WB 50.2 Zero 31 Sao Tome and Principe* High, in distress Jul-18 WB 81.3 Zero 32 Solomon Islands Moderate Oct-18 WB 12.1 Non-zero Loan by loan 33 South Sudan High, in distress Feb-17 WB 43.8 Zero 34 Tonga* High Jan-18 WB n/a Zero 35 Tuvalu* High Jun-18 WB 27.6 Zero 36 Vanuatu* Moderate Jun-19 WB 50.5 Non-zero Loan by loan 37 Yemen, Republic of High, in distress Sep-14 WB 63.2 Zero 3) No NCB ceiling 38 Rwanda Low Jul-19 WB 40.7 Non-zero Loan by loan PV, Nom. 39 Uganda Low Apr-19 WB 42.2 Non-zero Loan by loan PV, Nom. 40 Tanzania Low Jun-19 WB 36.0 Non-zero Loan by loan PV, Nom. *Countries with small states exceptions. Source: Joint WB IMF DSAs; Countries eligibility to Debt Limit Conditionality (WB-IMF) and WEO, IMF. 30 Table A1. 2. Inventory of NCBP Country Cases, 2007-2019 Loan Years IDA Amount Country Reviewed Cycle % of GDP DSA IDA Response 1 Angola 2007 IDA14 35 Moderate harden term 2 Ghana 2007 IDA14 8 Moderate harden term 3 Mali 2007 IDA14 1 Low Waiver 4 Mauritania 2007 IDA14 4 Moderate Waiver 5 Rwanda 2008 IDA14 2 Moderate Waiver 6 Cameroon 2009 IDA15 0.3 Low Waiver 7 Congo, Republic 2009 IDA15 0.3 High Waiver 8 DRC 2009 IDA15 1 High Waiver 9 Ghana 2009 IDA15 2 Moderate no response 10 Rwanda 2009 IDA15 1 Moderate Waiver 11 Senegal 2009 IDA15 1 Low Waiver 12 Chad 2010 IDA15 6 Moderate Volume reduction 13 Ghana 2010 IDA15 -- Moderate harden term 14 Mauritania 2010 IDA15 1 Moderate Waiver 15 Burundi 2011 IDA15 4 Moderate Waiver 16 Côte d'Ivoire 2011 IDA15 2 High Waiver 17 Ethiopia 2011 IDA15 2 Low Waiver 18 Ghana 2011 IDA15 9 Moderate no response 19 Ghana 2011 IDA15 0 Moderate no response 20 Cameroon 2010-12 IDA15/16 5 Low Waiver 21 Lao PDR 2010-14 IDA15/16 9 Moderate harden terms 22 Zambia 2011-12 IDA15/16 5 Low Waiver 23 Burundi 2013 IDA16 1 High Waiver 24 Comoros 2013 IDA16 7 High Waiver 25 Kyrgyz Republic 2013 IDA16 1 Moderate Waiver 26 Guinea 2012-13 IDA16 6 Moderate Waiver 27 Chad 2013-14 IDA16 5 High Waiver 28 Chad 2013-14 IDA16 9 High Waiver 29 Ethiopia 2013-14 IDA16 17 Moderate volume reduction + harden terms 30 Zambia 2013-14 IDA16 6 Moderate no response 31 Madagascar 2014 IDA16 1 Low Waiver 32 Sao Tome & Principe 2014 IDA16 10 High Waiver 33 Togo 2014 IDA16 1 Moderate Waiver 34 Ethiopia 2015 IDA17 2 Moderate hardening terms 35 Ethiopia 2016 IDA17 1 Moderate hardening terms 36 Mozambique 2016 IDA17 11 Moderate volume reduction + harden terms 37 Ethiopia 2017 IDA17 1 Moderate no response 38 Maldives 2017 IDA17 24 High hardening terms 39 Mozambique 2017 IDA17 1 High volume reduction 40 Benin 2018 IDA18 1 Moderate Waiver 41 Comoros 2018 IDA18 8 Moderate no waiver, ceiling suggested 42 Ethiopia 2018 IDA18 2 High hardening terms 43 Maldives 2018 IDA18 22 High hardening terms 44 Mozambique 2018 IDA18 -- High volume reduction 45 Tajikistan 2018 IDA18 7 High Waiver 46 Comoros 2019 IDA18 2.5 Moderate Euro 25m ceiling established 47 Ethiopia 2019 IDA18 -- High hardening terms 48 Maldives 2019 IDA18 6.4 High hardening terms 49 Mozambique 2019 IDA18 -- High no response 50 Mozambique 2019 IDA18 36.4 High Waiver 51 Tajikistan 2019 IDA18 -- High no response Source: NCBP Country Notes, World Bank. 31 Annex 2. NCBP Implementation – Survey of the World Bank (WB) Country Teams IDA administered a survey to the WB Country Teams on the experience from the implementation of the Non-Concessional Borrowing Policy (NCBP) and to gather inputs on adjustments needed to enhance the effectiveness of the framework. A total of 44 responses were received with 54 percent of them fully completed. This note considers respondents who are currently working and previously worked on countries subject to the NCBP. The survey comprises three sections on (1) experience and familiarity with the NCBP, (2) policy dialogue with the government on NCBP related issues, and (3) the experience with the NCBP implementation process. Survey results are presented below. 1. Eighty percent of respondents encountered NCBP related issues in policy dialogues with the government. The context of their discussion tends to be macroeconomic monitoring and/or development policy operations (Figure A2.1). The most common issues were non- concessional borrowing followed by debt management (Figure A2.2). When having discussions on addressing debt vulnerabilities with the government, sixty percent of the respondents indicated NCBP complemented the conversation. The main counterparts were from the MoF and the Debt Department and in some instances, the Central Bank and the President’s office. 2. Most respondents (81 percent) indicated that the NCBP played a constructive role in supporting debt sustainability. Those respondents indicated that the NCBP resulted in some countries seeking alternative sources of concessional finance and/or engaging in policy dialogue (Figure A2.3). When external factors contributed to debt burdens or when the NCBP is not well understood, the NCBP had little impact on debt sustainability. 3. Two-thirds responded that the NCBP lowered the risk of IDA grants potentially incentivizing overborrowing. Policy dialogue, alternative sources of concessional finance, and disincentive effect of a response were indicated as contributing factors (Figure A2.4). The NCBP did not play a role when external factors contributed to rising debt burdens or when countries are trying to maximize resources from different sources to finance needs. One respondent stressed that the Grant Allocation Framework (GAF) resulted in the wrong incentives by providing 100 percent grants to countries at high risk of debt distress. Figure A2.1. Context of Policy Dialogues Figure A2.2. Type of Issues Raised 100 100 80 80 Percent 60 60 40 Percent 20 40 0 20 0 Macro monitoring Technical Development Discussions on assistance policy operations non-conessional borrowing 32 Figure A2.3. NCBP and its role in debt Figure A2.4. How NCBP lowered the risk of sustainability IDA grants potentially incentivizing overborrowing 70 70 60 60 50 50 Percent Percent 40 40 30 30 20 20 10 10 0 0 Note: Other includes third party legal contract, and Note: Other includes third party legal contract notification on potential hardening of IDA terms in case of The NCBP did not play a role because of external factors non-concessional borrowing (NCB). The NCBP did not play contributing to rising debt burdens and countries have financing a role because of external factors contributing to rising debt needs and trying to maximize resources from different sources. burdens. 4. Eight out of 21 respondents thought the NCB guidelines23 were helpful in implementing the NCBP. The most helpful guidance was “afford IDA reasonable opportunity to exchange views with the Recipient on the matter” (88 percent). The reasons the guidelines were not helpful include finding the guidelines ex-post, difficulty in finding the guidelines, unclear guidelines, and counterparts not being aware of them. One suggestion is to have guidelines that are more flexible and realistic. 5. Nineteen respondents indicated some reasons that limited NCBP implementation in the countries they worked on (Figure A2.5). Other responses included clients’ lack of understanding of the NCBP, NCBP linked to SOEs, politically driven decisions, limited transparency, and time mis-match between the NCBP monitoring and data availability. 6. Addressing the hardening of terms and volume reductions were both difficult to address for eleven out of nineteen respondents (Figure A2.6). 23 Section 6.01c in General Conditions for IDA Financing 33 Figure A2.5. NCBP and its role in debt Figure A2.6. Which type of responses are sustainability more difficult to address? 70 Hardening of terms 60 10% None 50 16% Volume Percent 40 cuts 16% 30 20 10 Both 0 58% Limited Lack of Administrative Other - Write In information on comprehensive procedures debt terms borrowing plans 7. NCBP-related issues informed a DPO and/or other operations for six out of 21 respondents. For two respondents, the DPO was put on hold. For another, NCB would have had an impact on the macro framework adequacy for the DPO. 8. The majority of respondents (90 percent) thought that the focus of NCBP was current debt levels. One respondent indicated “maintaining debt sustainability” as the focus, while it was unclear for the other. 34 Annex 3. Outreach on Debt – Multilateral Development Bank’s (MDB) Survey on Sustainable Lending Practices IDA administered a survey on lending practices of 18 MDBs24 and International Financial Institutions (IFIs) and received ten completed responses, including all the Regional Development Banks. The survey aims to deepen information exchange on approaches to follow in response to the increased debt vulnerabilities in low-income counties and to identify key common principles and coordinated actions to enhance sustainable lending. The survey comprises four sections on (1) country allocation and terms, (2) non-concessional borrowing or debt limits policy, (3) transparency, and (4) coordination. Survey results are presented below. 1. All the responding institutions provide concessional financing (Figure A3.1). In addition, some provide partial risk guarantees, partial credit guarantees, grants with third party resources, and blending of resources. Three institutions do not provide financing on non- concessional terms while two do not provide grants. 2. Fifty percent of the respondent institutions have a formula-based country resource allocation system, all of whom factor debt sustainability in determining volumes and/or terms. For some of the other institutions, the Board decision alone determines the country resource allocation. IDA has a performance-based allocation (PBA) system that serves as a basis for allocation of IDA resources at the country level. Figure A3.1. Type of financing terms provided Figure A3.2. Number of institutions where by responding institutions debt sustainability considerations affect volumes and/or terms 100 Percent of respondents 80 60 40 20 0 Note: Other includes partial risk guarantees, partial credit guarantees, risk participation, grants with third party resources, blending of resources 24 African Development Bank (AfDB), Arab Fund for Economic and Social Development (AFESD/FADES), Arab Monetary Fund (AMF), Caribbean Development Bank (CDB), Central American Bank for Economic Integration (CABEI), Corporación Andina de Fomento (CAF), European Commission (EC), European Investment Bank (EIB), Inter-American Development Bank (IaDB), International Fund for Agricultural Development (IFAD), Arab Bank for Economic Development in Africa (BADEA), Islamic Development Bank, Kuwait Fund for Arab Economic Development (KF), Nordic Development Fund, Saudi Fund for Development (SF), Banque Ouest Africaine de Développement (BOAD) - West African Development Bank. 35 3. Two responding institutions factor debt sustainability in determining volume of allocations and financing terms; and three do not consider debt sustainability for either purposes (Figure A3.2). Debt sustainability is factored in the volume of country allocations through formulas only or formulas combined with a Board decision (Figure A3.3). IDA does not directly factor debt sustainability25 in determining core IDA financing although a country's performance in the areas of debt policy and management as reflected in its Country Policy and Institutional Assessment (CPIA) score influences financing volumes to a small degree and access to certain Windows. 4. A country’s risk of debt distress affects its concessionality for all the responding institutions that consider debt sustainability in determining financing terms (Figure A3.4). IDA uses a country's risk of debt distress ratings to determine the share of grants and loans in IDA's assistance to the country: IDA-only non-gap26 countries at high risk of debt distress receive their IDA allocations fully on grant terms; at moderate risk of debt distress, IDA concessional financing in a mix of 50 percent credit and 50 percent grant terms; and at low risk of debt distress, concessional IDA resources on credit terms. Figure A3.3. How debt sustainability is Figure A3.4. How a country’s risk of debt factored in the volume of country allocations distress affects the financing terms 70 100 60 80 Percent 50 60 Percent 40 40 30 20 20 0 10 0 Formula-based Board decision- Other based Note: Other include volume discounts and blending of Note: Other include a conceptual framework for providing concessional and non-concessional resources. debt relief to eligible countries. 5. All the responding institutions that consider debt sustainability use the joint World Bank-IMF Debt Sustainability Framework for Low Income Countries (LIC DSF). Half of them complement the DSF with internal analyses. The periodicity of adjustment varies - annually (2), biennially (1), and usually annually but ad hoc for DSF changes (1). 6. Responding institutions summarized what they hoped to address by changing the allocation framework (Figure A3.5). Key among them are to prioritize poorer countries, to achieve more equitable distribution of resources, to factor in specific needs such as infrastructure, and to provide more resources to support reforms. Planned or recently implemented changes include introducing different levels of allocations reflecting the country’s needs, absorption 25 As assessed under the joint World Bank-IMF Debt Sustainability Framework for Low Income Countries. 26 IDA countries with GNI per capita above the operational cutoff for more than two consecutive years are known as "Gap" countries. 36 capacity, and risks; rebalancing weights of specific sectors; and increasing elasticity to ensure new situations of fragility are captured. 7. Seventy percent are familiar with the World Bank’s NCBP, about half of whom factor it in their lending decisions. Thirty percent have a policy on non-concessional borrowing and all have conducted reviews of their polices. Their policies consider risk of debt distress rating most often, followed by public financial management capacity (Figure A3.6). Reponses to breaches to their non-concessional borrowing policy include a zero tolerance for country-level limit breaches (paired with a proactive approach where limits are checked prior to financing approvals) to reducing volumes or adjusting lending terms. Figure A3.5. Is your institution working on Figure A3.6. Factors that influence MDBs’ or has it recently implemented changes to NCBPs its allocation framework? 100 Implemented recently 80 10% Yes 60 Percent 40% 40 No 20 50% 0 Risk of debt Macro-fiscal Government Country Public Other distress rating situation investment program financial programs dialogue management capacity 8. All the responding institutions disclose information on commitments (Figure A3.7). The data on commitments are made available by country (100 percent), by lending instrument (70 percent), and some by sector and income. Several institutions are working on making more data and information available. 9. A majority (80 percent) of the responding institutions coordinate with other MDBs on debt issues (Figure A3.8). The World Bank is the most common partner (75 percent) followed by IMF (63 percent). Figure A3.7. Information Disclosure Figure A3.8. Coordination with other MDBs on Debt Issues 100% 100% 80% 80% 60% 60% 40% 20% 40% 0% 20% Country Commitments Terms and Lending terms Allocations conditions of by borrower loans/support 0% Disclosed Not disclosed 37 Annex 4. Public Debt Dynamics and Debt Vulnerabilities 1. Public debt levels in IDA countries have risen substantially over the past five years. Having fallen steadily for many years on the back of strong growth and debt relief starting from 2013 median public debt un IDA countries increased by 14 percentage points of GDP to reach about 50 percent of GDP in 2018 (see Figure A4.1). The increase in public debt levels was broad- based across IDA-eligible countries (see Figure A4.2), but slightly larger among commodity exporters. The increase in debt levels for countries affected by fragility, conflict and violence (FCV) was slightly lower than for countries not affected by conflict (nine percentage points, compared to 13 percentage points for non-FCS), while Small States experienced a modest increase. Figure A4.1. Public Debt in IDA Figure A4.2. Public Debt by IDA groupings, countries, median, percent of GDP percent of GDP 90 70 Average Median 80 60 50 70 40 60 30 50 20 40 10 30 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2007 2013 2018 2007 2013 2018 2007 2013 2018 2007 2013 2018 All IDA FCS non-FCS Small states Source: WEO, April 2019. Source: WEO, April 2019. 2. Rising debt levels and shifts in the composition of debt have contributed to increased debt vulnerabilities in IDA countries as a group. As noted in the April 2019 Paper on Debt Vulnerabilities in IDA countries, as of end-January 2019, 50 percent of IDA countries covered under the joint World Bank-IMF DSF are assessed at high risk of external debt distress or in debt distress (34 countries)27. This number falls to around 30 percent once small states are excluded. In 2018, more than 80 percent of IDA-only and 75 percent of Gap countries are assessed to be at high or moderate risk. Ten countries are assessed to be in debt distress. Most countries at moderate risk of debt distress have limited debt carrying capacity, with rising liquidity pressures. Countries at low risk of debt distress tend to have more diversified exports. Rising debt levels have also crowded out public expenditures as debt service obligations account for an increasingly larger share of public expenditures (see Figure A4.4). 27 See “Debt Vulnerabilities in IDA Countries: Policy Options for IDA19”. Paper for IDA19 Replenishment. April 2019. 38 Figure A4.3: Evolution of Debt Distress Risk Figure A4.4: A ratio of debt service to Ratings for IDA Countries, Percent of IDA expenditures in IDA countries, simple countries average 30 100 25 80 60 20 40 15 20 10 0 5 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Low Moderate High Source: World Bank/IMF LIC-DSA database. As of May 2019. Source: IMF WEO, April 2019. Note: Figures are in calendar year. High risk includes countries in Note: Average of these countries where data is available, debt distress. data not available for all IDA countries. 3. A decomposition of public debt dynamics across IDA countries points to several reasons for increased debt vulnerabilities. These include: (i) weaknesses in fiscal policy frameworks (including, weak domestic resource mobilization and efficiency of public expenditures); (ii) changing composition of debt towards more expensive and riskier sources of financing; and (iii) weak debt transparency. In addition, several countries were affected by internal conflict (Yemen and Burundi experienced increases in the public debt-to-GDP ratio) or by shocks (Ebola epidemics in the case of Liberia and Sierra Leone – where the public debt-to-GDP ratio increased by about 20 percentage points between 2013 and 2017). 39 Annex 5. Summary of Proposals Made in the 2015 Review 1. The main objective of the update paper was to inform Executive Directors about the implementation of the Non-Concessional Borrowing Policy (NCBP). The update provided detailed information about NCBP cases reviewed by IDA Management since the last update in 2010 and took stock of the lessons learned. It also provided information on developments and the debt outlook in countries subject to the NCBP, and efforts to strengthen debt management capacity. Lastly, the paper also noted the adjustments to the IMF’s Debt Limits Policy (DLP) and outlines how the implementation of the NCBP will be adapted to harmonize the implementation of the two policies28. 2. Since the last Progress Report in 2010, 40 cases in 20 countries have been discussed in the context of the NCBP. Several countries, such as Burundi, Chad, Ethiopia, Madagascar and Zambia had more than one case of non-concessional borrowing (NCB) for NCBP Committee consideration since 2010. Non- concessional loans were in line with IDA’s NCBP based on country- and loan-specific circumstances in the following countries: Burundi, Cameroon, Chad, Comoros, Côte d’Ivoire, Ethiopia, Ghana, Guinea, Kyrgyz Republic, Madagascar, Mauritania, São Tomé and Príncipe, Togo, and Zambia. For three countries, IDA’s financing terms and allocation volumes were adjusted in response to their NCB: Chad, Ethiopia, and Lao PDR. Chad’s IDA allocation was reduced 20 percent in FY11; Ethiopia’s grant portion for FY15 was converted to regular IDA credits, and the allocation was subject to a further 5 percent volume reduction; and 62 percent of Lao PDR’s grant allocation was converted to credit terms in FY15. The adjustments to financing terms applied to Ghana in the form of hardening of IDA terms since FY09 were discontinued in FY12. 3. Adjustments to the implementation arrangements for the NCBP have been introduced with the aim of enhancing flexibility, harmonizing with the IMFDLP, and augmenting transparency. These adjustments relate to several areas: ▪ Capacity assessment. A joint World Bank-International Monetary Fund (WB-IMF) streamlined capacity assessment narrows the focus from a wide range of debt and public financial management indicators to the authorities’ ability to record and moni tor external public and publicly guaranteed debt in a timely manner. Applying the streamlined capacity assessment, the number of countries with adequate capacity increased from four to approximately 15, compared to 39 countries that were under the NCBP29. ▪ Setting debt ceilings. To enhance further the flexibility of the policy, countries at low or moderate risk of debt distress that have adequate capacity may request ceilings on total external public and publicly guaranteed debt in present value terms. While more complex to monitor, the ceiling in present value terms removed the differentiation between concessional and non-concessional loans. This option was in addition to the “old approach” of nominal ceilings on non-concessional borrowing or loan-by-loan considerations. Furthermore, countries at high risk of debt distress would continue to be able to borrow non-concessionally based on “loan-by-loan” considerations. Finally, if a deterioration in the risk of debt distress occurs under an IMF arrangement, a justification for IDA’s grant allocation in the following 28 The paper establishing the NCBP was sent to Executive Directors for discussion (IDA.2006. “IDA countries and non - concessional debt: dealing with the ‘free rider’ problem in IDA14 grant -recipient and post-MDRI countries”, IDA/R2006- 0137/1, June), whereas the two subsequent updates (2008 and 2010) were sent to the Executive Directors for information only. 29 The number of countries subject to the NCBP has declined from 46 in FY15 to 42 in FY16 as four countries have shifted to gap status starting July 1, 2015 (Côte d’Ivoire, Lao PDR, Nicaragua and Zambia). 40 fiscal year would be based on a case-by-case assessment with the goal of promoting equal treatment across IDA clients. ▪ Transparency. In addition to regular Board notes on IDA measures taken based on the NCBP as well as detailed descriptions of all NCBP cases in Board updates, IDA measures would be reported as part of OP3.10 Annex D, comprising a table with aggregated loan information starting July 1, 2016. Annex D is a public document. In addition, a real-time table with agreed ceilings and IDA decisions will be established on IDA’s external website. Lastly, borrowing plans would become part of the Country Notes sent to the Board. 41 Annex 6. Principles that Guide Exceptions to Non-Concessional Borrowing Ceilings30 1. The concessionality benchmark proposed for the purposes of identifying cases of free riding has been a proven benchmark in Poverty Reduction and Growth Facility (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 (NCB) in countries (which are often zero). In the context of the PRGF, these limits can be overridden in one of three ways: (i) agreement ex- ante within program criteria by defining a sub-ceiling to accommodate a specific non-concessional loan, (ii) finding 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. 2. Similar to considerations that feed into decisions on NCB limits in the PRGF, a number of country-specific and loan-specific factors would be taken into account in the free rider context to assess whether an exception to the zero-ceiling using the proposed benchmark is warranted. Although many proposed loans may have merit on 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 of the country. A modest level of overall borrowing by the country on the basis of the Debt Sustainability Analyses to accommodate a particular investment may warrant consideration. For such a consideration, clear reporting of overall borrowing plans is needed, and enhanced creditor coordination through the Debt Sustainability Framework would facilitate this possibility. ▪ Impact of borrowing on the macroeconomic framework. Whether or not the borrowing would have a deleterious effect on the macroeconomic framework would influence the consideration of an exception. ▪ Impact on the risk of debt distress. The current risk classification, and whether or not the loan is likely to lead to a higher risk of debt distress will be a key consideration. Given their lower-risk of debt distress, and generally better performance, more flexibility is envisaged for “green light” countries. In addition, “yellow light” countries could benefit from somewhat greater flexibility than “red light” ones. ▪ Strength of policies and institutions, especially public expenditure management and debt management. The Board paper makes clear that policies and institutions, in particular, those governing the efficiency of public investment are critical31. Without these, even high return projects may fail to meet objectives. 30 See Box 3 in: IDA.2006. “IDA countries and non -concessional debt: dealing with the ‘free rider’ problem in IDA14 grant - recipient and post-MDRI countries”. IDA/R2006-0137/1, June 31 See: World Bank. 2006. “Fiscal Policy for Growth and Development: An Interim Report”. DC2006 -0003, April 6. 42 Loan-specific: ▪ Development content and potential impact of the loan, i.e., investment will unlock a proven bottleneck to development as determined by analytical work such as a Public Expenditure Review. ▪ Estimated economic, financial and social returns to investment of the project, weighted by the probability that the project will succeed. ▪ Lender equity stake in the project. ▪ No additional costs associated with the loan, i.e., collateralization, hidden costs. ▪ No other sources of more concessional financing are available. ▪ Concessionality of the overall financing package for a particular investment.