Mitigating the Inlpact of MDRI Netting Out on New IDA Country Allocations: Additional Options

Transcription

1 IDA16 Mitigating the Inlpact of MDRI Netting Out on New IDA Country Allocations: Additional Options International Development Association IDA Resource Mobilization Department (CFPIR) February 2010

2

3 Abbreviations and Acronyms DSF HIPC IDA IMF MDG MDRI MTR NCBP NPV ada PBA SDR Debt Sustainability Framework Heavily Indebted Poor Country International Development Association International Monetary Fund Millennium Development Goal Multilateral Debt Relief Initiative Mid-Term Review Non-Concessional Borrowing Policy Net Present Value Official Development Assistance Performance Based Allocation Special Drawing Rights

4

5 Table of Contents Executive Summary... i I. Introduction... 1 II. A Framework for Weighing Key Trade-offs... 3 III. Additional Scenarios for Mitigating the Allocation Impact of MDRI Netting Out... S A. Limiting the MDRI netting out to 50 percent of annual debt service foregone... 6 B. Limiting the MDRI netting out to 30 percent of gross PBA allocation... 7 C. Introducing a minimum per capita allocation of US$2.S (SDR1.6)... 8 D. Placing a moratorium on MDRI netting out for the IDA I6 period... 9 IV. Comparing the options V. Key Issues for discussion... IS Tables Table I: Impact of MDRI netting out on the IS top beneficiaries vs. IS most negatively affected countries during IDAIS-IDA Table 2: Allocation and equity impact of capping MDRI netting out at 50 percent of debt service foregone...6 Table 3: Allocation and equity impact of capping netting out at 30 percent of gross PBA allocation...7 Table 4: Allocation and equity impact of a minimum per capita allocation of US$ Table 5: Allocation and equity impact of putting a moratorium on netting out mnning through the IDA16 period Table 6: Comparison of Options: 15 top beneficiaries vs. 15 negatively affected countries Figures Figure 1: Allocation impact of the various scenarios on the 15 most negatively affected countries relative to the baseline Figure 2: Equity impact of the various scenarios on the 15 top beneficiaries relative to the baseline Annexes Annex 1: Results of Scenarios MTRI to MTR4 in the IDA15 Mid-Term Review Paper (Reproduced) Annex 2: Results of Additional Scenarios 1 to Annex 3: Top 15 Countries Currently Benefiting from the MDRI Netting Out Annex 4: Comparison ofthe Allocation Impact of All Scenarios Annex 5: Comparison of the "Equity" Impact of All Scenarios... 29

6

7 EXECUTIVE SUMMARY i. The Multilateral Debt Relief Initiative (MDRI) which provides for 100 percent cancellation of eligible debt owed to IDA by countries reaching the Completion Point for the Heavily Indebted Poor Country (HIPC) Initiative has been under implementation since July Under the MDRI, IDA's "gross assistance flows" to countries benefiting from debt relief are reduced by the amount of annual debt service forgiven, an allocation process known as the "MDRI netting out". The rationale for this is to help reduce moral hazard and promote equity of treatment among all low-income countries. Ii. The equity and moral hazard principles underlying the MDRI are practically implemented by linking the distribution of the "compensatory" resources to country performance and using the PBA system to distribute the resources to all IDA countries. As a result, several non-mdri countries gain from the current practice of MDRI netting out. Under the baseline scenario, a total of 29 countries benefit from the MDRI netting out. Three countries among them-bangladesh, Nigeria, and Kenyacapture about 55 percent of the benefits during IDA The 15 top beneficiaries are projected to gain about 11 percent of their gross annual allocation, or about SDROA per capita per year, between IDA On the other hand, the implementation experience to date clearly indicates that the MDRI netting out has reduced new IDA allocations for some countries receiving debt relief, and it is projected to lead to even more significant declines in the future. A paper presented at the IDA15 Mid-Term Review (MTR) showed that under the current practice ofmdri netting out, 29 IDA countries' allocations are negatively impacted. Of these, 15 are severely affected and a number of countries-including Chad, Guinea, Guyana, Haiti, Mauritania, the Republic of Congo, and The Gambia-will receive little or no new IDA allocations starting as early as IDAl7. The decline in new IDA allocations will have serious consequences for the countries affected: IDA will be unable to support these countries with fresh resources; and its presence on the groundincluding in undertaking policy dialogue and knowledge and diagnostic work-will be significantly reduced. This will in turn negatively impact its ability to playa leadership role in donor coordination and aid effectiveness. The countries themselves risk reversing many years of hard-won development and poverty reduction gains. iv. At the IDA15 MTR, held in November 2009, IDA Deputies discussed the negative allocation consequences of the MDRI netting out mechanism and its implications for IDA's operational engagement in affected countries. While the MTR paper considered four scenarios for mitigating the allocation impact, IDA Management recommended eliminating the MDRI netting out as the best course of action. However, prior to making a decision, IDA Deputies requested that Management present additional options that would strike a better balance between preserving the principles of MDRI and mitigating the negative allocation consequences of the MDRI netting out.

8 - ii - v. Accordingly, this paper presents four additional scenarios. In addition, one of the scenarios in the MTR paper (Scenario MTR2) is also included for comparison. For each scenario, the paper discusses the allocation impact on the most negatively affected countries as well as the extent to which the equity of treatment is preserved for the top beneficiary countries. vi. These options perform differently on the objective of striking a reasonable balance between increasing allocations to the most negatively affected countries on the one hand and preserving the equity of treatment for the top beneficiary countries on the other. Capping the MDRI netting out at 50 percent of gross PBA allocation (Scenario MTR2) mitigates the zero or negligible allocation impact for the most affected countries. An important strength of this option is that it links the MDRI netting out to the size of initial PBA allocation, and ensures that countries with small PBA allocations (but high debt service foregone) are not affected unduly by the MDRI netting out. However, the 50 percent capping may be too high a threshold to allow most of the affected countries to benefit significantly. Indeed, only 8 of the 15 most affected countries are projected to have debt service foregone in excess of 50 percent of their PBA allocation during IDAI5-20, with 7 of the countries not benefiting if this measure were chosen. Capping the MDRI netting out at 50 percent of debt service foregone (Scenario 1) mitigates the zero or negligible allocation impact for all of the IS most adversely affected countries. However, it has the highest equity tradeoff for the top beneficiaries, in part because the MDRI netting out is simply reduced by half without targeting that benefit to those countries with the lowest PBA allocations. Hence, although it performs better allocation wise, it may not be the best option from the equity perspective. Capping the MDRI netting out at 30 percent of gross PBA allocation (Scenario 2) mitigates the zero or negligible allocation impact for all of the 15 most adversely affected countries. As in the MTR2 scenario, this option links the MDRI netting out to the size of the initial PBA allocation, helping to limit any excessive impact of the netting out on the PBA allocation. Further, the threshold at 30 percent of the PBA allocation will allow all the 15 most affected countries to benefit from this option (as their average debt service foregone is projected to be above 30 percent of their PBA allocation during IDAI5-20), without imposing excessive equity cost on the beneficiaries (with only about 2 percent loss in the PBA allocation for the top 15 beneficiaries relative to the baseline). Introducing a minimum per capita allocation of US$2.5 (Scenario 3) to meet the fixed costs of country engagement increases allocation for only 9 of the 15 most adversely affected countries; and, 6 ofthe 15 countries most adversely affected by the MDRI netting out would receive a lower allocation than under the baseline as they would have to be "taxed" in order to keep all countries at the

9 - iii - minimum per capita level. Even though these 6 countries are not among those who face the problem of negligible or zero allocation due to netting out (and the difference in allocation is small), the minimum per capita floor can be seen as exacerbating, rather than reducing, the MDRI related allocation problems for these countries. Although it is an option with the best impact in maintaining equity for the top beneficiaries, its allocation impact is very limited. Putting a temporary moratorium on the MDRI netting out running through the IDA16 period (Scenario 4) and to be reviewed in the context of the IDA 17 replenishment fully mitigates the zero or negligible allocation impact for all of the 15 most adversely affected countries. Indeed, it has the highest allocation impact on the 15 most affected countries (for IDA16), but also the highest equity tradeoff for the top 15 beneficiaries. vii. All options involve making important trade-offs between higher allocation for the most affected countries and equity of treatment for the non-mdri lowincome countries. IDA Deputies are requested to provide guidance on the option that they think would balance these tradeoffs. With regard to the temporary moratorium option, Deputies may consider this option as either a self-standing option or as a supplementary option complementing whatever other option IDA Deputies agree upon to address the allocation impact of the MDRI netting out. Vlll. IDA Deputies may want to consider the following questions: Do IDA Deputies agree that the options discussed above have sufficiently captured the allocation impact and the equity impact of MDRI netting out? Which of the options strikes the best balance between the allocation impact and the equity impact of MDRI netting out? Do IDA Deputies wish to consider a MDRI netting out moratorium for the period running through IDAI6 as a temporary measure to enhance the allocations to the most negatively affected countries (starting with FY 11), while leaving open the possibility to review it during the IDA17 replenishment? If so, do IDA Deputies wish to consider this as a self-standing or supplementary option?

10

11 I. INTRODUCTION 1. In the spring of 2006, donors and shareholders approved IDA's participation in the Multilateral Debt Relief Initiative (MDRI), which provides 100 percent cancellation of eligible debt owed to IDA by countries reaching the Completion Point for the Heavily Indebted Poor Country (HIPC) Initiative. Following that, IDA began implementing MDRI in July 2006 as approved by IDA's Executive Directors and Board of Governors and has, as of June 2009, provided SDR28.6 billion (US$43.S billion) of debt reliefto 26 HIPC countries. In order to safeguard IDA's long-term capacity to finance international development, donors have agreed to provide "dollar-for-dollar" compensation for the debt relief provided by IDA. 2. In addition to improving long-term debt sustainability by providing irrevocable debt relief for IDA countries with heavy debt burdens, the MDRI was designed to address equity of treatment and moral hazard issues associated with debt relief. IDA's "gross assistance flows" to countries benefiting from debt relief under the Multilateral Debt Relief Initiative (MDRI) are reduced by the amount of annual debt service forgiven. The MDRI debt relief mechanism affects new IDA allocations through a two-step process. In the first step, an MDRI eligible country's annual debt service foregone is deducted from its annual PBA allocation (also known as MDRI "netting out"). In the second step, compensatory donor resources that IDA receives in lieu of the debt service are then reallocated to all IDA-only countries, including those that do not benefit from the MDRI, using the performance based allocation (PBA) system. The rationale for this is to help reduce moral hazard and promote equity of treatment among all low-income countries. 3. The principles underlying the MDRI netting out mechanism need to be contrasted with the implementation experience to date. While netting out has reduced allocations, it is projected to lead to even more significant declines in new IDA allocations for some countries. A paper presented at the IDAIS Mid-term Review l showed that under the current practice of MDRI netting out, 29 IDA countries' allocations are negatively impacted. Ofthese, 15 are severely affected and a number of countries-including Chad, Guinea, Guyana, Haiti, Mauritania, the Republic of Congo, and The Gambia-will receive little or no new IDA allocations starting as early as IDA17? Most of the countries negatively affected by the MDRI netting out are among the poorest countries in Sub-Saharan Africa and Latin America and the Caribbean. 4. The significant decline in new IDA allocations will have serious consequences for the countries affected. In cases where new IDA allocations fall to negligible levels or even zero, IDA will be unable to support these countries with fresh resources; and its See IDA, "Options for mitigating the impact of MDRI netting out on new IDA country allocations." IDAl5 Mid-Term Review Paper. November. Alternative assumptions around country graduations from IDA-in particular of India and Vietnamand continued growth of IDA resources (through replenishment) will not alter this trend significantly.

12 - 2 - presence on the ground-including through policy dialogue and knowledge and diagnostic work-will be significantly reduced. This will in turn negatively impact its ability to playa leadership and coordinating role (i.e. "platform" role),3 undermining aid coordination and aid effectiveness. The countries themselves risk reversals of many years of hard-won development and poverty reduction gains both because of reduced fresh resources as well as IDA's limited engagement or possible disengagement. 5. At the IDA15 Mid-Term Review (MTR), held in November 2009, IDA Deputies discussed the negative allocation consequences of the MDRI netting out mechanism and its implications for IDA's operational engagement in these countries. The MTR paper considered four scenarios for mitigating the allocation impact: (i) eliminating the MDRI netting out; (ii) capping MDRI netting out at 50 percent of gross PBA allocation; (iii) capping MDRI netting out at 75 percent of gross PBA allocation (and netting out in an NPV-neutral way); and (iv) Redistributing the compensatory donor resources within MDRI-only countries. For ease of reference, the results of the four scenarios in the MTR paper are reproduced in Annex 1 of this paper. 6. At the MTR, IDA Management recommended eliminating the MDRI netting out as the best course of action. Several IDA Deputies supported Management's proposal to eliminate the MDRI netting out in order to increase fresh resources towards the countries most affected, while others opposed the elimination of MDRI netting out on the ground that it undermines the principles of fair treatment and moral hazard that underpinned the MDRI agreement. While there was broad recognition of the negative allocation impact of MDRI netting out, Deputies requested that Management explore additional options that would strike a better balance between preservation of the principles of MDRI on the one hand and the mitigation of the negative allocation consequences of the MDRI netting out on the other. They requested Management to return with recommendations at the first meeting of IDA Accordingly, this paper explores four additional scenarios for mitigating the negative allocation impact of MDRI netting out: (i) Capping the MDRI netting out at 50 percent of annual debt service forgone (rather than gross PBA allocation); (ii) Capping the MDRI netting out at 30 percent of the gross PBA allocation; (iii) Introducing a minimum per capita allocation that would allow IDA to meet the fixed cost of country engagement, and thereby help to deal with the adverse impact of the MDRI netting out; and (iv) Placing a moratorium on MDRI netting out for the period lasting through IDA The rest of the paper is organized as follows. Section II discusses the key tradeoffs associated with the mitigation options for the MDRI netting out. Section III and IV discuss the additional scenarios and presents a comparison of the all scenarios (including those from the MTR). Section V concludes by setting out key issues for discussion. See IDA, "The Role of IDA in the Global Aid Architecture: Supporting the Country-Based Development Model", IDA Resource Mobilization Department (CFPIR), June. See "Chairperson's Summary, IDAI5 Mid-Term Review Meeting, Washington DC, November 18-20, 2009",

13 - 3 - II. A FRAMEWORK FOR WEIGHING KEY TRADE-OFFS 9. At the IDAI5 MTR discussions, many IDA Deputies pointed out that the twin principles underpinning the MDRI --equity and moral hazard-- remained important and requested that Management provide additional options to better balance the key trade-offs between these principles and the allocation impact of MDRI netting out. This section discusses the key trade-offs involved in seeking to address the adverse allocation impact of the MDRI netting out. 10. The MDRI agreement was underpinned by two key principles: equity-oftreatment and reduction of moral hazard. 5 The equity-of-treatment principle emphasized the need to maintain equity between HIPC countries that are eligible for MDRI and non-hipc low-income countries that do not receive debt relief because they did not accumulate unsustainable debt in the first place. The principle emphasized that the latter group should also benefit from the compensatory resources donors provide to IDA for debt relief. The moral hazard principle emphasized that debt relief should not reward, through high future aid flows, countries that had accumulated high and unsustainable debt in the past. Nor should debt relief create an incentive for countries to rapidly re-accumulate new debt, which would defeat the original purpose of the MDRI. 11. Since the original MDRI agreement of 2006, IDA has adopted several measures that have contributed towards minimizing the moral hazard associated with debt relief. These have included developing and implementing, jointly with the IMP, the low-income country debt sustainability framework (DSP); designing and implementing IDA's grant allocation system linked to DSP risk ratings; developing and implementing IDA's Non-Concessional Borrowing Policy; and developing tools and providing technical assistance aimed at strengthening borrowers' debt management capacity. These measures have greatly contributed to reducing moral hazard. 12. The principles underlying the MDRI are practically implemented by linking the distribution of the "compensatory" resources to country performance and using the PBA system to distribute the resources to all IDA (including non-hipe) countries. As a result, several non-mdri countries benefit from the current practice of MDRI netting out. Under the baseline scenario, a total of 29 countries benefit from the MDRI netting out. Three countries among them-bangladesh, Nigeria, and Kenyacapture about 55 percent of the benefits during IDA15-20 (Annex 3). These include the Republic of Yemen (14 percent), Lao PDR (13 percent), Tajikistan (13 percent), Nepal (12 percent), Cambodia (12 percent), Kyrgyz Republic (12 percent), Lesotho (11 percent), Bangladesh (11 percent), Nigeria (11 percent), Kenya (11 percent), Kosov0 6 (11 6 See: 08 Finance Ministers' Conclusions on Development, London, June 2005; and "Technical Note: 08 proposal for RIPC debt cancellation" in "The 08 Debt Relief Proposal: Assessment of Costs, Implementation Issues, and Financing Options" by IDA (September, 2005). It is assumed to stay as an IDA country under the base case, although there is a possibility for Kosovo to graduate from IDA because of its high income (in which case it won't receive compensatory allocations),

14 - 4- percent), Afghanistan 7 (11 percent), Mongolia (10 percent) and Djibouti (9 percent). 8 On average, these countries will gain about 11 percent of their gross annual allocation, or about SDRO.4 per capita per year, between IDA (Table I). Most of the options considered to mitigate the allocation impact of the MDRI netting out will affect the allocation of these countries negatively, although differently under the different scenarios. Table 1. Impact of MDRI netting out on the 15 top beneficiaries vs. 15 most negatively affected countries during IDAI5-IDA20 (SDR million) Impact of MDRI Netting Out Top 15 Top 15 Negatively Beneficiaries Affected Countries Average Annual Allocation (GainILoss) Average Annual Allocation (% ofpba allocation) Average Allocation Per Capita per year (GainlLoss) Note: 11 The latest population figures for FYlO are used in the calculation. The FY 10 population data are estimated at 518 million for the top 15 beneficiaries and at million for the 15 most negati vely affected countries. 13. The equity of treatment principle should, however, be weighed against the severe adverse allocation impact on some of the poorest countries. If nothing is done and MDRI netting out continues in its present form (i.e. under the baseline), it is projected to lead to significant declines in new IDA allocations over IDAI5-20. As shown in the IDA15 MTR Paper, the MDRI netting out adversely affects 29 countries overall; and 15 of them are severely affected. The 15 most negatively affected countries, as percent of their PBA allocation, are: Chad (-100 percent), Guyana (-79 percent), Guinea (-76 percent), Mauritania (-63 percent), Haiti (-56 percent), The Gambia (-56 percent), Bolivia (-54 percent), Honduras (-51 percent), Congo Republic (-48 percent), Central African Republic (-41 percent), Togo (-40 percent), Senegal (-37 percent), Cote d'ivoire (-34 percent), Guinea Bissau (-33 percent), and Zambia (-32 percent). Together, they will on average lose about 46 percent of their gross PBA allocations, or about SDR2.1 per capita per year, over IDA15-20 (Table 1). The significant declines in new IDA allocations will in tum have two operational implications: These countries will receive little or no fresh resources for development from IDA. Some countries (e.g., Togo and Cote d'ivoire) were in arrears, and have not yet seen real budgetary benefits from debt relief. Others such as Haiti have had small allocations to begin with, and exceptional access under the re-engaging countries category is ending (not to mention the vast need for recovery now), When MDRI netting out is added on top of this, fresh resources received for development would be minimal. It is assumed to reach the RIPC completion point by 2011 when the MDRI netting out willldck-in. Myanmar is another country projected to benefit from the MDRI netting out. While it is currently inactive, it is assumed to reengage with IDA by 2015 under the baseline scenario, after which it is projected to benefit from the MDRI netting out.

15 - 5 - Small or no allocations limit IDA's ability to engage in these countries - IDA's country presence, knowledge and diagnostic work, and policy dialogue all depend on whether meaningful level of resources are available to support the country. As country allocations fall to zero or to a negligible level, it becomes increasingly difficult to justify the administrative and overhead costs of engagement, including staff costs. Reduced engagement or complete disengagement will in tum negatively impact IDA's ability to playa leadership and coordinating role (i.e. "platform" role), since that leadership essentially emanates from country presence, knowledge and diagnostic work, and partnerships. 14. As requested by IDA Deputies, the options presented for consideration assess the impact on the key trade-offs discussed above: (i) allocation impact on the most negatively affected countries; and (ii) equity and fairness for non-mdri low-income countries. For all scenarios, the allocation impact is proxied by the impact on the 15 most negatively affected countries, while the "equity" impact is approximated by the allocation impact on the top 15 gainers from netting out. In addition to these two key tradeoffs, it would be important to consider how the mitigation scenarios would impact the PBA system. The MDRI netting out has already added considerable complexity to the PBA system,9 and any further options considered for mitigating the allocation impact should, to the extent possible, not add any further complexity. In the next section, the additional scenarios are examined in light of these tradeoffs. III. ADDITIONAL SCENARIOS FOR MITIGATING THE ALLOCATION IMPACT OF MDRI NETTING OUT 15. This section explores four additional scenarios for mitigating the negative allocation impact of MDRI netting out: (i) Capping the MDRI netting out at 50 percent of annual debt service forgone (rather than gross PBA allocation); (ii) Capping the MDRI netting out at 30 percent of the gross PBA allocation; (iii) Introducing a minimum per capita allocation that would allow IDA to meet the fixed cost of country engagement; and (iv) Placing a renewable moratorium on MDRI netting out for the period running through ID A 16 and to be reviewed in the context of the IDA 17 replenishment. Following the framework above, for each scenario, we will discuss the allocation impact and the equity effects. Annex 2 presents the detailed results pertaining to these scenarios. 16. The baseline scenario, which remains the same as the in the MTR paper, reflects projected new allocations under the MDRI netting out system with the following key assumptions: IDA resource envelope growth of 6 percent per replenishment; graduation from IDA of Armenia and Azerbaijan in FY12 and Georgia and Bosnia-Herzegovina in FY16; and relative performance of countries held constant at the FYlO level. 9 The MDRI netting out has added two additional steps to the PBA system--one for deducting debt service forgone and another for distributing compensatory resources. In addition, explaining annual variations in country allocations has become more difficult as the variations are no longer a function only of changes in performance but also changes in the amount of debt service foregone (i.e. the amount of netting out).

16 -6- A. Limiting the MDRI netting out to 50 percent of annual debt service foregone 17. This scenario limits the MDRI netting out to 50 percent of a country's annual debt service forgone. All of the 29 countries that are negatively affected in the baseline would receive a higher allocation under this scenario, with a total increase amounting to an average of about SDR191 million per year. The individual results are given in Annex 2, Scenario 1. Table 2. Allocation and equity impact of capping MDRI netting out at 50 percent of debt service foregone (A verage annual allocation; SDR million) IDAl5 IDAl6 IDAl7 IDA18 IDAl9 IDA20 15 Most Negatively Affected Countries Scenario 1: Cap netting out at 5 of debt service foregone Change from the baseline % 47% 58% 72% 6 All 29 Adversely Affected Countries Scenario 1: Cap netting out at 5 of debt service foregone Change from the baseline % 7% 8% 9% 12% 13% 15 Top Beneficiaries Scenario 1: Cap netting out at 5 of debt service foregone Change from the baseline % A.2% -4.8% -5.2% -6.6% -6.6% 18. Allocation impact. The 15 most negatively affected countries would receive, on average, a higher allocation of about SDR120 million per year during IDA 15-20, or about 50 percent of their average annual allocation (Table 2), The problem of zero or negligible allocation is fully overcome for the most negatively affected countries. 19. "Equity" impact. However, because the amount of compensatory resources available for allocation through the PBA system is now smaller (in fact 50 percent smaller), non-mdri countries that currently benefit in the baseline scenario would lose out in this scenario. The top 15 beneficiary countries would lose an average of about SDR150 million per year during IDA15-20, or about 5 percent of their average annual PBA allocation (Table 2). However, this scenario preserves the equity and moral hazard principles -and the link between compensatory resources and the PBA--although in a weakened form relative to the baseline.

17 - 7 - B. Limiting the MDRI netting out to 30 percent of gross PBA allocation 20. This scenario caps the maximum MDRI netting out at 30 percent of gross PBA allocation for MDRI countries whose debt service forgone exceeds that threshold (and continues with full netting out for those remaining under the threshold). Under this scenario, only 18 out of the 29 countries that are negatively affected by MDRI netting out would receive higher allocation (as only 18 of them have their debt service exceeding 30 percent of their PBA allocation), with a total increase in allocation to these countries of about SDR96 million per year, or about 4 percent of their average annual PBA allocation. The full results are given in Annex 2, Scenario 2. Table 3. Allocation and equity impact of capping netting out at 30 percent of gross PBA allocation (A verage annual allocation; SDR million) IDA15 IDA16 IDA17 IDA18 IDA19 IDA20 15 Most Negatively Affected Countries Scenario 2: Cap netting out at 3 of PBA allocation Change from the baseline % 25% 47% 66% 92% 73% All 29 Adversely Affected Countries Scenario 2: Cap netting out at 3 of PBA allocation Change from the baseline % 2% 4% 5% 7% 7% 15 Top Beneficiaries Scenario 2: Cap netting out at 3 of PBA allocation Change from the baseline % -1.3% % -3.1% -2.9% 21. Allocation impact. All of the 15 most negatively affected countries would receive a higher allocation under this scenario (as all have their debt service foregone above 30 percent of allocation). Their allocation would increase, on average, by about SDR131 million per year, or about 52 percent of their average annual allocation (Table 3). The problem of zero or a negligible allocation is fully overcome for the most severely affected countries. 22. "Equity" impact. As in the previous option, non-mdri countries that currently benefit from netting out would see their allocations reduced under this scenario. The top 15 beneficiaries would lose, on average, about SDR62 million per year over IDA 15-20, or about 2 percent of their average annual PBA allocation (Table 3).

18 - 8 - C. Introducing a minimum per capita allocation of US$2.5 (SDR1.6) 23. Another potential option to consider is to set a minimum per capita annual allocation below which country allocations are not allowed to fall. The basis for this is that IDA engagement and operational effectiveness requires a fixed minimum allocation per country. While IDA currently has a base allocation of SDRl.5 million per year per country (which benefits most small states with a population of less than 1.5 million), it is obviously not adequate in preventing country allocations from falling to negligible or zero due to MDRI netting out. 24. We consider the option of introducing a minimum per capita annual allocation of US$2.5 in addition to maintaining the current base allocation. lo Overall, 12 countries would receive higher allocations than under the base case scenario, only 9 of which are among the 15 most affected. The full results are given in Annex 2, Scenario 3. Table 4. Allocation and equity impact of a minimum per capita allocation of US$2.5 (Average annual allocation; SDR million) IDAI5 IDA16 IDA17 IDA18 IDA19 IDA20 15 Most Negatively Affected Countries Scenario 3: Per capita floor of US$ Change from the baseline % 12% 15% 22% 25% 19% All 29 Adversely Affected Countries Scenario 3: Per capita floor of US$ Change from the baseline % 1% 2% 2% 2% 15 Top Beneficiaries Scenario 3: Per capita floor of US$ Change from the baseline -S % -O.S% 0.8% -0.6% -0.6% -O.S% 25. Allocation impact. While no country receives zero or a negligible allocation under this scenario, only 9 of the 15 countries receive higher allocations. The total increase in allocation to these countries is, on average, about SDR28 million per year, or about 1 percent of their average annual PBA allocation (Table 4). 10 We also considered a variant of this option where the current base allocation (of SDR1.S million per year) is replaced altogether by the minimum per capita allocation of US$2.S, but the results were found to be unsatisfactory on three grounds: First, only 10 of the 29 countries (and only 5 of the IS most affected) that are negatively affected in the baseline would receive higher allocations. Secondly, six small countries would receive less than SDR2 million. Finally, this would weaken the PBA system in that the share of minimum allocation (not contingent on performance) would rise to about 21 percent of IDA envelope from the current level of 1 percent.

19 However, 6 of the 15 countries most adversely affected by MDRI netting out (Bolivia, Honduras, Togo, Senegal, Cote d'ivoire and Zambia) would receive lower allocations under this scenario as they would have to give up some of their allocations in order to "finance" the 12 countries whose allocations need to be increased to meet the minimum per capita allocation level ofus$2.s. Even though the difference in allocation is small, the minimum per capita floor exacerbates, rather than reduces, the MDRI related allocation problems for these countries. 27. "Equity" impact. The allocations of these countries would be reduced only marginally under this scenario-by about an average of SDR9 million over IDA 15-20, or about 0.3 percent of their average annual PBA allocation (Table 4). Of the options considered in this paper, this scenario involves the minimum equity trade-off, i.e. the sacrifice made by the top countries currently benefiting from the MDRI netting out is the least (and correspondingly the allocation gains of those adversely affected is the smallest). 28. While the zero or negligible allocations are fully overcome for the most affected countries, a key problem with this scenario is that 6 of the 15 countries most negatively affected are actually made worse off, receiving lower allocations (than under the baseline). D. Placing a moratorium on MDRI netting out for the IDA16 period 29. Finally, we consider an option where a temporary moratorium is placed on MDRI netting out running through the period of IDAI6. The moratorium would continue to address equity and moral hazard over the long-term but would temporarily suspend these underlying principles for the IDA 16 replenishment period. Many of the countries most negatively affected by MDRI netting out not only have high debt service foregone but they are also relatively weak performers where rapid performance improvements are rare. Yet, these countries will need increased ada, including through IDA allocations, to deal with the impact of the global crisis in the short-term and grow out of poverty, and accelerate progress towards the MDGs over the medium term. 30. A temporary suspension of the MDRI netting out, where the netting out would be restored once the crisis has ended, and growth has returned to the pre-crisis level, could move IDA in the direction of being able to provide more financial support for these poorest IDA countries so they are better able to weather the impact of the global crisis. Furthermore, Deputies could consider this option for implementation as early as FYII, extending through the IDA16 period. The moratorium could then be reviewed by the IDA Deputies in the course of the IDA 17 replenishment, with the possibility of extending it should Deputies decide there is still a need. 31. The results for this scenario are presented in Annex 2, Scenario 4. All the 29 adversely affected countries receive higher allocation during IDA16, with the increase amounting to an average of SDR286 million per year.

20 Allocation impact. The problem of zero or negligible allocation is fully overcome for the period of IDA16 only, and not for any period beyond IDAI6. The 15 most negatively affected countries would receive, on average, a higher allocation of about SDR187.5 million per year during IDAl6, or about 62 percent of their average annual allocation (Table 5). 33. "Equity" impact. These countries would see their allocations reduced by an average of about SDR211 million per year relative to the baseline as there will be no redistribution of compensatory resources for the IDA 16 period. In absolute terms, Bangladesh (-SOR13 million), Nigeria (-SORIO million), and Ethiopia (-SDR6.5 million) would lose the most in annual allocation. Table 5. Allocation and equity impact of putting a moratorium on netting out running through the IDA16 period (Average annual allocation; SDR million) IDA15 IDA16 IDA17 IDA18 IDA19 IDA20 15 Most Negatively Affected Countries Scenario 4: Putting a moratorium on netting out Change from the baseline % All 29 Adversely Affected Countries Scenario 4: Putting a moratorium on netting out Change from the baseline % 15 Top Scenario 4: Putting a moratorium on netting out Change from the baseline Because a moratorium would not address the medium term problem with zero and negative allocations, especially beginning in IDAl7 when the zero or negative allocations would become prominent, it ideally should be considered for renewal at that point, or combined with whatever final option is chosen to mitigate the allocation impact of netting out. Combining it with any final option would give a temporary additional relief to the countries most negatively affected by the MDRI netting out. IV. COMPARING THE OPTIONS 35. In this section, we compare the scenarios presented above and one from the MTR paper (reproduced in Annex 1) in terms of the key trade-offs. The objective is, as stated above, to strike a reasonable balance between mitigating the allocation impact on the most affected countries with maintaining as much fairness for non-mori low-

21 incorne countries as possible, preferably without adding any further cornplexity to the PBA systern. Annex 4 provides a cornparison of all the eight scenarios (the four presented above and the four frorn the MTR paper) in terrns of their allocation irnpact on the 15 countries rnost adversely affected as well as all 29 countries affected by MDRI netting out. Annex 5 cornpares the equity irnpact of all options (on the 15 top beneficiaries). With the exception of the cornplete elirnination of netting out, all scenarios rnaintain sorne degree of equity and rnoral hazard rnitigation in that not only do the non-ripc countries receive allocations frorn the "cornpensatory" resources but also that countries are not rewarded for accurnulating unsustainable debt in the past. I I 36. Of the four options presented at the MTR, only one option is included for consideration below: Based on the discussions and feedback received at the MTR, the option of elirninating the MDRI netting out is not included arnong the cornparisons presented here. Further, the MTR paper showed that Scenario MTR3 (capping the MDRI netting out at 75 percent of PBA allocation) and Scenario MTR4 (lirniting the redistribution of cornpensatory resources to MDRI-only countries) do not help address the zero or negligible allocation irnpact on the rnost affected countries-a key problern the scenarios are intended to address. In the case of Scenario 3, this is because rnost of the countries would not reach that capping lirnit. In the case of Scenario 4, the allocation irnpact on the 15 countries rnost affected is not large - resources sirnply shift to the better perforrners within the MDRI countries. For that reason, in cornparing the options, this paper retains scenario 2 frorn the MTR paper plus the new scenarios discussed above. 37. This leaves five scenarios for comparison: (i) Capping the MDRI netting out at 50 percent of gross PBA allocation (Scenario MTR2); (ii) Capping the MDRI netting out at 50 percent of debt service foregone (Scenario 1); (iii) Capping the MDRI netting out at 30 percent of gross PBA allocation (Scenario 2); (iv) introducing a minirnurn per capita allocation of US$2.5 (in addition to the current base allocation) (Scenario 3); and (v) Placing a ternporary renewable rnoratoriurn on MDRI netting out for the period running through IDA16 and to be reviewed in the context of the IDAl7 replenishrnent (Scenario 4). 38. Figure 1 presents a comparison of the allocation impact of tbese five scenarios for the 15 most adversely affected countries. Relative to the baseline, placing a rnoratoriurn on netting out yields the rnaxirnurn allocation irnpact (for IDAI6), followed by capping the MDRI netting out at 30 percent of PBA allocation (Scenario 2), capping the netting out at 50 percent of debt service foregone (Scenario 1), capping the MDRI II Combined with new instruments like the DSF and NCBP, the moral hazard concern is addressed and therefore not considered in detail under the individual scenarios.

22 12 - netting out at 50 percent of PBA allocation (Scenario MTR2), and introducing a minimum per capita allocation of US$2.5 (Scenario 3). 39. Figure 2 presents a comparison of the equity effects of the five scenarios. Relative to the base case, introducing a minimum per capita allocation of US$2.5 (Scenario 3) performs best in terms of equity (and minimizing losses to the top 15 gainers), followed by capping the MDRI netting out at 50 percent of PBA allocation (Scenario MTR2), capping the MDRI netting out at 30 percent of PBA allocation (Scenario 2), and capping the netting out at 50 percent of debt service foregone (Scenario 1). Note that the 15 top beneficiaries recei ve about 10 times that of the 15 most affected in terms of their average annual PBA allocation. The relative allocation gains (as percent of the PBA allocation) of those negatively affected are therefore generally much smaller than the equity losses of the top gainers. Figure 1: Allocation impact of the various scenarios on the 15 most negatively affected countries (Avera e annual allocation; SDR million) 500 II S Scenario MTR2: Cap MDRI netting out at 5 of PBA allocation Scenario 1 : Cap netting out at 5 of debtservice foregone Scenario 2:Cap netting out at 3 of PBA allocation Scenario 3: Set per capita floo r atus$ , IDA15 IDA16 IDA17 IDA18 IDA19 IDA20 Scenario 4: Putting a temporary moratorium on the MDRI netting outrunning through the IDA16 period

23 Figure 2: Equity impact of the various scenarios on the 15 top beneficiaries (Avera e annual allocation; SDR million) B --+- Base ine A- Scenario MTR2: Cap MDRI netting out at 5 of PBA allocation Scenario 1: Cap netting out at 5 of debt service foregone * "Scenario 2: Cap netting out at 3 of PBA allocation Scenari03: Set per capita floor at US$ IDA15 IDA16 IDA17 IDA18 IDA19 IDA20 Scenario 4: Putting a temporary moratorium on the MDRI netting out running through the IDA These options perform differently on the objective of striking a reasonable balance between increasing allocations to the most affected countries on the one hand and preserving the equity of treatment for the top beneficiary countries on the other (Table 6). Note that the impact on PBA complexity is similar for all scenarios (they all add one additional step of having to monitor the capping or the minimum per capita allocation), except the moratorium option, which helps temporarily simplify the PBA system The only two options that help simplify the PEA are: (i) eliminating the MDRI netting out (Scenario MTRI), and (ii) putting a temporary moratorium on the MDRI netting out (Scenario 4). The rest will all add one more step to the PEA system.

24 - 14- Table 6. Comparison of Options: 15 top beneficiaries vs. 15 negatively affected countries (Average annual allocation impact relative to the baseline during IDA 15-20; SDR million) Scenarios Capping at 5 Capping at Capping at Minimum per Putting a of gross PBA 5 of debt 3 of capita temporary allocation service gross PBA allocation of moratorium (Scenario MTRl) foregone allocation US$2.5 running through (Scenario 1) (Scenario 2) (Scenario 3) IDA16 (Scenario 4l Allocation Impact 15 most negatively affected countries 15 top beneficiaries Equity & Moral hazard PBA Complexity Preserved, but weaker One additional step Diluted One additional step Preserved, but weaker One additional step Preserved, with Principles of minimal equity are departure preserved but temporarily sus2ended One additional PBA system step simplified--le. two steps reduced (netting out plus redistribution of compensatory Capping the MDRI netting out at 50 percent of gross PBA allocation (Scenario MTR2) mitigates the zero or negligible allocation impact for all of the 15 most adversely affected countries. An important strength of this option is that it links the MDRI netting out to the size of the initial PBA allocation, and ensures that countries with a small PBA allocation (but high debt service foregone) are not hurt unduly by the MDRI netting out. However, the 50 percent capping may be too high a threshold to allow most of the affected countries to benefit from it. Indeed, under the baseline scenario, only 8 of the 15 most affected countries are projected to have debt service foregone in excess of 50 percent of their PBA allocation, with 7 of the countries not benefiting if this measure were chosen. Capping the MDRI netting out at 50 percent of debt service foregone (Scenario 1) mitigates the zero or negligible allocation impact for all of the 15 most adversely affected countries. However, it has the highest equity cost for the top beneficiaries, in part because the MDRI netting out is simply reduced by half without targeting that benefit to those countries with the lowest PBA allocations. In that sense, although it performs better allocation wise, it may not be the best option from an equity perspective.

25 Capping the MDRI netting out at 30 percent of the gross PBA allocation (Scenario 2) mitigates the zero or negligible allocation impact for all of the 15 most adversely affected countries. As in the MTR2 scenario, the strength of this option is that it links the MDRI netting out to the size of initial PBA allocation, helping to limit any excessive impact of netting out on the PBA allocation. Further, the threshold at 30 percent of the PBA allocation will allow all the 15 most affected countries to benefit from this option (as their average debt service foregone is projected to be above 30 percent of their PBA allocation during IDA 15-20), without imposing excessive equity cost on the beneficiaries (with only about 2 percent loss in the PBA allocation for the top 15 beneficiaries relative to the baseline). Introducing a minimum per capita allocation of US$2.5 (Scenario 3) increases allocation for only 9 of the 15 most adversely affected countries; and 6 of the 15 countries most adversely affected by MDRI netting out would receive lower allocation than under the baseline since they would have to be "taxed" in order to keep all countries at the minimum per capita level. Even though these 6 countries are not among those who face the problem of negligible or zero allocation due to netting out (and the difference in allocation is small), the minimum per capita floor can be seen as exacerbating, rather than reducing, the MDRI related allocation problems for these countries. For that reason, although it is an option with the best impact in maintaining equity for the top beneficiaries, it is not an option that seems to balance the two tradeoffs. Putting a temporary moratorium on the MDRI netting out running through the IDA16 period (Scenario 4) fully mitigates the zero or negligible allocation impact for all of the 15 most adversely affected countries. Indeed, it has the highest allocation impact on the 15 most affected countries, and the highest equity cost for the top 15 beneficiaries. The principles of equity are preserved but temporarily suspended as compensatory resources will not be distributed back to all IDA countries, but instead go back to the MDRI countries. Because a moratorium running through IDA16 would not address the medium term problem especially beginning in IDA17 when the zero or negative allocations would become prominent, it should ideally be considered for renewal at that point, or combined with another option chosen to mitigate the allocation impact of netting out over the medium-term. v. KEY ISSUES FOR DISCUSSION 41. If the current practice of MDRI netting out continues, a number of countriesincluding Chad, Guinea, Guyana, Haiti, Mauritania, the Republic of Congo, and The Gambia-will receive, starting from as early as IDAI7, little or no new IDA allocations. Alternative assumptions around country graduations from IDA-in particular of India and Vietnam-and continued growth of IDA resources (through replenishment) will not alter this trend significantly. The decline in new IDA allocations will in tum have