73363 . IDA16 MID-TERM REVIEW REVIEW OF IDA’S GRADUATION POLICY IDA Resource Mobilization Department Concessional Finance and Global Partnerships October 2012 Acronyms and Abbreviations Fiscal Year July 1 – June 30 AfDB African Development Bank AfDF African Development Fund ADB Asian Development Bank ADF Asian Development Fund CAS Country Assistance Strategy CFA African Financial Community CPS Country Partnership Strategy DeMPA Debt Management Performance Assessment DSA Debt Sustainability Analysis FSO Fund for Special Operations FY Fiscal Year GAVI Global Alliance for Vaccines and Immunization GNI Gross National Income GNP Gross National Product HIPC Heavily Indebted Poor Country IDB Inter-American Development Bank IBRD International Bank for Reconstruction and Development IDA International Development Association IFC International Finance Corporation IFI International Financial Institution IMF International Monetary Fund LIC Low-Income Country MDGs Millennium Development Goals MDRI Multilateral Debt Relief Initiative MIGA Multilateral Investment Guarantee Agency MTDS Medium-Term Debt Strategy OCR Ordinary Capital Resources ODA Official Development Assistance OECD Organization for Economic Cooperation and Development PBA Performance Based Allocation PREM Poverty Reduction and Economic Management PRGT Poverty Reduction and Growth Trust RDB Regional Development Bank SBL Single Borrower Limit WEO World Economic Outlook Table of Contents Executive Summary ..................................................................................................................... i I. Introduction ...................................................................................................................... 1 II. IDA Eligibility Criteria .................................................................................................... 2 Eligibility criteria ........................................................................................................... 2 Access to IDA Resources ............................................................................................... 4 Terms of IDA Financing ................................................................................................ 4 III. Graduation from IDA to IBRD ........................................................................................ 6 IV. IDA Graduates–Characteristics and Lessons for the Future ............................................ 9 V. Graduation from IDA ..................................................................................................... 14 Implications of Graduation ........................................................................................... 14 Potential IDA Graduates .............................................................................................. 15 Impact of Possible Graduations on IDA Resources ..................................................... 19 VI. IDA Graduation: Forward-Looking Issues .................................................................... 20 Alternative Measures of Relative Poverty.................................................................... 20 IDA Transitional Support ............................................................................................. 22 VII. Issues for Discussion...................................................................................................... 24 Charts Chart 1: Evolution of IDA Thresholds FY89-FY12 ...................................................................... 3 Chart 2: Graduation and Reverse Graduation Events .................................................................. 11 Chart 3: GNI per Capita in Year of Graduation ........................................................................... 12 Chart 4: Share of Concessional Debt in External Public Debt .................................................... 13 Chart 5: IDA Commitments as a Share of World Bank Group Commitments ............................ 17 Chart 6: IDA as a Share of External Public Debt Stocks 2010 .................................................... 18 Chart 7: World Bank Commitments as a Share of Total Public and Publicly Guaranteed Debt Commitments ................................................................................................................................ 18 Chart 8: Number of Poor in IDA Countries as of 2008 ............................................................... 23 Tables Table 1: Countries that Have Graduated from IDA ..................................................................... 10 Table 2: Blend Countries that Have Reverse Graduated to IDA-only Status .............................. 11 Table 3: Current Blend Countries ................................................................................................ 16 Table 4: 2011 Human Development Groupings for IDA- eligible Countries ............................. 21 Annexes Annex A: The Atlas method of calculating GNI and GNI per capita .......................................... 25 Annex B: Small States – Exceptional Access to IDA Resources ................................................ 26 Annex C: Graduation Policies from Concessional Financing – Select International Agencies .. 28 Executive Summary i. At the request of the IDA Deputies, this paper reviews IDA’s current graduation policy, including both criteria and process. IDA was established in recognition of the fact that, for many of the poorest countries, private capital and market based multilateral sources of financing were not adequate. From the outset, demand for IDA resources outstripped supply, hence the need for criteria to determine country eligibility for access to IDA resources and for graduation at the appropriate time. The eligibility and graduation criteria are constructed around: (a) the absence of creditworthiness; and (b) the concept of relative poverty, as measured by GNI per capita below an agreed threshold (US$1,195 for FY13). Both criteria have evolved over time reflecting–among others–the availability of IDA resources and developments in the global economy. The terms of IDA financing have also evolved becoming more differentiated in recognition of the varying stages of economic development of IDA recipients. ii. Graduation from IDA is a flexible process that typically extends over several years. The ultimate objective of IDA‟s graduation approach is to help countries make a successful and lasting exit from dependence on concessional resources and to avoid reversals back to IDA, which are costly for recipient countries and donors. Consequently, graduation has not been driven by a mechanistic approach but has relied on careful case-by-case evaluations. The current process begins with a positive assessment of creditworthiness and reclassification of a country from IDA-only to blend IBRD/IDA (“blend�) status and concludes with a reclassification from blend status to IBRD-only borrower with no further access to IDA resources. In the interim years, IBRD and other market based sources of financing are phased in and IDA and other concessional assistance to the country in question are phased out. The graduation process is normally triggered when a blend country exceeds the IDA operational cutoff for at least two consecutive years but IDA policies allow for exceptions to the criteria under special circumstances. iii. Since IDA’s inception, 36 countries have graduated from IDA and become IBRD- only borrowers. In 11 of these countries, adverse developments subsequent to graduation resulted in them becoming a “reverse graduate� and once again eligible for IDA resources. Almost 80 percent of the graduations took place prior to 1990 with 20 graduations occurring before 1980 and eight more in the subsequent decade 1980-1990. Of the countries that graduated before 1980, 17 did so successfully and three reverted to IDA eligibility. In sharp contrast, all of the countries that graduated in the period 1980-1990 subsequently reversed to IDA eligibility. There have been no reversals among the eight countries that graduated since 1990. iv. This paper examines the experience with past IDA graduations, including issues associated with graduation and how they have been addressed. The paper points out some marked differences observed in the characteristics of the countries that graduated before 1980 and those that did so in the subsequent decade. Pre-1980 graduates were, on average, richer at time of graduation and had limited dependence on IDA and relied to a much greater extent on non-concessional sources of financing prior to graduation. In contrast, the countries that graduated in the 1980s were poorer, more dependent on concessional financing and had more narrowly based economies dependent on commodity exports (often a single one). In addition, reverse graduations experienced by this group of countries were often the result of poor macro- ii economic management combined with volatile international commodity prices, over-borrowing on non-concessional terms in boom years (fueled by an abundance of commercial bank lending in the 1980s) and a downturn in the global economy. v. Graduation from IDA clearly marks a major milestone in a country’s progression through the process of economic and social development. It signals the important economic advances that the country has made, including having access to a broader range of credit sources and borrowing instruments. However, the combined impact of having to make accelerated repayments on IDA credits, which immediately increase debt servicing costs, and the likelihood of a lower volume of financing from the World Bank than in the pre-graduation phase – when both IBRD and IDA financing were available – may pose challenges to maintaining resource transfer commensurate with the graduating country‟s development financing needs. vi. Going forward, six countries look poised to graduate over the medium term. There are currently 11 IDA-eligible countries (excluding countries under the small island economies exception) classified as blend countries. Of these, Armenia, Bosnia and Herzegovina, Georgia, India, and Vietnam look poised to graduate over the IDA17 or IDA18 periods under current GNI per capita income projections. Angola, which has been confirmed as creditworthy and is in the process of being formally reclassified from IDA-only to blend, could also graduate during the same period. For these countries growth prospects are positive and their level of public and external debt is currently assessed as sustainable. These countries, however, differ significantly from prior graduates (even those that graduated during 1980-1990). Their dependence on concessional flows and IDA in particular, and more generally on the World Bank, is much higher. Consequently, some of these countries may be severely challenged to keep up the development momentum once graduation occurs as they would need to substitute for lost IDA resources and absorb the impact of accelerated repayments against a backdrop of continued global economic and financial uncertainty. vii. The paper examines the feasibility of using alternative measures of relative poverty to complement the per-capita income criterion. Notably, the paper discusses the merits of using a social index in conjunction with the per capita income criterion in order to identify the most vulnerable countries and to adapt IDA‟s response accordingly. However, it finds that this approach would be difficult to implement due to designing issues already identified in 2001 review of IDA graduation policy. viii. A number of potential IDA graduates have expressed the need for financing options that would smooth their transition from IDA-eligibility to IBRD-only status. The paper discusses the broad parameters that could guide such transitional support. The provision of transitional support has significant implications for the upcoming replenishment. Following discussions at the IDA16 Mid-Term Review, Management will undertake early discussions with IDA Deputies and IDA countries in the context of the IDA17 replenishment process to further elaborate the specifics of these options, and to assess their potential in terms of providing support to graduating countries. I. Introduction 1. The origin of IDA lies in the recognition that, for many of the poorest countries, private capital and market-based multilateral sources of financing existing at the time of its inception were inadequate. IBRD could not lend to these countries without damaging its own credit position in international markets and the countries themselves could not afford the debt servicing costs of loans on market terms to finance their development objectives. From the outset, demand for IDA resources outstripped supply, hence the need for criteria that determine eligibility for access to IDA resources, the level and terms of IDA‟s financing to eligible countries and graduation from IDA eligibility at the appropriate time. 2. At the request of IDA Deputies, this paper provides an overview of IDA’s graduation policy. The paper builds on the last review of IDA graduation policies conducted in 2001 and is particularly timely given that several countries could potentially graduate from IDA in the near to medium term.1 Currently, there are 16 IDA-eligible countries that have also access to IBRD (so-called “blend� countries). Within this group, 13 countries have a GNI per capita in excess of the IDA operational cutoff for more than two consecutive years, including five small islands with an exceptional access to IDA. In addition, Angola and Timor-Leste have been confirmed as creditworthy and are in the process of being formally reclassified from IDA-only to blend status and Nigeria may shortly be confirmed creditworthy. Overall, the experiences of volatility and fragility of economic and political progress in IDA countries suggest that graduation cannot be driven solely by mechanical formula but must be evaluated carefully on a case-by-case basis. In the past, 36 countries have graduated from IDA and become IBRD-only borrowers. Of these graduates, 11 have been granted renewed access to IDA due to adverse developments subsequent to their graduation. Similarly, adverse economic developments have also resulted in close to 20 countries, primarily in Sub-Saharan Africa, which were reclassified from blend status to IDA-only status. In addition, economic vulnerabilities, such as those faced by some small island economies, also mean that IDA graduation policy needs to be applied with flexibility and due regard of the diversity of country circumstances to ensure that development progress can be sustained. At the same time, it is expected that relatively richer IDA countries would exit in a timely manner to release IDA resources for other less wealthy IDA countries. 3. The paper summarizes IDA’s current graduation policy, including both the criteria and process involved. The paper also identifies key challenges and issues posed by forthcoming graduations. It is formulated as a discussion paper to facilitate the debate on graduation and – as this issue will be considered during the IDA17 replenishment negotiations – it does not present issues for decision. It is organized as follows. Section II sets out the eligibility criteria for access to IDA resources and IDA lending terms. Section III elaborates the process by which recipient countries are graduated from IDA, including how countries are assessed for creditworthiness and access to IBRD lending. Section IV reviews the history of IDA graduations and lessons to be drawn for future IDA graduations. Section V examines the impact of graduation on borrowing options and costs, presents potential IDA graduates in the near to medium term and what the graduation of these countries may mean in terms of IDA‟s 1 International Development Association: “IDA Eligibility, Terms and Graduation Policies,� January 2001. 2 evolving client base and geographic orientation. Section VI looks at options for alternate graduation criteria structured around indicators other than GNI per capita and examines options for transitional financing for creditworthy IDA-eligible countries. Section VII sets out issues for discussion. II. IDA Eligibility Criteria Eligibility criteria 4. “Graduation� refers to the formal determination that a country has attained a level of development that renders it ineligible for new IDA resources. The legal cornerstone for graduation is reflected in IDA‟s Articles of Agreement which prevent IDA from providing assistance if: (a) financing is available from private sources on terms that are reasonable for the recipient; or (b) could be provided by IBRD.2 Since IDA‟s inception in 1960, access to IDA resources has been determined by eligibility and graduation criteria constructed around the concept of relative poverty, as measured by GNI per capita, and absence of creditworthiness. 5. The relative poverty criterion for graduation has evolved over time. The threshold value of gross national income (GNI – referred to as GNP at the time) per capita for IDA eligibility was initially set at US$250 in 1964 (now known as the “historical cutoff�). By the early eighties, rising costs and a series of economic crises around the globe meant that available IDA resources were not adequate to fund programs for all countries below this eligibility ceiling. As a result, IDA ceased to lend to countries at the upper end of this per capita income scale, thereby creating a second and lower “operational cutoff�. IDA donors formally recognized the operational cutoff in IDA8 and have reaffirmed it in each subsequent replenishment round. Both the operational and the historical cutoff are reset annually on the basis of GNI per capita calculations made by the World Bank, in accordance with the Atlas conversion factor methodology. The purpose of the conversion factor is to maintain the real value of GNI by: (a) taking into account the impact of international inflation in the cross-country comparison of national incomes; and (b) smoothing out year-on-year fluctuations. It is understood that the GNI per capita does not completely summarize a country‟s level of development but it has proved to be a useful and easily available indicator of relative poverty.3 Annex A provides detailed information on how GNI per capita figures are derived in accordance with the Atlas methodology. 6. The operational cutoff was introduced in 1989 and set at US$580 (at 1987 prices). This was approximately 38 percent below its equivalent historical level and since then this differential has remained broadly unchanged: the same adjustment applies to both the historical and operational cutoffs. For FY13, the operational and historical cutoffs (based on 2011 data) 2 See Article 5, Section 1 of IDA‟s Articles of Agreement: http://www.worldbank.org/ida/articles-agreement/IDA-articles-of-agreement.pdf 3 See: The World Bank Group, Development Economics. “Crossing the threshold: an empirical review of the IBRD graduation threshold,� September 2010. 3 are US$1,195 and US$1,945, respectively.4 Relative to the global GNI per capita, the operational cutoff is much lower now than when it was established: it has not been adjusted to reflect the progression in global per capita income which has risen substantially. Whereas the operational cutoff in FY89 was equivalent to 18 percent of the global GNI per capita, in FY13 it is estimated to be equivalent to 13 percent. If the operational cutoff had maintained its relative value since FY82, in relation to the global GNI per capita, its nominal value would now be US$1,722, or 44 percent above the current value. This raises the question of whether the operational cutoff is set at an appropriate level and calls for prudence in the application of the per capita income criterion when considering countries‟ graduation from IDA. Chart 1: Evolution of IDA Thresholds FY89-FY12 2,500 2,000 1,500 1,000 500 0 2011 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2012 Operational cutoff Operational cutoff-to-World GNI per capita ratio constant over time Historical cutoff 7. The creditworthiness criterion that governs access to IBRD financing, and a country’s ability to graduate from IDA, has also evolved over time. This reflects the increased complexity of developments in the global economy and in emerging economies, and the evolving nature, depth and requirements of the international capital markets in which IBRD funds its operations. In the first decade or so of IDA operations, IDA credits, for countries with GNI per capita below the IDA eligibility threshold, were largely seen as complementary to IBRD financing. Very few countries were limited to only IDA financing and effectively classified as IDA-only borrowers: of the 20 countries that graduated from IDA prior to 1980 only four had periods in which they received only IDA resources and these periods were typically short – on average three to six years. Similarly, countries that borrowed from IBRD prior to the inception of IDA retained access to IBRD after IDA came into effect. This applied even in countries such as India, which at the time had a GNI per capita among the lowest in the world. Criteria for access to IBRD financing were, however, more stringently applied to countries that became a member of the World Bank after IDA was put in place. Oil price shocks of the seventies, widespread debt crises in the late seventies and the eighties, incidents of payments arrears to IBRD and reverse graduation of countries back into IDA served as catalysts for sharpening and 4 See Annex D of OP 3.10 on Financial Terms and Conditions of IBRD Loans, IBRD Hedging Products, and IDA Credits: http://siteresources.worldbank.org/OPSMANUAL/Resources/OP310_Annex_D_July_24_2012_Rev_Final.pdf 4 strengthening the formulation of creditworthiness criteria. The IBRD creditworthiness assessment process is described more fully below in Section III. Access to IDA Resources 8. Currently 81 countries are eligible for IDA resources.5 The majority of these countries meet both of the criteria for IDA-eligibility: lack of creditworthiness and a per capita income below the operational cutoff. However, some IDA-eligible countries have a per capita GNI above the operational cutoff but are not creditworthy (the so-called “gap countries�). Access to IDA resources for these countries is deemed temporary and the borrower in question is expected to undertake adjustment efforts designed to establish creditworthiness as rapidly as possible. IDA also recognizes the principle that no World Bank borrower should be left without access to either IDA or IBRD, provided that its performance is adequate and it is in good standing with the institution, i.e., with no payment arrears to IDA or IBRD. Therefore, IDA policies allow for exceptions under special circumstances. First, some small island economies are granted exceptional access to IDA resources even though they have a per capita GNI above the operational cutoff and – in some cases –have access to IBRD or other market based sources of financing. Annex B provides information on the rationale for this exception and the countries currently benefiting from it. Second, an allocation of IDA resources to middle income countries may be agreed by IDA donors when special circumstances prevail. Such instances, determined on a case-by-case basis, are rare and have been restricted to: (a) countries emerging from a prolonged period of conflict with unsustainable levels of public debt, protracted payment arrears, including to IBRD, and massive reconstruction needs (for example Serbia and, more recently, Iraq); or (b) newly created states with a fragile macro-economic structure, reconstruction needs, high or unsustainable levels of public debt, and obligations to IBRD, assumed upon independence as a separate state (for example Bosnia and Herzegovina). Terms of IDA Financing 9. The terms of IDA financing have also progressed, becoming more differentiated over time. Initially IDA had a single, 50-year maturity, lending instrument for all borrowers. Hardening of terms for a blend country, i.e., an IDA eligible country assessed as creditworthy for IBRD financing, was achieved through the mix and relative volume of IBRD and IDA financing. However, starting with IDA8 a differentiation in lending terms was introduced, in recognition of the varying stages of economic development of IDA recipients. As a result of the new policy, blend countries and gap countries began to have the level of concessionality of their IDA credits reduced. Until IDA15, the differentiation in terms remained relatively modest, particularly with regard to blend countries. However, a marked hardening of lending terms in conjunction with the IDA16 replenishment significantly increased the differentiation. The grant element of a regular IDA credit is 62 percent as compared to 35 percent for the newly introduced blend credit.6 IDA‟s current lending terms are summarized below: 5 See the list of countries in Annex D of OP 3.10 Op. Cit. 6 For a detailed review of the evolution of IDA‟s lending terms, pleas e see the IDA16 Financial Sustainability Working Group note “A Review of the History of IDA‟s Finances,� (August 2011). 5  Grants: IDA-only recipients, excluding gap countries, with a high risk of debt distress (as measured through the annual Debt Sustainability Analysis – DSA) receive their assistance on grant terms, and those with a moderate risk of debt distress receive 50 percent of their assistance on grant terms. Grants were first introduced in IDA13 and since then have accounted for about 20 percent of IDA commitments.7  Regular IDA credits: IDA-only recipients with GNI per capita below the cutoff and at low or moderate risk of debt distress receive all or 50 percent of their assistance, respectively, in regular IDA terms – 40-year maturity with a ten-year grace period, 0.75 percent service charge and a variable commitment charge;8, 9  Blend credits: For blend and gap countries – 25-year maturity with a 5-year grace period, 0.75 percent service charge, a variable commitment charge and an interest charge of 1.25 percent;  Hard-term credits: Countries classified as blend, excluding small islands, are eligible for hard-term credits – 25-year maturity with a 5-year grace period, 0.75 percent service charge, variable commitment charge and an interest charge corresponding to the IBRD fixed rate equivalent minus 200 basis points.10, 11 10. An Accelerated Repayment Clause has been included in IDA credits since 1987.12 It requires an accelerated repayment of IDA credits once the borrowing country has been assessed as creditworthy for IBRD loans and its GNI per capita has surpassed the eligibility GNI per capita threshold for a prescribed time period. The clause aimed at “enhancing IDA‟s ability, through quicker recycling of resources, to concentrate its resources on countries most in need.� The original provisions called for a doubling of scheduled principal payments when the GNI per capita of the borrowing country has remained above the historical cutoff for five consecutive years subject to a minimum grace period of ten years (the „old clause‟). This was subsequently amended in 1996 to a doubling of repayments to IDA when GNI per capita has exceeded the operational cutoff for three consecutive years subject to minimum grace period of 5 years (the 7 The grant portion of a recipient‟s allocation is discounted by 20 percent to address moral hazard and equity concerns and to compensate IDA for the charge income it forgoes when it provides grants. The charge-related portion (7 percent) is used to compensate IDA for forgone charges and is lent out through the hard-term window. The incentive-related portion (13 percent) is re-allocated to all IDA-only countries in proportion to their regular PBA allocation. Post-conflict countries are not subject to the incentive-related portion of the discount (but also do not benefit from the reallocation round). 8 All countries eligible for IDA under the small island exception, including those classified as a blend country, receive regular IDA credits regardless of their GNI per capita level. 9 The commitment charge is set by the Executive Directors on an annual basis between 0 and 0.5 percent. For FY13, the commitment charge is set at zero percent. 10 The hard-term credit window is funded with the charge-related portion of the volume discount on grants. Hard- term credits are additional to the regular PBA allocation. Their annual allocation is determined in proportion to an eligible country‟s PBA allocation. 11 The hard-term interest charge is set on an annual basis and is fixed at that rate for all credits approved during a fiscal year. For FY13, the hard-term window interest charge is set at 1.5 percent. 12 See “IDA Financial Policy Review,� IDA/R87-26, March 12, 1987. 6 „new clause‟).13 The new clause does not apply retroactively. Therefore, for credits issued between 1987 and 1996, the old clause is applicable; and for credits issued after 1996 the new clause is applicable. Credits issued prior to 1987 or those extended on hardened terms are not subject to the accelerated repayment clause.14 Instead of doubling the principal repayments, the borrower may request that IDA substitute an interest charge for some or all of the higher principal repayments, provided the new terms have a grant element equivalent to that resulting from doubling of the principal payments alone. Exercising the acceleration clause is subject to Board approval, after due consideration of the developments in the recipient country economy. During IDA16, the accelerated repayment clause in qualifying IDA credit agreements was exercised for eight graduates: Albania, Azerbaijan, China, Egypt, Equatorial Guinea, Indonesia, FYR of Macedonia and St. Kitts and Nevis.15 III. Graduation from IDA to IBRD 11. The graduation criteria are founded on the assumption that the capacity to meet the higher debt service costs associated with IBRD and other market based financing increases as countries move along the development path and national income rises. The World Bank recognizes three categories of borrowers:  IBRD-only: countries that are creditworthy and have no access to IDA resources.  IDA-only: countries that are not creditworthy and have no access to IBRD loans.  Blend: countries eligible to borrow from both IDA and IBRD. 12. Graduation from IDA is a flexible process that typically extends over several years. During these years, IBRD and other market based sources of financing are phased in and IDA and other concessional assistance to the country in question are phased out. It begins with a positive assessment of creditworthiness and reclassification of a country from IDA-only to blend status and concludes with a reclassification from blend status to IBRD-only borrower with no further access to IDA resources. The graduation process normally starts when a blend country exceeds the IDA operational cutoff for at least two consecutive years. 13. Prior to borrowing from IBRD a country must first be assessed as creditworthy for IBRD lending. A comprehensive credit assessment is conducted by IBRD. The assessment is based on an evaluation of eight broad components of creditworthiness: political risk, external debt and liquidity, fiscal policy and public debt burden, balance of payment risks, economic 13 See “Modification to the Accelerated Repayment Terms of IDA Credits,� IDA/R96 -145, July 23, 1996. 14 Starting in IDA13 (July 2002), IDA intr oduced “hardened term� credits for so called gap countries. The IDA credit maturity for gap countries was shortened to 20 years with a ten-year grace period, with the standard service and commitment charges, but for these credits the accelerated repayment clause did not apply. Starting in IDA16 (July 2011), hardened terms were discontinued and gap countries receive 25-year blend credits that are subject to accelerated credit repayments. 15 See “Acceleration of Credit Repayments to IDA and New Policy Framework for Voluntary Prepayments,� IDA/R2010-0351, November 11, 2010; and “Update on Acceleration of Credit Repayments to IDA,� IDAR2012-0178, June 26, 2012. 7 structure and growth prospects, monetary and exchange rate policy, financial sector risks, and corporate sector debt. This includes a comprehensive analysis of short and long term vulnerabilities against the backdrop of individual country circumstance and its links to the global economy. IBRD, in contrast to most private lenders, does not charge a risk premium to borrowers and its loans have a much longer maturity. Its assessment of creditworthiness may differ from the market‟s perception of the creditworthiness of a country. The creditworthiness assessment may entail detailed desk work as well as a mission to the country for the credit risk analysts to meet with government officials, the donor community, academics and the private sector. Executive Directors are typically informed of creditworthiness decisions during board discussions on a new Country Assistance Strategy (CAS)/Country Partnership Strategy (CPS) or CAS/CPS progress report, for example the April 2012 Country Partnership Strategy for Mongolia.16 Changes in country classifications are reflected in Annex D of OP 3.10 following such Board discussions. 14. Creditworthiness assessments are rigorous. History has shown that pushing countries into market based borrowing before their economies are sufficiently robust to absorb the associated higher debt servicing costs is counter-productive and likely to lead only to situations where countries have to reverse graduate into IDA. Unlike most private lenders, IBRD does not differentiate its loan pricing across borrowers of varying creditworthiness but extends broadly the same terms to all its borrowers. Therefore, IBRD must mitigate risk through prudent lending practices, alignment of lending volumes and comprehensive assessment of the perceived risk factors for each IBRD borrower. These are central to maintaining the highest credit standing and IBRD‟s AAA rating. 15. Following a positive assessment of creditworthiness and the process described in paragraph 13 above, a country is reclassified to blend status. Blend status typically increases World Bank resources to a country and also expands the types of financial products and services available. In addition to its performance-based IDA allocation a blend country gains access to IBRD resources. The level of IBRD and IDA assistance during this period is set out in a CAS with the combined IBRD and IDA lending volume set in relation to the country‟s creditworthiness, its policy and portfolio performance and its need for development finance. As with all other IDA recipients, the IDA lending envelope for a blend country is individually determined through the performance-based allocation (PBA) system. For some blend countries, e.g., India and Pakistan, the IDA allocation is “capped� because of the size of the economy and population, and to reflect their broader financing opportunities. 16. While in blend status, countries are typically supported by economic and sector work and technical support from the World Bank Group aimed at improving their access to market based sources. The International Finance Corporation (IFC) plays a critical role in assisting with private sector development through its investment and advisory services and its broad suite of financial products including loans, equity, trade finance, structured finance, syndicated loans and risk management instruments. Similarly, the Multilateral Investment 16 See “International Bank for Reconstruction and Development and International Development Associat ion and International Finance Corporation and Multilateral Investment Guarantee Agency Country Partnership Strategy for Mongolia for the Period FY2013-2017,� Report No. 67567-MN. http://imagebank.worldbank.org/servlet/WDSContentServer/IW3P/IB/2012/04/27/000333037_2012042701024 0/Rendered/PDF/675670CAS0P1250Official0Use0Only090.pdf 8 Guarantee Agency (MIGA) can provide a range of advisory services and investment insurance products to attract foreign investors into critical sectors of the economy. 17. Effective and sustainable public debt management is key to countries’ successful transition to IBRD-only borrower status. The complexity of managing the public debt portfolio and formulation of a coherent debt management strategy increases significantly in parallel with the shift to non-concessional sources of financing. The World Bank places a priority on supporting countries in this critical area and offers a wide range of technical and advisory services in all aspects of public debt management, grounded in international best practice and extensive cross-country experience. These include the Debt Management Performance Assessment (DeMPA), institutional reform programs and Medium-Term Debt Management Strategy (MTDS) and advisory services and training seminars in debt management, capital market development and asset and liability management.17 18. The volatility and fragility of economic and political progress in many IDA countries warrant careful case-by-case evaluation of a country’s capacity to sustain the move from concessional to market based sources of financing. Consequently, there is no mechanical formula that sets a timetable for graduation from IDA and dictates the length of the transition phase. It depends on several factors including individual country circumstance and access to financing from other bilateral and multilateral sources. However, the aim is for relatively richer IDA countries to exit from IDA in a timely manner so as to release IDA resources for the other poorer IDA countries. The expectation is that countries will transition from blend status to IBRD-only borrower within a reasonable timeframe when their GNI per capita exceeds the operational cutoff for more than two consecutive years. For example, Albania was declared creditworthy and reclassified to blend status in FY06 and graduated to an IBRD- only borrower in FY09 after one three-year CAS cycle. 19. Nothing precludes a country from graduating from IDA before its per capita income reaches the operational cutoff if it has been assessed as creditworthy and is able to meet its financing needs from IBRD and other commercial sources. Such countries would normally have strong export earnings and large international reserves, a good credit rating and a demonstrated track record of borrowing in international capital markets. Historically, a few countries graduated from IDA before their GNI per capita reached the IDA eligibility threshold. China, the most recent example, ceased to borrow from IDA at end-FY99 when its GNI per capita was US$840, as compared to the then prevailing operational cutoff of US$925. From time to time countries have also graduated from IDA on an accelerated basis because of factors that favorably impacted the assessment of their creditworthiness, often in conjunction with a rapid rise in their national income levels. Equatorial Guinea graduated from IDA in FY99, after the discovery of substantial petroleum reserves significantly improved the country‟s debt servicing capacity and pushed up GNI and GNI per capita sharply, aided by the very small size of the population. 20. The graduation policies of other multilateral institutions in respect of their concessional financing facilities are, for the most part, closely aligned with those of IDA . The International Monetary Fund (IMF) and regional development banks, such as the African 17 See http://go.worldbank.org/A6WVE6SDV0 and http://treasury.worldbank.org/index.html 9 Development Fund (AfDF), the Asian Development Bank (ADB), and the Inter-American Development Bank (IDB), all employ an income and a creditworthiness criterion. The IMF also takes into account short term vulnerabilities. Following revisions to the graduation policies of the Poverty Reduction and Growth Trust (PRGT) in 2010 the IMF differentiates between the income and market access criterion. Under the new policies a country is no longer eligible for the resources if it: (a) has a persistently high level of income; or (b) durable and substantial access to financial markets; and does not face serious short-term vulnerabilities regardless of its per capita income level. In common with IDA, each institution regularly reviews and updates its income criterion to reflect changes in national income levels in recipient countries. Without exception, these institutions base their income criterion on the World Bank GNI per capita data, calculated according to the Atlas methodology. Annex C presents a detailed review of the graduation policies of the concessional financing facilities of select multilateral agencies. 21. The graduation policies of bilateral aid agencies differ from one donor country to another and are found to be indirectly aligned with those of IDA. That said, reclassification by the World Bank of a country from low to middle income, or graduation of a country from IDA typically serves as the trigger for a gradual phasing out of Official Development Assistance (ODA) from bilateral donors and hardening of terms on which bilateral loans are extended. This is evidenced by information reported by OECD member countries to the Development Assistance Committee and the OECD Creditor Reporting System on aid and other official financing. In addition, the scope and scale of technical assistance, capacity building and advisory services received from donors may also be reduced. IV. IDA Graduates–Characteristics and Lessons for the Future 22. Since IDA’s inception, 36 countries have graduated from IDA and become IBRD- only borrowers. The majority of these countries were never classified as IDA-only borrowers and graduated from blend status to IBRD-only borrowers. Table 1 lists the countries and the fiscal year in which they received their last IDA credit. In 11 of these countries adverse developments subsequent to graduation resulted in them becoming “reverse graduates� and once again eligible for IDA resources. Three countries, Egypt, Indonesia and the Philippines, subsequently regained creditworthiness and became IBRD-only borrowers. They first graduated from blend status to IBRD-only borrower in FY79 (Philippines), FY80 (Indonesia) and FY81 (Egypt) and were reclassified to blend status in FY90 (Philippines), FY91 (Egypt) and FY99 (Indonesia). The Philippines regained IBRD-only status within two years in FY93, Egypt after eight years in FY99 and Indonesia after a decade in FY09. Among other reverse graduates, Papua New Guinea was reclassified from IBRD-only to blend status in FY03 and retained that classification thereafter. The other seven “reverse graduates� are currently not considered creditworthy and are still classified as IDA-only countries.18 18 Nigeria may shortly be confirmed as creditworthy and reclassified to blend status. Zimbabwe is an exception. It is classified as a blend country but this is notional since the country has no access to IBRD or IDA resources on account of its protracted payment arrears to both IBRD and IDA. 10 Table 1: Countries that Have Graduated from IDA Country Fiscal year Year of Country Fiscal year Year of of last IDA Reverse- of last IDA Reverse- credit Graduation credit Graduation to IDA to IDA Chile FY61 Thailand FY79 Costa Rica FY62 Philippines FY79; FY93 FY91 Colombia FY62 Honduras FY80 FY91 Nigeria FY65 FY89 Indonesia FY80; FY08 FY99 Dominican Republic FY73 Cameroon FY81 FY94 Korea FY73 Egypt FY81; FY99 FY91 Côte d'Ivoire FY73 FY92 Nicaragua FY81 FY91 Turkey FY73 Congo Republic FY82 FY94 Botswana FY74 Papua New Guinea FY83 FY03 Ecuador FY74 Zimbabwe FY83 FY92 Syria FY74 Equatorial Guinea FY93 Mauritius FY75 St. Kitts and Nevis FY94 Swaziland FY75 China FY99 Morocco FY75 Macedonia, FYR FY02 El Salvador FY77 Serbia FY07 Paraguay FY77 Albania FY08 Tunisia FY77 Montenegro FY08 Jordan FY78 Azerbaijan FY11 23. There are another 17 IDA-only countries that were, at one point in time, assessed as creditworthy for IBRD financing and classified as a blend country, but which subsequently reversed to IDA-only status (Table 2). Many of these countries encountered difficulties in servicing their obligations to external creditors and required extraordinary support from donors, up to and including in the context of the Heavily Indebted Poor Countries (HIPC) Initiative and Multilateral Debt Reduction Initiative (MDRI). The Special Program of Assistance to Low- Income, Debt-Distressed Countries in Sub-Saharan Africa launched in 1987 included a provision for supplemental concessional financing, primarily from IDA, to offset debt service owed to IBRD (the Fifth Dimension).19 24. The history of IDA graduation provides useful insights on the factors associated with successful (or unsuccessful) graduation outcomes. A review of the countries that have graduated from IDA serves to identify some of the reasons why, in some cases, the process has not worked causing countries to reverse graduate to IDA. Outcomes are found to be closely linked to: (a) country circumstance at the time of graduation; (b) the level of reliance on IDA funding; (c) the track record of access to international capital markets; and (d) economic structure and vulnerability to exogenous shocks. 19 For more information on the Special Program of Assistance to Low-Income, Debt-Distressed Countries in Sub- Saharan Africa see Gamarra, Pollock, and Braga: “Debt Relief to Low-Income Countries: A Retrospective� in “Debt Relief and Beyond: Lessons Learned and Challenges Ahead,� 2009, from Braga and Doemeland (Editors). 11 Table 2: Blend Countries that Have Reverse Graduated to IDA-only Status Year of Last IBRD Year of Last IBRD Country Commitment Country Commitment Burundi 1966 Malawi 1987 Congo Dem.Rep. 1969 Mauritania 1965 Ethiopia 1973 Senegal 1981 Ghana 1979 Sierra Leone 1976 Guinea 1975 Sudan 1978 Haiti 1963 Tanzania 1981 Guyana 1982 Togo 1978 Liberia 1983 Zambia 1985 Madagascar 1976 Note: Does not include IBRD loans extended to countries prior to independence under the guarantee of the then colonial power 25. Almost 80 percent of IDA graduations took place prior to 1990 with 20 graduations occurring before 1980 and eight more in the subsequent decade 1980-1990 (see Chart 2). Of the countries that graduated before 1980, 17 did so successfully and only three (Côte d‟Ivoire, Nigeria and the Philippines) subsequently experienced a downturn in their macro-economic situation and reverted to blend status and, in the case of Côte d‟Ivoire and Nigeria, subsequently to IDA-only status. In sharp contrast to this, all of the countries that graduated in the eighties subsequently reversed to IDA-eligibility. There have been no reversals among the eight countries to graduate since 1990. Included in this group are China and six independent republics of the former Soviet Union or former Yugoslavia that gained access to IDA largely on account of the transformational shock related to transition from command to market economy. Their graduation from IDA took place once they went through market reforms. Chart 2: Graduation and Reverse Graduation Events Chart 2.A. Graduation Events Chart 2.B. Reverse Graduation Events 20 10 Countries that subsequently reverse graduated Countries that kept IBRD-only 8 15 status 6 10 4 5 2 0 0 1960s 1970s 1980s 1990s 2000s 2010s 1960s 1970s 1980s 1990s 2000s 2010s 26. There are some marked differences in the characteristics of the countries that graduated from IDA before 1980 and those that did so in the subsequent decade. Pre-1980 graduates were, on average, richer at the time of graduation than those that graduated during the 1980s (see Chart 3). At the time of graduation, most countries graduating prior to 1980 had GNI 12 per capita significantly above the IDA eligibility threshold. In contrast, most of those that graduated in the 1980s did so prior to reaching the GNI per capita threshold or when it was only moderately above. Exceptions are the Republic of Congo which had a GNI per capita 60 percent above the threshold when it graduated in FY82 and Zimbabwe which graduated in FY83, following independence in 1980, with GNI per capita of US$1,080 compared to a threshold of US$795. On an inflation-adjusted basis the difference in the relative wealth of the two groups of graduates is even more marked. Comparisons with more recent IDA graduates are less instructive as many of them are atypical in that they were granted access to IDA under the small island exception (St. Kitts and Nevis) or as newly emerging states after prolonged period of conflict (FYR of Macedonia, Montenegro, Serbia). In such circumstances it is to be expected that GNI per capita at time of graduation may be high. Chart 3: GNI per Capita in Year of Graduation 1,400 GNI per-capita at the time of graduation 1,200 1,000 800 (US$) 600 400 200 0 FY70 FY71 FY72 FY73 FY74 FY75 FY76 FY77 FY78 FY79 FY80 FY81 FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 Historical cutoff Note: Dots represent countries that graduated during the 1970s. Triangles represent countries that graduated during the 1980s. 27. Most of the early graduates had very limited dependence on IDA, and relied to a much greater extent on non-concessional sources of financing prior to graduation than countries that graduated in the 1980s (see Chart 4). For countries graduating from IDA before 1980, concessional debt, as a share of outstanding external public debt at graduation, was on average a little over 20 percent, down from 26 percent five years prior to graduation. In sharp contrast, the comparable share for countries graduating in the eighties was on average almost 35 percent at graduation down from 42 percent five years prior. In part this can be explained by an economic structure which was typically much more diversified, and with some manufacturing base, for countries that graduated in the seventies. Those that graduated in the 1980s were, without exception, countries with narrowly based economies dependent on commodity (often single commodity) exports such as oil or agricultural products. In addition, the fact that IDA was in its start up phase, and consequently the line between access to IDA and IBRD financing more fluid, was also an important factor. The first IDA graduates (Chile, Colombia and Costa Rica) 13 each received a single IDA credit before the IDA eligibility per capita income criterion came into effect in FY64. Similarly, the majority of the pre-1980 graduates had been IBRD borrowers prior to the inception of IDA and never classified as IDA-only. Those that were, for example Botswana or Korea, were IDA-only borrowers for a very limited time and received proportionally far more IBRD than IDA financing once creditworthiness was established. Chart 4: Share of Concessional Debt in External Public Debt in Years Prior to IDA Graduation 45 40 35 (percent) 30 25 20 G-5 G-4 G-3 G-2 G-1 Graduation Countries graduating in the 70s Countries graduating in the 80s 28. Additional factors account for why some IDA graduates were not able to sustain IBRD-only borrower status and reversed to IDA-eligibility. Often it was the result of a juxtaposition of circumstances arising from poor domestic macro-economic management combined with a downturn in the global economy, for example the reverse graduation of Indonesia in 1999 in the wake of the Asian crisis. For Egypt and the Philippines a determining factor was the first Gulf War which pushed up the cost of oil imports and led to a loss of vital inflows of workers‟ remittances. In the main however, the cause can be traced to sharp fluctuations in commodity prices coupled with exuberant over-borrowing in boom years, fueled in turn by an abundance of commercial bank lending in the eighties and, with benefit of hindsight, an over-optimistic view of the macro-economic prospects of many developing countries. The 1992 parity change in the value of the CFA franc exacerbated the problem for countries such as Côte d‟Ivoire, Cameroon and Congo. With the exception of Papua New Guinea countries that reverse graduated to IDA were over-indebted and most had already restructured their external claims with bilateral and commercial creditors by the time of reverse graduation and in subsequent years received debt relief from official and commercial creditors, including five countries eligible for debt relief under the HIPC Initiative and the MDRI (Cameroon, Republic of Congo, Côte d‟Ivoire, Honduras and Nicaragua). 14 V. Graduation from IDA 29. Graduation from IDA marks a major milestone in a country’s progression through the process of economic and social development. It signals the important advances that the country has made and opens up its development financing options to a much broader range of credit sources and borrowing instruments. However, it may also add to the challenge of maintaining a transfer of resources commensurate with the graduating country‟s development financing needs. Implications of Graduation 30. Upon graduation new IDA commitments will cease. The inflow of IDA resources will decline as the existing pipeline is drawn down and turns negative once this process is complete, typically within four to five years after graduation. Thereafter, the country will be making a net transfer of resources to IDA. 31. Post graduation, overall borrowing costs will rise. This is due to the shorter maturity and higher interest costs associated with market based sources of financing. The country may also expect to be subject to more rigorous borrowing criteria and much closer scrutiny by creditors of its macro-economic prospects and debt servicing capacity. Markets‟ assessment of these factors will be reflected in credit risk ratings which, in turn, will drive the risk premium associated with financing raised in international capital markets. 32. Access to external financing may be constrained by market perception or the country’s debt servicing capacity and as a result external financing may be much lower than the concessional inflows received prior to graduation. Particularly in cases where concessional flows were large, in relation to the recipient country‟s GNI, the country may find it is not possible to borrow an equivalent amount on non-concessional terms and retain public debt levels within prudent and sustainable limits. 33. The graduating country is also likely to have access to a lower volume of financing from the World Bank than in the pre-graduation phase when both IBRD and IDA financing were available.20 Unlike most private lenders, IBRD does not differentiate across borrowers with varying creditworthiness but instead extends broadly the same terms to all its borrowers. This requires IBRD to mitigate for risk to its portfolio through prudent lending practices, and alignment of country lending volumes with the perceived risk factors of each IBRD borrower: the latter determined through regular (at a minimum annual) credit risk assessments of all IBRD borrowers. Moreover, the advantageous loan terms IBRD is able to offer, a direct reflection of its AAA rating, make it an attractive source of financing for borrowers and competition for IBRD resources is strong. This is especially so in periods of global economic downturn when many countries look to IBRD as a lender of last resort. 20 In the period prior to being declared blend, however, the country would have had access only to IDA resources. The level of World Bank support when the country was IDA-only compared with the level of support it would receive when IBRD-only may not show a significant decline. 15 34. Graduation will also trigger the IDA accelerated repayment clause, lowering the net transfer of resources from the World Bank and raising debt servicing costs. The rationale for the IDA accelerated repayment clause, described in section II, is firmly anchored in policies to ensure that IDA resources are allocated equitably, particularly given the needs of the countries with the lowest income. IDA graduates have much stronger growth prospects and higher GNI per capita than IDA eligible countries and therefore are judged to be in a position to reimburse IDA more rapidly and, in turn, raise the level of net reflows to IDA. From the perspective of the IDA graduate, however, the impact of increased debt service costs may be severe, particularly for countries where the level of new external financing from IBRD or other market sources is constrained, and the share of IDA in total external debt service payments is high. Considering its potential impact, the current practice is to notify countries one year in advance of the actual date of the acceleration of repayments to IDA. Potential IDA Graduates 35. A cluster of IDA eligible countries look poised to graduate from IDA in IDA17 or IDA18. This section identifies these potential graduates and takes a closer look at their characteristics and what graduation might imply for them, especially in terms of resources from the World Bank. It also looks more broadly at the resource implications for IDA as more countries pass the important graduation milestone. 36. Currently, there are 11 IDA-eligible countries (excluding five countries under the small island economies exception) that have blend status. In addition, Angola and Timor- Leste have been confirmed as creditworthy and are in the process of being formally reclassified from IDA-only to blend status and Nigeria may shortly be confirmed creditworthy. Of the group of 11 blend countries eight already have GNI per capita in excess of the IDA operational cutoff for more than two consecutive years (Table 3). 37. There is considerable variation in the length of time countries have been in blend status and their access to IBRD resources. At one end of the spectrum are India and Pakistan, among the earliest IBRD borrowers and reclassified as blend following the inception of IDA in 1961. At the other end are Mongolia and Sri Lanka, reclassified from IDA-only to blend in FY12. Bosnia and Herzegovina gained access to IDA under exceptional circumstances following prolonged conflict and subsequent independence from the former Federal Republic of Yugoslavia in 1996. The country obtained its first financing from IBRD in FY09. Bolivia has only notional blend status. Since it was classified as a blend borrower in FY02 it has contracted only one IBRD loan for US$15 million, of which all but US$150,000 was subsequently cancelled. Similarly, Papua New Guinea has not borrowed from IBRD since it reversed graduated from IBRD-only to blend status in FY03. Uzbekistan was an IBRD country until FY02 when it gained access to IDA and was classified as blend because its GNI per capita was below the IDA operational cutoff and creditworthiness had deteriorated. There was no IBRD financing from FY03 until FY11. By contrast, IBRD financing to Sri Lanka in the first year of blend status was US$213 million, equivalent to the average annual financing from IDA in the prior two years. 16 Table 3: Current Blend Countries GNI per capita Years above Year reclassified to WEO growth operational cutoff blend IBRD/IDA forecast (2011, US$) status (2012-17, percent) IBRD/IDA Blend Countries (16) 1) Non-Small Islands Exception (11) Armenia 3,360 8 FY07 4.0 Bolivia (Plurinational State of) 2,040 6 FY02 5.0 Bosnia and Herzegovina 4,780 15 FY96 2.5 Georgia 2,860 8 FY09 5.7 India 1,420 3 FY61 6.3 Mongolia 2,320 5 FY12 11.4 Pakistan 1,120 0 FY61 3.5 Papua New Guinea 1,480 3 FY03 8.3 Sri Lanka 2,580 8 FY12 6.6 Uzbekistan 1,510 2 FY02 6.3 Vietnam 1,260 1 FY10 6.5 2) Small Islands Exception (5) Possible IBRD/IDA Blend Countries (2) Angola 4,060 8 5.6 Nigeria 1,200 2 6.7 Timor-Leste n.a. n.a. n.a. Memorandum Items Operational cutoff 1,195 Historical cutoff 1,945 Sources: Global Development Finance and World Economic Outlook. Notes: 1/ With the exception of Bosnia and Herzegovina, the number of years a country has remained above the operational cutoff reflects the figures used as part of the IDA allocation exercise. For Bosnia and Herzegovina, the latest available figures in GDF were used. 2/ The number of years in blend status are counted since the year of last reclassification from one category of Bank borrower to another. 3/ Angola and Timor-Leste have been confirmed as creditworthy and are in the process of being formally reclassified from IDA-only to blend status. 4/ Nigeria may shortly be confirmed creditworthy. Nigeria reverse graduated to IDA in FY89 as a blend country and was reclassified as an IDA-only country in FY2005. 38. The growth prospects for these countries are positive. The latest World Economic Outlook (WEO) forecast for the 11 blend countries is for annual real GDP growth to average six percent over the period 2012-17. However, there are downside risks that should be considered. For Armenia, Bosnia and Herzegovina and Georgia, which are still coping with the impact of the 2009 global recession, the outcomes are intrinsically tied to those of the Euro-zone to which their economies are closely linked. Thus, a deeper and longer recession in the Euro-zone constitutes a major downside risk to the forecast. Many have economies heavily dependent upon one or more primary commodities with few linkages to other areas of the economy. Price fluctuations constitute a major downside risk for these economies. The countries with some export manufacturing base, for example India and Vietnam, are vulnerable to shortfalls in global demand. A slow recovery or further downturn in global economic conditions may also have a knock on effect on workers‟ remittances on which countries like Armenia, Pakistan and Sri Lanka depend heavily. 39. Most of these countries have a sustainable level of public and external debt, a key indicator of capacity to borrow on non-concessional terms. A review of the most recent debt sustainability analysis, conducted jointly by the IMF and the World Bank, indicates that, on average, debt levels are moderate, measured against key macro-economic variables and that for most countries the risk of debt distress is low even under adverse conditions and combined shocks. Half of the countries in the group have accessed international capital markets in the past two years and placed ten year bonds at fixed coupon rates averaging around six percent. 17 40. Based on the above considerations, six countries could potentially graduate to IBRD-only status during the period covered by IDA17 and IDA18. While graduation will be determined based on an assessment performed with the latest available information, six countries (Angola, Armenia, Bosnia and Herzegovina, Georgia, India and Vietnam) could potentially graduate in the upcoming two replenishments. At this point, these countries present a confluence of factors, including their growth prospects as well as their overall macro-economic and debt sustainability outlook. Bolivia and Uzbekistan may be candidates for graduation but in the case of Bolivia, any consideration would be subject to the outcome of a creditworthiness assessment scheduled to take place by end 2012. Pakistan has blend status but its GNI per-capita is still below the IDA operational threshold. Furthermore, its graduation could be hampered by low growth rates and internal security problems. The other three countries are unlikely to graduate during the next two replenishment periods on account of their lack of access to IBRD (Papua New Guinea) or their newly acquired blend status (Mongolia and Sri Lanka). 41. With the exception of Angola, these potential graduates differ significantly from earlier ones in that they are more dependent on IDA for external financing. This is measured as a share of financing from the World Bank and as a share of overall public and publicly guaranteed external debt. Consequently, potential graduates may be severely challenged to keep up the development momentum once graduation occurs and they are confronted with the need to substitute for lost IDA resources and, in parallel, meet much higher debt service payments to IDA on account of implementation of the accelerated repayment clause. If graduations take place against a backdrop of global economic and financial uncertainty, these challenges may be further exacerbated. For countries graduating in the 1970s and 1980s, IDA commitments, as a share of total commitments from the World Bank, averaged close to 13 percent and 22 percent, respectively, in the year of graduation and five preceding years (Chart 5.A). In sharp contrast, over the past six years (2006-2011) for the potential graduates mentioned in paragraph 40 – excluding Angola – IDA has averaged close to 60 percent of financing from the World Bank Group (Chart 5.B). In addition, IDA – as a share of these countries‟ total external public debt as at end-2010 – is on average 25 percent but varies from 45 percent in Armenia to 24 percent for India (Chart 6). Chart 5: IDA Commitments as a Share of World Bank Group Commitments Chart 5.A. To Prior IDA Graduates Chart 5.B. To Potential Graduates 100% 100% 80% 80% 60% 60% 40% 40% Countries graduating in the 80s 20% 20% Countries graduating in the 70s 0% 0% G-5 G-4 G-3 G-2 G-1 Graduation 2006 2007 2008 2009 2010 2011 18 Chart 6: IDA as a Share of External Public Debt Stocks 2010 Armenia Bosnia and Herzegovina Georgia Vietnam India 0% 20% 40% 60% 80% 100% IDA IBRD Other creditors 42. The potential graduates, excluding Angola, are also more dependent on World Bank financing than the earlier ones. For countries that graduated in the 1970s and 1980s, World Bank commitments, as a share of total public and publicly guaranteed debt commitments, averaged close to 18 percent and 13 percent, respectively, in the year of graduation and five preceding years (Chart 7.A). This compares to an average of 25 percent over the period 2005- 2010 for the potential graduates (Chart 7.B). It is important also to note that these countries‟ dependence on the World Bank has significantly increased over the last few years, resulting in the World Bank representing, on average, slightly more than one third of these countries total external public debt as at end-2010. Chart 7: World Bank Commitments as a Share of Total Public and Publicly Guaranteed Debt Commitments Chart 7.A. To Prior IDA Graduates Chart 7.B. To Potential Graduates 50% 50% 40% 40% 30% 30% Countries graduating in the 70s 20% 20% 10% Countries graduating in the 80s 10% 0% 0% G-5 G-4 G-3 G-2 G-1 Graduation 2005 2006 2007 2008 2009 2010 Note: Chart 7.A. excludes Korea due to lack of data. 19 43. India faces the added challenge of the IBRD Single Borrower Limit (SBL). The SBL mitigates the risk of concentration of the IBRD portfolio by capping IBRD exposure to IBRD borrowing countries that are large in terms of economic size and highly creditworthy with all other borrowing countries capped below this limit. Currently the SBL for India is set at US$17.5 billion.21 Under ordinary circumstances it is unlikely that IBRD financing to a country newly graduated from IDA would be constrained by the SBL. However, it may occur when a large IDA-eligible country was assessed as creditworthy for IBRD when its GNI per capita was low, in relation to the IDA eligibility threshold, and then the country attains a high degree of creditworthiness prior to graduation from IDA. Among past IDA graduates only China came close to falling into this category. At graduation from IDA in 1999 its accumulated outstanding obligations to IBRD were equivalent to almost 80 percent of the SBL limit (then set at US$13.5 billion). 44. India has had uninterrupted access to IBRD since its first IBRD loan in 1948. Outstanding obligations to IBRD at end FY12 amounted to US$11.6 billion. India is also the largest and longest IDA recipient: its outstanding obligations to IDA totaled US$26.2 billion at end FY12. In the past five fiscal years India has received on average US$1.8 billion per year in IDA commitments. The latter is high in volume but very low in per capita terms when compared to other IDA recipients due to the “capping� of IDA allocations to India. OECD DAC statistics also show that a very large share of aid to India over the past decade has come from IDA. The high outstanding IDA credit balance means that implementation of the IDA accelerated credit repayment clause would also have a significant impact. Principal repayments to IDA currently account for around 15 percent of principal payments on public and publicly guaranteed debt. If the accelerated credit repayment clause were to take effect as of IDA17, it would raise principal repayments to IDA by, on average, US$0.5 billion per annum in FY15-FY20 from an estimated average of US$1.1 billion per annum under the current repayment schedule to US$1.6 billion per annum after the application of the accelerated repayments clause. Impact of Possible Graduations on IDA Resources 45. As countries graduate from IDA, more resources could become available for other IDA countries. The discussion of the impact of graduations on IDA resources will need to take account of the following three elements:  First, the amount of resources freed from graduations. In the case of the six countries that could potentially graduate, this amounts to US$9.8 billion in IDA16, of which India and Vietnam represent more than 90 percent.  Second, IDA countries‟ needs and the size of the replenishment. Past experience, including with large IDA graduates, has shown that the resource needs of the remaining IDA countries exceeded the potential savings generated by the graduations. Going forward, IDA Deputies will need to determine the volume needed for achieving the objectives of the next replenishment. This may include a reduced or expanded agenda of 21 In FY11, the SBL for India was increased from US$16.5 billion to US$17.5 billion. Furthermore, in April 2012, the Executive Board agreed with the issuance of Special Private Placement Bonds that would allow for an increase in IBRD lending to India by US$4.3 billion through FY16. 20 actions. For example, in recent replenishments there has been an expanded support for fragile and conflict-affected countries and regional programs. In addition, IDA has been called upon to scale up support for the MDGs, and as such additional financing was required.  Third, the discussion will need to take account of the broader strategy and objectives of the World Bank Group, including the capacity of other Group members (i.e. IBRD, IFC and MIGA) to provide continuous support to graduating countries as IDA support is phased out. VI. IDA Graduation: Forward-Looking Issues 46. The ultimate objective of IDA’s graduation is for countries to make a successful and lasting exit from dependence on concessional resources to fund their development finance needs. Reversals back to IDA are costly for both recipient countries and donors alike. The fact that one third of past IDA graduates reversed back to IDA and that of these only three have since re-graduated to IBRD-only status highlights the volatility and fragility of the development progress in these countries. From IDA‟s perspective, this raises the broader question of whether the GNI per capita criterion could be complemented by alternative measures of poverty to more fully capture a country‟s poverty profile. In addition, there is a question on how best to adapt IDA‟s financing to ensure a smooth transition towards graduation. These issues are discussed in the following section. Alternative Measures of Relative Poverty 47. The concept of relative poverty, a key criterion of eligibility for access to IDA resources, has traditionally been measured by GNI per capita. This has proven to be a robust and useful indicator of a country‟s overall level of development and a proxy for individuals‟ potential access to income to enhance their wellbeing. But the GNI per capita only partially captures a country‟s level of development as it does not indicate how equitabl y income is shared within a country, the extent of existing poverty or where a country stands with regard to non-income indicators of wellbeing and international development goals such as those included in the Millennium Development Goals (MDGs). This raises the question of whether the use of per capita income is sufficient as the sole indicator of relative poverty, and whether other indicators could complement it or better capture relative poverty. However, as discussed below, alternative approaches involve significant implementation challenges with limited or adverse impact. 48. One approach to better identify relative poverty would be measuring GNI in purchasing power parity (PPP) terms rather than by the Atlas methodology as this would improve its correlation with non-income indicators of wellbeing. However, the PPP approach has limitations, in particular related to data availability and methodological issues that remain to be resolved. In contrast, long-term data under the Atlas methodology is available (see 21 Annex A for a description of the Atlas methodology).22 In addition, country rankings by GNI per capita and estimate of country‟s real growth rates are similar using either of the methodologies. Table 4: 2011 Human Development Groupings for IDA-eligible Countries High Human Development Blend IBRD/IDA countries Grenada 0.748 Dominica 0.724 Armenia 0.716 Bosnia and Herzegovina 0.733 Saint Lucia 0.723 Georgia 0.733 Saint Vincent and the Grenadines 0.717 IDA-only countries Tonga 0.704 Medium Human Development Blend IBRD/IDA countries Sri Lanka 0.691 Uzbekistan 0.641 India 0.547 Bolivia (Plurinational State of) 0.663 Viet Nam 0.593 Mongolia 0.653 Cape Verde 0.568 IDA-only countries Samoa 0.688 Honduras 0.625 Ghana 0.541 Maldives 0.661 Kiribati 0.624 Congo 0.533 Moldova (Republic of) 0.649 Vanuatu 0.617 Lao People's Democratic Republic 0.524 Micronesia (Federated States of) 0.636 Kyrgyzstan 0.615 Cambodia 0.523 Guyana 0.633 Tajikistan 0.607 Bhutan 0.522 Low Human Development Blend IBRD/IDA countries Pakistan 0.504 Zimbabwe 0.376 Papua New Guinea 0.466 IDA-only countries Solomon Islands 0.510 Haiti 0.454 Ethiopia 0.363 Kenya 0.509 Lesotho 0.450 Mali 0.359 Sao Tome and Principe 0.509 Uganda 0.446 Guinea-Bissau 0.353 Bangladesh 0.500 Togo 0.435 Eritrea 0.349 Timor-Leste 0.495 Comoros 0.433 Guinea 0.344 Angola 0.486 Zambia 0.430 Central African Republic 0.343 Myanmar 0.483 Djibouti 0.430 Sierra Leone 0.336 Cameroon 0.482 Rwanda 0.429 Burkina Faso 0.331 Madagascar 0.480 Benin 0.427 Liberia 0.329 Tanzania (United Republic of) 0.466 Gambia 0.420 Chad 0.328 Yemen 0.462 Sudan 0.408 Mozambique 0.322 Senegal 0.459 Côte d'Ivoire 0.400 Burundi 0.316 Nigeria 0.459 Malawi 0.400 Niger 0.295 Nepal 0.458 Afghanistan 0.398 Congo (Democratic Republic of) 0.286 Notes: 1/ Angola has been confirmed as creditworthy and is in the process of being formally reclassified from IDA-only to blend status. 2/ Nigeria may shortly be confirmed creditworthy. Nigeria reverse graduated to IDA in FY89 as a blend country and was reclassified as an IDA-only country in FY2005. 49. An alternative approach would be to design an appropriate and comparable social index to be used in conjunction with the per capita income criterion in order to identify the most vulnerable countries socially, and to adapt IDA’s response accordingly.23 This would also be consistent with IDA‟s efforts to better impact and monitor social indicators through Poverty Reduction Strategy Papers (PRSP), and the Bank‟s overall mission of poverty alleviation 22 See for example The World Bank Group, Development Economics (September 2010), Op. Cit. 23 This issue has been explored in IDA: “IDA Eligibility, Terms and Graduation Policies,� January 2001. 22 and helping countries reach the MDGs. However, as the 2001 review of IDA graduation policies identified, designing an appropriate social index would be a complex undertaking and a number of practical considerations in constructing such an index would need to be addressed. These would encompass what dimensions of poverty and social indicators should be included, how statistical and data problems would be addressed, and how aggregation problems would be solved. The link between the social index and the PRSP would need to be determined, in order for PRSPs to identify specific monitoring indicators, consistent with the MDGs. 50. If a suitable social index were to be developed, it could complement the GNI per capita estimates. However, this approach could alter significantly the set of IDA-eligible countries as there are countries currently not IDA-eligible with very low social index rankings and high per capita income levels which could have access to IDA under the revised criteria. This would in effect divert some of IDA‟s resources away from the poorest countries. The implications of this approach can be illustrated if the UN Human Development Index (HDI) was used (see Table 4). At the Very High Development level (country ranking 1-47) there are no IDA-only or blend countries. At the High Development level (country ranking 48-94) there are one IDA-only and seven blend countries. At the Medium Development level (country ranking 95-141) there are 15 IDA-only countries, seven blend countries and 22 IBRD-only borrowers. At the Low level of Development (country ranking 142-187) there are 2 blend countries (excluding Zimbabwe). The remaining 42 countries at the Low level of Development are IDA- only. IDA Transitional Support 51. A number of IDA-eligible countries that are poised to graduate in the near to medium term have expressed the need for financing options that would smooth the transition from IDA-eligibility to IBRD-only status. Of particular concern, especially in the case of heavily populated countries such as India and Vietnam, is the availability of sufficient levels of financing, on appropriate terms, to address the needs of the poorest segments of society. This seems particularly relevant as these countries are among those with the largest number of poor (see Chart 8).24 Donors have recognized that the sharp fall in concessional financing, or cessation of such financing that occurs when a country graduates from IDA, risks slowing the development momentum, raising the likelihood of reverse graduation back to IDA and compromising progress towards the MDGs. They have suggested that options for providing some form of transitional support be explored. 24 Almost half of the absolute poor in IDA countries (and a third of the poor worldwide) are in India alone. 23 Chart 8: Number of Poor in IDA Countries as of 2008 (in million, below US$1.25 per day) number of poor (million) 0 100 200 300 400 500 India Nigeria Bangladesh Congo, Dem. Rep. India Pakistan 351 Tanzania 445 Other blend countries Kenya Other IDA Vietnam 166 countries Madagascar Uganda Mozambique Ethiopia Angola Nepal Malawi Note: Bars in black are the potential graduates during IDA17 and ID18. Bars in dark gray are current and possible blend countries. The remaining IDA countries are in light gray. 52. The challenges of designing and funding an appropriate financing mechanism are considerable and will need to be carefully examined in close consultation with donors and potential IDA graduates. As part of early discussions, some donors and potential graduates have expressed views on how such financing could be framed:  Access should be limited to countries that have graduated from IDA and be time- bound and limited to a pre-specified period after graduation (three to five years).  There is range of views on the possible uses of transitional financing, including targeting resources to support projects with systemic or transformational impact on the poorest groups, lagging sub-national areas or the social sectors, support for fiscal and external sustainability, projects with positive spillover impacts at the regional level and innovation and piloting of new approaches that would lead to leveraging and scale up.  The terms on which financing is extended should be at near-market terms to reflect the positive assessment of creditworthiness on which the decision to graduate the country from IDA was based and the broader financing options that would be available. However, the terms should be more concessional than IBRD lending 24 terms. This approach would also be consistent with the financing terms established in IDA16 for blend countries (see section II) and with IDA‟s Articles of Agreement.25 53. As part of the options being considered to broaden the financing possibilities of IDA, concessional donor loans could help finance IDA17. This would represent a clear evolution of IDA‟s business model and would provide all IDA donors an additional avenue to contribute to IDA and respond to IDA countries‟ needs. In this context, these resources could partly be used to finance transitional support to IDA graduates. 54. Clear rules for granting access to transitional support will need to be defined. The diversity of situations of IDA countries in general, and that of the possible upcoming graduates in particular, calls for a case-by-case approach. Elements that could inform these decisions include the country‟s: i) level of dependence on IDA financing; ii) debt sustainability and macro- economic prospects; and iii) level of access to alternative sources of funding and its terms. In addition, these decisions could also be informed by aspects linked to poverty and the GNI per- capita level. In that case, the provision of transitional support would require clear signals demonstrating the country‟s commitment to address poverty issues as part of a strong national plan. In addition, further discussion would be needed to determine whether the transitional financing would support broad country priorities or be more targeted. VII. Issues for Discussion 55. Staff would welcome the Deputies’ views on:  IDA‟s current graduation policy.  The provision of transitional support to graduating countries and how best to approach it in the upcoming IDA17 replenishment discussions. 25 Article I requires that IDA provide financing on terms that “bear less heavily on the balance of payments than those of conventional loans�. 25 Annex A: The Atlas method of calculating GNI and GNI per capita 1. Gross national income (GNI), and GNI per capita, in U.S. dollars applicable to IDA graduation criteria (and other operational purposes) is calculated using the World Bank Atlas methodology. The purpose of this methodology is to incorporate a conversion factor that reduces the impact of exchange rate fluctuations in the cross-country comparison of national incomes and to smooth year-on-year fluctuations. 2. The Atlas conversion factor, for any year, is derived by taking the average of a country‟s exchange rate for that year and its exchange rates for the two preceding years, adjusted for the difference between the rate of inflation in the country, and the international rate of inflation. Up to and including 2000, the international rate of inflation was based on five countries, France, Germany, Japan, the United Kingdom, and the United States. For 2001 onwards, it is based on the Euro Zone, Japan, the United Kingdom, and the United States. A country‟s inflation rate is measured by the change in its GDP deflator. 3. The international rate of inflation rate is measured by the change in the Special Drawing Rights (SDR) deflator. This is calculated as a weighted average of the selected countries‟ GDP deflators in SDR terms: the weights being the amount of each country‟s currency in one SDR unit. These weights may vary over time because both the composition of the SDR and the relative exchange rates for each currency change. The deflator is calculated first in SDR and then converted to U.S. dollars using the SDR to dollar Atlas conversion factor. The Atlas conversion factor is then applied to a country‟s GNI. The resulting GNI in U.S. dollars is divided by the midyear population to derive GNI per capita. If official exchange rates are deemed to be unreliable or unrepresentative of the effective exchange rate during a period, an alternative estimate of the exchange rate is used. 4. The following formula shows the calculation of the Atlas conversion factor for year t: And the calculation of GNI per capita in U.S. dollars for year t: where et* is the Atlas conversion factor (national currency to the U.S. dollar) for year t, et is the average annual exchange rate (national currency to the U.S. dollar) for year t, pt is the GDP deflator for year t, pt S$ is the SDR deflator in U.S. dollar terms for year t, Yt $ is the Atlas GNI per capita in U.S. dollars in year t, Yt is current GNI (local currency) for year t, and Nt is the midyear population for year t. 26 Annex B: Small States – Exceptional Access to IDA Resources 1. In 1985, six small island countries (Dominica, Grenada, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, and Tonga) obtained access to IDA resources even though their per capita incomes were above the IDA-eligibility criterion.26 Since that time, the number of countries eligible for IDA under this “small island exception� has risen to thirteen. Eight other small island countries saw their per capita income increase above the operational cutoff (Cape Verde, Kiribati, Maldives, Marshall Islands, Micronesia, Samoa, Vanuatu and Tuvalu) while one of those originally eligible graduated from IDA (St. Kitts and Nevis). 2. The decision to introduce the small island exception was based on several considerations. First, small island economies share many of the problems of low income developing countries including export concentration, small domestic markets, high infrastructure costs, limited skill base, and weak institutions. Second, research has determined that these economies suffered from certain circumstances particular to their size and geography such as isolation, lack of natural resources, and unavailability of commercial credit. Third, many of these economies are highly vulnerable to natural disasters. The rationale for differential treatment of the small island countries remains valid. 3. An important advantage of eligibility for IDA under the small island exception is not only continued access to IDA but the concessionality of the lending terms. As agreed during IDA16, all eligible countries, including those classified as blend, receive IDA credits on the same terms as those extended to IDA-only recipients, regardless of GNI per capita level. Decisions on whether to grant a country continued eligibility to IDA under the small island exception is evaluated on a case by case basis but since the inception of the policy all small island economies whether classified as IDA-only or blend have been accorded access to the exception when their GNI per capita surpassed the IDA eligibility threshold. St. Kitts and Nevis lost its eligibility to the exception in 1994 when it graduated from IDA on the basis of its access to IBRD financing and its GNI per capita: it then stood at US$5,010, over six times the prevailing IDA operational cutoff. Two other small islands, São Tome and Principe and Timor-Leste, both currently classified as IDA-only, have GNI per capita above the operational cutoff in FY12 and FY13. 4. Countries that benefit from the small island exception account for only a very small share of IDA lending, and consequently the impact on overall availability of IDA resources to the poorest countries is not an important consideration. However, it does give rise to issues of equitable treatment across IDA-eligible countries, and in particular other small states, defined by the World Bank as those with a population of 1.5 million or less. While it is recognized that small islands are particularly vulnerable to natural disasters, the challenges associated with small populations and economies (diseconomies of scale, capacity and institutional constraints, open economies unable to easily diversify, remoteness and high transport costs) are common to other non-island small states. 5. Of the thirteen countries currently eligible for IDA under the small island exception, eight are classified as IDA-only. The other five countries have blend status and do borrow limited amounts from IBRD. The average per capita incomes of the four blend countries in the 26 See “Terms of Lending to Small Island Economies Graduating from IDA,� IDA/R85 -134, November 18, 1985. 27 Caribbean that benefit from the exception now exceed almost six times the operational cutoff. One of them, Grenada has a GNI per capita above the operational guideline to initiate graduation from IBRD (US$7,035 for FY2013). Two others, Dominica and St. Lucia have a GNI per capita close to the IBRD graduation cutoff. The extent to which the GNI per capita of countries eligible for the exception classified as IDA-only exceeds the operational cutoff is more moderate, but not insignificant. It ranges from well over five times for the Maldives, the richest country in this group, to a little less than two times in the case of Kiribati, the poorest (see Annex B. Table 1). Annex B. Table 1. GNI per-capita levels of countries under the Small Islands Exception (US$) Blend IBRD/IDA Countries Cape Verde 3,540 Dominica 7,090 Grenada 7,220 St. Lucia 6,680 St. Vincent and the Grenadines 6,100 IDA-only Countries Kiribati 2,110 Maldives 6,530 Marshall Islands 3,910 Micronesia, FS 2,900 Samoa 3,190 Tonga 3,580 Tuvalu 5,010 Vanuatu 2,870 Memorandum items: IDA operation cutoff 1,195 IDA historical cutoff 1,915 IBRD graduation cutoff 7,035 6. In 1985 when the exception was approved, the only small states that qualified were all islands. Currently, however, there are four small states classified as IDA only with per capita incomes above the eligibility threshold that are not islands: Bhutan, Djibouti, Guyana and Lesotho. Due to lack of creditworthiness these small states retain access to IDA resources but are subject to terms harder than countries under the small islands exception. On average, their GNI per capita (US$1,868) is much lower than that of small island states eligible for the small island exception (US$4,672). 28 Annex C: Graduation Policies from Concessional Financing – Select International Agencies 1. This annex presents a review of the graduation policies of the concessional financing facilities of the IMF and three major regional development banks (RDBs – AfDF, ADB and IDB). It also includes information on the eligibility and threshold limits set by the Global Alliance for Vaccines and Immunization (GAVI) for the grants it extends to developing countries. A comparison matrix of the criteria used by IDA and each of these institutions is given in Annex C. Table 1. 2. The review confirms that the graduation policies of these agencies are, for the most part, closely aligned with those of IDA. Both the IMF and the RDBs employ an income criterion and a creditworthiness criterion and the IMF also takes into account short term vulnerabilities. Creditworthiness is not a concern for GAVI since it does not make loans and it relies solely on an income criterion. As regards the IMF, however, revisions to the graduation policies of the Poverty Reduction and Growth Trust (PRGT) in 2010 differentiate between the income and market access (creditworthiness) criterion. Under the new policies a country is no longer eligible for the resources if it: (a) has a persistently high level of income; or (b) durable and substantial access to financial markets; and does not face serious short-term vulnerabilities regardless of its per capita income level. In common with IDA, each institution regularly reviews and updates its income criterion to reflect changes in national income levels in recipient countries. Without exception, the GNI per capita data used are based on the calculations made by the World Bank according to the Atlas methodology. A. International Monetary Fund 3. The IMF provides concessional lending to member countries through the PRGT and historically PRGT eligibility has been closely aligned with that of IDA. In 2010, as part of the revamp of its lending instruments, the IMF revised the eligibility and graduation criteria of the PRGT and introduced some important differences between the conditions of entry and exit. A key objective of the change was to minimize the risk of untimely graduation decisions and subsequent need for reversal and re-access to the PRGT. 4. Under the new policies, eligibility for PRGT remains almost identical to that of IDA; countries would become eligible if: (a) their per-capita income is below the IDA cutoff; and (b) they do not have durable and substantial access to financial markets, i.e., they are not creditworthy. As regards to graduation, however, a country is no longer eligible for PRGT resources if it: (a) has a persistently high level of income; or (b) durable and substantial access to financial markets; and does not face serious short-term vulnerabilities. The assessment of vulnerabilities focuses, in particular, on risks of a sharp decline in income, loss of market access, and debt vulnerabilities. Specifically:  The income criterion requires that a country has: i) per capita GNI above the IDA operational cutoff in the last five years; ii) per capita GNI in the last year at least twice the IDA operational cutoff, or at least three times the cutoff for small countries; and iii) per capita GNI not declining in the last five years. 29  The market access criterion requires that a country has: i) per capita GNI above 80 percent of the IDA operational cutoff; and ii) durable and substantial access to foreign bond markets and external commercial loans of at least 100 percent of quota in the last five years, or case-specific assessments of the country‟s market access to be sufficient based on the volume and terms of recent actual borrowing in international capital markets and the sovereign credit rating.  The short-term vulnerability criterion requires that short-term vulnerabilities must be limited with regard to income decline, the loss of market access, and debt distress as indicated by the most recent DSA. In this context, where there is a LIC-DSA, risks of debt distress should be moderate or less. 5. Following the 2010 review, Angola and Azerbaijan graduated based on the income criterion given that their short-term vulnerabilities appear manageable and India, Sri Lanka, and Pakistan graduated under the market access creditworthiness criterion and their short-term vulnerabilities appear manageable. The 2012 review found that eight countries met the income or the market access criterion for graduation from PRGT eligibility, but none was proposed for graduation at this time given their serious short-term vulnerabilities, which are exacerbated by the current elevated risks in the global economy (Armenia, Dominica, Georgia, Ghana, Grenada, Maldives, St. Vincent and the Grenadines, and Vietnam). The 2012 review also proposed to increase the population threshold used to define small states from 1 to 1.5 million to align with the definition adopted by the World Bank. This change would not result in any additional countries becoming PRGT-eligible now, but it could do so in the future, and it would allow the IMF to expand PRGT eligibility to vulnerable small countries. 6. For the next review in early 2014 the IMF intends to conduct a more comprehensive review of the PRGT-eligibility framework, including the entry and graduation criteria. B. African Development Fund 7. The relationship between the AfDB and its concessional lending window, the AfDF, is almost identical to that between IBRD and IDA. The eligibility and graduation criteria that govern the African Development Fund are aligned to those of IDA and reviewed on an annual basis. Graduation from access to AfDF resources is normally triggered when a country‟s GNI per capita exceeds the IDA operational cutoff for more than two consecutive years and has been assessed as creditworthy for market based financing of the type provided by the AfDB or IBRD. As with IDA there is nothing to preclude a country that is assessed as creditworthy and able to meet its financing needs from market sources to graduate but in practice this has not been the case. The process of graduation is managed in much the same way as the World Bank. Countries pass through a transition phase during which concessional AfDF financing is phased out and market based AfDB financing is phased in. The length of the transition period is determined on a case-by-case basis but typically lasts from two to five years. The list of countries eligible for AfDF is identical to that of IDA with the exception of Angola which is being graduated from AfDF. The classification of countries between AfDF-only and AfDF- AfDB blend follows that of IBRD and IDA with the exception of Nigeria which the World Bank 30 reclassified from blend to IDA-only in 2006 and which the AfDB retained as a blend country largely to avoid crowding out of limited ADF resources to smaller African countries. C. Asian Development Bank 8. The criteria that govern graduation from the ADF, the concessional window of the ADB, mirror those of IDA and include an income and creditworthiness component. The income criterion is comparable to the one used for graduation from access to IDA resources and the process of determining creditworthiness for access to ADB‟s Ordinary Capital Resources (OCR), market-based financing equivalent to IBRD, follows much the same process as that of the World Bank. A three way classification system is also employed: “adequate creditworthiness�, equivalent to an IBRD borrower; “limited creditworthiness�, equivalent to a blend country; and “lack of creditworthiness�, which equates to an IDA-only borrower. As with IDA, the policies governing ADF eligibility are reviewed periodically during replenishment rounds. 9. Graduation from access to ADF financing is normally triggered when a country exceeds the IDA operational cutoff (income criterion) and is determined to have “ adequate creditworthiness� to borrow from the OCR of the ADB or access comparable market-based financing. Four IDA-only countries have access to OCR resources (Bangladesh, Marshall Islands, the Federated States of Micronesia, and Timor-Leste). The graduation process from ADF is conducted in close consultation with the borrower, and normally extends over a three to four year period during which ADF financing is phased out and replaced with borrowing from the OCR. As is the case for IDA, there is no impediment to a borrowing country with adequate creditworthiness graduating from ADF while its per capita income is below the income criterion threshold. India has never exercised its eligibility for concessional ADF resources and has borrowed only from ADB OCR. D. Inter-American Development Bank 10. The IDB delivers its concessional financing through blending highly concessional resources from the Fund for Special Operations (FSO) and non-concessional resources from the Ordinary Capital (OC) account. FSO resources are directed at the poorest and least developed countries in the region (D2 countries) that are also IDA eligible countries: Bolivia, Guyana, Haiti, Honduras, and Nicaragua. In the case of Haiti all financial support is provided in the form of grants. Two IBRD borrowers, Guatemala and Paraguay currently have access to a small allocation of FSO resources, in addition to borrowing from the OC account. The IDB does not have a formal graduation policy for D2 countries, but it uses a GDP per capita criterion (US$2,430 in constant 2005 prices) to set a threshold for access to FSO resources. Based on this criterion Ecuador, El Salvador and Suriname graduated from access to FSO resources in 2009. E. Global Alliance for Vaccines and Immunization 11. The focus of the GAVI is on the world's poorest countries, and eligibility for support from the GAVI Alliance is determined solely on the basis of national income. From 2000-2005, all countries with a GNI per capita equal to or below US$ 1,000 (based on World Bank 1998 calculations Atlas methodology) qualified for support. Seventy-five countries qualified for the 31 GAVI Alliance, including Timor-Leste which was added to the list of eligible countries in 2002 when it became an independent state. The threshold was reset for 2006-2010 on the basis of the World Bank 2003 GNI per capita calculations Atlas methodology and as a result four countries, Albania, Bosnia and Herzegovina, China and Turkmenistan, ceased to be eligible for GAVI Alliance resources and one country, Kiribati, was added to the list of eligible countries. In total, 72 countries were eligible for GAVI Alliance support during this period. 12. In 2009, the GAVI Alliance Board approved changes to the eligibility criterion. Under the new policy, eligibility will continue to be based solely on national income but the threshold will be automatically adjusted for inflation at the start of each calendar year on the basis of the World Bank‟s GNI per capita data Atlas methodology. The new policy went into effect on January 1, 2011 with eligibility threshold for 2011 set at a GNI per capita equal to or below US$1,500 based on the World Bank 2009 data published in July 2010. A total of 56 countries are currently eligible for GAVI Alliance support. The Board also endorsed a revision of the budget cap for India, subject to fund availability, for the period 2012-2015. In addition, the Board endorsed a proposal that GAVI Alliance explore the options for providing some financial support, for a limited period, when countries surpass the threshold and cease to be eligible for GAVI Alliance support in order to ease the transition process. 32 Annex C. Table 1: Summary of the Graduation Policy in Concessional Financing in IFIs III. Other Graduation Frequency of Recent IFIs I. Income criterion II. Creditworthiness criterion Exceptions considerations decision rule review graduates Meet I and II in GNI per capita exceeds the Adequacy of general, but Azerbaijan IDA operational cutoff for Creditworthiness to sustain sufficient Small Review every IDA market-based possible to (2011); Albania more than two consecutive IBRD financing islands three years financing graduate based (2009) years on II The capacity to access international capital markets on a durable and substantial basis The absence of Small (i) GNI per capita exceeds is based on two tests (i) public sector short-term countries the IDA operational cutoff issuance or guaranteeing of external bonds vulnerability to (the income Albania, Angola, for at least the last five or external PPG loans during at least three Meet I or Meet IMF- income decline, ceiling is Review every Azerbaijan; India, years; (ii) non-declining of the last five years and in a cumulative II and III PRGT loss of market raised to two years Pakistan, and Sri income in five years; (iii) amount ≥100% of its IMF quota; or (ii) combined access, and debt three times Lanka (2010) latest GNI per capita is case-specific assessments of the market distress of the above twice the IDA cutoff access adequacy; and having GNI per cutoff) capita ≥80% IDA cutoff and non-declining in the last five years Graduating countries are Meet I and II in GNI per capita exceeds the eligible for a general, but Angola (2011); IDA operational cutoff for Creditworthiness to sustain AfDB/IBRD gradually declining Annual review AfDF possible to No Egypt, Equatorial more than two consecutive financing amount of AfDF in practice graduate based Guinea (1999) years resources on blend on II terms for two to five years. Meet I and II in Review during China and India general, but No replenishment GNI per capita exceeds the “Adequate creditworthiness� to OCR or Assessment of the (de facto since ADF possible to available discussions IDA operational cutoff market-based financing country‟s readiness 1999); and Cook graduate based info every three Islands (2008) on II years GDP per capita exceeds a Allow more threshold (US$2,430 in Unspecified Ecuador, El flexible blend of No formal IDB-FSO constant 2005 dollars); but No specific rule No periodic Salvador, and FSO with non- policy no specific rule for FSO- review Suriname (2009) concessional OC only countries Albania, China, GNI per capita exceeds a Three-year Annual review Bosnia and Herz., GAVI threshold (US$1,500 in Not relevant Income only No transition from 2011 Turkmenistan 2010) (2006) 33 References African Development Bank, 2011, Transition Framework for Graduating and Reversing Countries, Revised. Asian Development Bank, 2008, Review of the 1998 Graduation Policy of the Asian Development Bank. Asian Development Bank, 1999, ADF VII: Progress Report- Report through the Board of Directors to the Board of Governors. International Development Association, 2001, IDA Eligibility, Terms and Graduation Policies. International Monetary Fund, 2012, Eligibility to Use the Fund’s Facilities for Concessional Financing. International Monetary Fund, 2010, Eligibility to Use the Fund’s Facilities for Concessional Financing. Inter-American Development Bank, 2011, Fund for Special Operations: Review of the Implementation of the Debt Sustainability Framework and Enhanced Performance-Based Allocation. Global Alliance for Vaccines and Immunization (GAVI), 2009, GAVI eligibility graduation procedures. GAVI website http://www.gavialliance.org/support/apply/countries-eligible-for-support/, accessed on July 20, 2011.