Abstract. Acknowledgements

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Analysis of Recipient Executed Trust Funds March 2010 CFP Working Paper Series No. 5 Abstract In recent years, bilateral aid channeled through multilateral organizations like the World Bank has become an important vehicle for concessional finance for development. At the end of FY09, the World Bank held US$23.9 billion in trust, an amount that has doubled since FY05. While much of the growth reflects the increase in large Financial Intermediary Funds (FIFs), there has also been a rapid increase in Recipient Executed Trust Funds (RETFs). This paper aims to inform internal Bank audiences and stimulate discussion on some selected issues primarily the growing importance of RETFs as a source of finance and relevance to Bank s country level operations. The report also examines briefly the link between RETF allocations and country performance. The report finds that disbursements of RETFs grew by twenty percent over FY03-FY09. Over this period the trends in regional disbursement shares shifted away from East Asia and Pacific and South Asia Regions to Africa Region. Recipient country concentration is evident: fifty percent of the disbursements have been from six countries Afghanistan, West Bank and Gaza, Ethiopia, Vietnam, Indonesia and China. RETF disbursements, as a share of the combined IBRD, IDA, and RETF, grew from 5 percent in FY03 to 12 percent in FY08, while that of IBRD and IDA fluctuated. The bulk of the disbursements from RETF grants were from IDA eligible countries. RETFs appear to have substituted for IDA in fragile and conflict affected states. Disbursements from RETF grants and IDA credits were broadly aligned by sector. There may be some correlation between RETF grants allocation and country performance but the results are not conclusive. The report identifies some areas for further examination: whether RETFs should complement or substitute for Bank operations in specific cases; and whether RETF allocations should be made on the basis of some need or performance based country priorities, outcomes and results. Acknowledgements This report was prepared by Wahida Huq (CFPPM). Soonhwa Yi (CFPPR) contributed to the analysis of the RETFs and country performance. Zafar Ahmed (Consultant) assisted with the data analysis. Data compilation was done by Jyoti Dhingra (ACTTN), Buena Cabanela, Zhimei Xu, Mahmut Aydogmus (CFPTO) and Augustina Nikolova (HDN). Peer reviewers were: Robert Utz (CFPIR) and Sajjad Shah (EAPCO). Overall guidance was provided by Roberto Tarallo (CFPPM), Rocio Castro (CFPVP) and David Potten (CFPTO). Mohamed Diaw and Milagros Reyes assisted with the formatting of the report. Patricia Rogers edited the report. THE WORKING PAPER SERIES The CFP Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development finance, aid architecture, and aid effectiveness. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s). They do not necessarily represent the views of the Word Bank Group, its Executive Directors, or the countries that they represent and should not be attributed to them. For more information, please contact Angela Furtado, telephone , afurtado@worldbank.org or visit where copies are available in pdf format. The author may be contacted at whuq@worldbank.org.

2

3 CFP Working Paper Series No. 5 Analysis of Recipient Executed Trust Funds Wahida Huq March 2010 The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s). They do not necessarily represent the views of the World Bank Group, its Executive Directors, or the countries that they represent and should not be attributed to them.

4 Abbreviations and Acronyms AFR ARTF BETF CFPVP CAS CPIA CPS EAP ECA FIF GEF GPP IBRD IDA IFC LCR MDG MDTF MIGA MNA PBA QAG RETF SAR TF WB&G Africa Region Afghanistan Reconstruction Trust Fund Bank-executed trust fund Concessional Finance and Global Partnerships Vice Presidency Country Assistance Strategy Country Policy and Institutional Assessment Country Partnership Strategy East Asia and the Pacific Region Europe and Central Asia Region Financial intermediary fund Global Environment Facility Global Programs and Partnerships International Bank for Reconstruction and Development International Development Association International Finance Corporation Latin America and the Caribbean Region Millennium Development Goal Multidonor trust fund Multilateral Investment Guarantee Agency Middle East and North Africa Region Performance-Based Allocation Quality Assurance Group Recipient-executed trust fund South Asia Region Trust fund West Bank and Gaza

5 CONTENTS Executive Summary... i I. Introduction... 1 A. Objectives and Scope of the Study... 2 II. Trends in RETF Disbursements... 3 A. Regional Disbursement Patterns... 3 B. Disbursements to Fragile and Conflict-Affected States... 4 C. Major Country Recipients... 5 D. Disbursements of Recipient-Executed MDTFs... 6 E. Disbursements of Recipient-Executed Programmatic and Free-Standing TFs F. Disbursements of RETF Grants by Bank Operational Products... 7 G. Disbursements of RETF Grants by Usage... 8 III. RETFs Supporting GPPs... 9 IV. RETFs and Bank Lending A. RETF Disbursements by Country Groups B. Disbursements of RETFs, IDA Credits, and IBRD Loans C. Disbursement Patterns of RETFs and IDA in Fragile and Conflict-Affected States D. Disbursement Patterns of RETFs and IDA Credits and IBRD Loans by Sector V. RETFs and Country Performance VI. Conclusions Boxes Box 1.1: World Bank s Roles in the Management of Trust Funds... 1 Box 2.1: RETF Portfolio in Ethiopia... 3 Box 2.2: RETF Portfolio in Vietnam... 4 Figures Figure 2.1: RETF Grant Disbursements in Fragile and Conflict-Affected States, FY03-FY09 (in US$M and as percent of total RETF disbursement)... 5 Figure 2.2: Total RETF and MDTF Disbursements (in US$M) by Fiscal Year and as Percent of Total RETFs... 6 Figure 4.1: IBRD, IDA, and RETF Disbursements, FY03-FY

6 Tables Table 2.1: Six Recipients with Highest Cumulative RETF Disbursements (FY03-FY09)... 5 Table 2.2: RETF Grant Disbursements by Operational Product Category as Percentage of Total RETF Disbursements, FY03-FY Table 2.3: Share of RETF Disbursements by Usage (as percentage of total annual disbursements) Annexes Annex 1: Hierarchy and Typology of World Bank-Managed Trust Funds Annex 2: Fragile States Annex 3: Global Programs and Partnerships Annex 4: Are RETF Commitments Related to Country Performance? Annex 5: Statistical Data

7 Executive Summary 1. In recent years, as the international aid system has become more complex, bilateral aid channeled though multilateral organizations like the World Bank (Bank) has become an important vehicle of concessional finance for development. At the end of FY09, the Bank held US$23.79 billion in trust, 1 an amount that had doubled since FY05. While much of the growth reflects the increase in large Financial Intermediary Funds (FIFs), there has also been a rapid increase in Recipient-executed trust funds (RETFs) and Bank-executed trust funds (BETFs). Excluding FIFs, the stock of trust funds (TFs) held by the Bank amounted to US$7.7 billion at the end of FY09. As TF stocks have increased, so have TF disbursements at an average annual growth rate of 19 percent between FY03 and FY This report focuses primarily on RETFs, which are becoming increasingly important as a source of finance and relevant to Bank operations. Specifically, the report examines the overall composition of and trends in RETF disbursements during the period FY03-FY09 and the link between RETF commitments and country performance in a sample of 74 IDA countries. Trends in RETF Disbursements 3. RETFs disbursements increased from US$981 million in FY03 to US$2,811 million in FY09, for an annualized disbursement rate of 19 percent. This growth rate is comparable to 17 percent for FIFs during this period. 4. Disbursement patterns in the regions were mixed except Africa Region (AFR) where there have been significant shifts in the RETF disbursement shares (as percent of total). AFR accounted for the third-highest share of disbursement (15-19 percent) until FY06, and then from FY07 to FY09 the Region disbursed about one-third of total RETF disbursements an increase that was largely due to several large-disbursing programs in Ethiopia. South Asia Region s (SAR) share had remained fairly stable at around 26 percent until FY08 and then rose to 30 percent in FY09; the bulk of SAR s share has been from the Afghanistan Reconstruction Trust Fund. The third-largest disbursement share was from the East Asia and Pacific Region (EAP) region whose share ranged between 15 and 23 percent during FY03-FY08, and then in FY09 fell to 13 percent, mainly because of the closure or winding down of large TFs, such as those for Timor-Leste and Aceh (Indonesia). By FY09, because of the closure of several post-conflict trust funds, the share of the Europe and Central Asia Region (ECA) had declined by more than half from 12 percent in FY03. The disbursement share for Middle East and North Africa (MNA) Region fluctuated between FY03 and FY08 but increased to 16 percent in FY09 largely due to the trust funds for Iraq. Use of RETFs has been the least in Latin America and the Caribbean Region (LCR) on average the share was only 6 percent. 1 Includes cash and promissory notes. 2 Excluding FIFs. i

8 5. About half of RETF disbursements during FY03-FY09 were concentrated in six countries: Afghanistan (22 percent), West Bank and Gaza (9 percent), Ethiopia (7 percent), Vietnam (6 percent), Indonesia (5 percent), and China (4 percent). Nearly 42 percent of the total RETF disbursements during FY03-FY09 were in fragile and conflict-affected states, with Afghanistan accounting for 51 percent (US$2.7 billion) and West Bank and Gaza for 21 percent (US$1.1 billion). Most of the RETFs in this country category are funded by multiple donors. 6. In a move to harmonize development priorities and policies, donors are increasingly pooling their resources into multi donor trust funds (MDTFs) to finance post-conflict, postdisaster reconstruction activities and provision of global/regional public goods. Disbursements of recipient-executed MDTFs grew rapidly, from US$457 million in FY03 to US$2.2 billion in FY09. The share of MDTFs to total RETF disbursements almost doubled, from 47 percent in FY03 to 80 percent in FY09. MDTFs in SAR disbursed, on average, about one-third of total recipient-executed MDTFs, and those in AFR 21 percent; however, if MDTFs for Afghanistan are excluded, SAR s share falls to 3 percent and AFR s rises to 26 percent. 7. Most of the MDTFs are Programmatic TFs which support a number of activities through multiple grants over several years. More than two-thirds of the total RETF disbursements from FY03-FY09 were from Programmatic TFs and the remainder from non-programmatic, or freestanding TFs. The highest Programmatic TFs are the Afghanistan Reconstruction Trust Fund, Global Environment Facility and West Bank and Gaza. RETFs Supporting Global Programs and Partnerships 8. Some TFs mainly financed by multiple donors support Global Partnerships and Programs (GPPs). RETFs supporting GPPs represented 32 percent of total disbursements in FY03 and grew to 59 percent in FY07. Disbursements from these TF grants were significant among programs disbursing more than US$50 million annually. There were noticeable shifts in the Regional distribution of disbursements from RETFs supporting GPPs over the FY03-FY07 period. In FY03, EAP accounted for the highest disbursement share (28 percent), ECA for 25 percent, and AFR for 19 percent. In FY07, by contrast, AFR accounted for 32 percent, EAP for 23 percent, and ECA for 9 percent; and SAR s share had increased significantly from the FY03 level. RETFs supporting GPPs are not significant in fragile states. RETF and IBRD/IDA Disbursements 9. During the period FY03-FY09, the annualized disbursement rate of RTEF, IBRD and IDA were 19 percent, 8 percent and 4 percent respectively. RETF disbursements, as a share of combined IBRD, IDA, and RETF, grew steadily from 5 percent in FY03 to 12 percent in FY08 while that of IBRD and IDA fluctuated. IBRD disbursement share declined from 59 percent in FY03 to 47 percent in FY08 and rose to 61 percent in FY09 mainly due to the increased disbursement in the wake of the financial crisis. IDA disbursement share increased from 36 percent in FY03 to 44 percent in FY05 but fell to 36 percent in FY09. ii

9 10. About 57 percent of RETF disbursements were in IDA-only countries during FY03- FY09, and the share increases to 62 percent when blend countries are included. IBRD countries accounted for about 21 percent. The remaining 17 percent was from RETFs supporting global and regional programs. 11. Are RETF disbursements aligned with those of IDA and IBRD financing? A comparative analysis for FY03 and FY09 yields the following results: The regional distribution of IDA disbursements remained relatively constant in both FY03 and FY09, with AFR accounting for the highest share. In comparison, the highest share of RETFs shifted away from SAR in FY03 to AFR in FY09, thus becoming more aligned with IDA. In the case of IBRD countries, however, such alignment is not apparent. IBRD loans were concentrated in LCR, while RETF disbursements were mostly in EAP in both FY03 and FY09. RETFs have substituted for IDA in fragile and conflict-affected states. Of the total amount disbursed in these countries, IDA s FY03 share of 76 percent dropped to 58 percent in FY09. In contrast, RETF s share in IDA-eligible fragile and conflict-affected states almost doubled, from 24 percent in FY03 to 42 percent in FY09. IDA credits and RETFs were broadly aligned by sector. The Public Administration and Law, Health, and Education sectors accounted for the highest shares for both RETF and IDA credits in both FY03 and FY09. In IBRD countries, the sectoral shares of RETFs were not totally consistent with those for IBRD loans. The sectors with the highest IBRD disbursements in both years were Public Administration and Law, Transportation, and Finance. In contrast, he sectors with the highest RETF disbursements were Industry and Trade in FY03 and, in FY09, Energy and Mining, with Agriculture and Industry and Trade close behind. RETF Allocations and Country Performance 12. Is there any link between RETF allocations and country performance? 3 Overall, the results of simple regression analysis suggest that there may be some correlation between RETF grant commitments per capita and CPIA ratings but the relationship is not strong. However, the question remains whether some ex-ante allocation criteria for determining country shares would be possible. Would such a measure go against the spirit of flexibility and need-based prioritization that is the hallmark of TFs? How would the Bank achieve a balance between equity and efficiency and a country performance-based allocation? These issues are worthy of further examination. 3 Approved RETF grant amounts have been used as a proxy for commitments and CPIA ratings for country performance. iii

10 Conclusions 13. TFs have grown substantially to provide large volumes of financial resources that the Bank now manages to support country programs, and RETF activities are a significant part of the Bank s operational portfolio. This growing resource envelope and portfolio need to be managed well. The Bank needs to continue applying to RETFs the same operational, fiduciary, social, and other safeguards that apply to regular Bank operations. In particular, RETFs, which mimic regular Bank lending operations in critical ways, are a special category for consideration. Yet, at the same time, it is important to preserve the flexibility, nimbleness, and innovativeness that often categorize RETF activities. Thus the particular question is whether RETFs should complement or substitute for Bank operations in specific cases. iv

11 I. Introduction 1.1 Within the international aid system, bilateral aid channeled through multilateral organizations like the World Bank (Bank) occupies a particular niche. The Bank has comparative advantages for the effective delivery of aid. It has a presence in most of the lowand middle-income countries, including many of the conflict-affected and fragile states, 4 and it often plays a coordinating role in channeling donor assistance to countries in crisis. In addition, it carries out rigorous analytic work that provides a strong foundation for member countries formulation of policies and programs for poverty reduction and economic growth. Thus it is useful to analyze to what extent the bilateral and multilateral aid entrusted to the Bank for recipient execution is aligned with the Bank s own activities and performance orientation. 1.2 Trust funds (TFs) managed by the Bank have become an important vehicle for channeling concessional finance. In recent years, the stock of trust funds held by the Bank has doubled, from US$11.56 billion at the end of FY05 to US$23.79 billion at the end of FY09, 5 with a little over 1,000 trust funds active at the end of that time. While much of the growth reflects the increase in large Financial Intermediary Funds (FIFs), the stock of Recipient-executed Trust Funds (RETFs) and Bank-executed Trust Funds (BETFs) has also increased rapidly, from US$3.6 billion at the end of FY05 to US$7.7 billion at the end of FY09. 6 As TF stocks have increased, so have disbursements. Box 1.1: World Bank s Roles in the Management of Trust Funds World Bank-managed TFs are executed by the Bank, by a third party (e.g., country recipients), or under some special financial partnership arrangements. Funds the Bank passes to a third-party recipient for development activities are referred to as RETF grants. The Bank has a supervisory oversight role for these trust funds. Funds over which the Bank has spending authority are identified as BETF grants. Arrangements under which the Bank provides specific administrative or financial services with limited fiduciary or operational roles or complex financial schemes are referred to as FIFs. Examples: Global Fund to Fight AIDS, Tuberculosis and Malaria, and International Finance Facility for Immunization. Note: See Annex 1 for the hierarchy and typology of TFs managed by the World Bank. 1.3 Over the period FY03-FY09, disbursements of Bank-administered TFs (excluding FIFs) grew at an annual average rate of 19 percent, from US$1.24 billion to about US$3.52 billion. Including FIFs, the annualized disbursement rate is 18 percent. The average annual growth rates of FIFs (17 percent) and RETFs (19 percent) during this period are comparable 4 As weak institutions and high risks of conflict constrain national poverty reduction and service delivery and create potential negative spillovers for neighboring countries, strengthening assistance to fragile countries remains a priority for the World Bank as well as other donors. Fragile states represent a broad spectrum of situations and characteristics, and they include conflict-vulnerable or post-conflict countries with fragile institutions, countries with weak or dysfunctional institutions and policies, and strong regimes with a particularly poor governance record. The Bank defines fragile states as those countries that score lowest on the average Country Policy and Institutional Assessment (CPIA), as well as non-member territories and countries without CPIA data; Annex 2 lists these countries. 5 Includes cash and promissory notes Partnerships and Trust Fund Annual Report (forthcoming). 1 Page

12 (Annex 5, Table 1). During this period, RETFs accounted for 37 percent of total TF disbursements, FIFs 53 percent, and BETFs only 6 percent. 1.4 A growing share of Bank-administered TFs supports Global Programs and Partnerships (GPPs): total GPP disbursements more than quadrupled between FY03 and FY07, from US$831million to US$3,410 million. 7 Even as total Bank-administered TFs have grown, the proportion of TFs supporting GPPs has grown significantly, from 32 percent in FY03 to 59 percent in FY As an important source of concessional funds for many country recipients, RETFs play an increasingly important role in the overall global aid architecture. They support, among other things, Global Environment Facility (GEF) projects, debt servicing operations, and other stand-alone projects, or they may be directly associated with Bank-financed projects and operations. RETFs serve as a flexible financing instrument to support countries with varied characteristics and needs; and they offer countries an opportunity to enhance their institutional capacity by managing the funds independently. A. Objectives and Scope of the Study 1.6 In 2007, the World Bank s Concessional Finance and Global Partnerships Vice Presidency (CFPVP) initiated analytic work on aid architecture to position the Bank in the rapidly changing aid environment, particularly with respect to its role in concessional financing. The analysis of the operational links between RETFs and Bank-financed operations was identified as one of the key areas of investigation. TF grants of this type are either explicitly linked to Bankfinanced operations or designed as stand-alone funds, and are most similar to the International Bank of Reconstruction and Development (IBRD) and International Development Association (IDA) lending programs, which support country-based initiatives. RETFs typically finance investment and recurrent operating costs for service delivery, capacity building, and technical assistance. For these funds, the Bank normally plays an operational role appraising and supervising funded activities. 1.7 This paper aims to inform internal Bank audiences and stimulate discussion by building on previous analytic work to focus on some selected issues primarily the growing importance of RETFs as a source of finance and relevance to Bank s country level operations. Specifically, it examines the composition of RETF disbursements during the period of FY03-FY09, giving particular attention to the role of RETFs in countries that are eligible for IBRD and IDA funds. It also briefly examines the link between RETF allocations and country performance in a sample of 74 IDA countries. 1.8 The report is organized as follows. Section II analyzes trends and patterns in RETF disbursements by the Bank s operational regions, fragile and conflict-affected states, and country concentrations, and through Multidonor trust funds (MDTFs) and Programmatic and Freestanding TFs. Section III discusses the disbursement trends of TFs supporting GPPs. Section IV analyzes the distribution of RETF grant disbursements by country groups 8 and compares them with disbursements of IDA credits and IBRD loans for both regions and sectors. Section V examines the relationship between RETFs and country performance and Section VI presents 7 The analysis of trust funds supporting GPPs covers FY03-FY07, as disaggregated data for later years are not available. 8 Countries are grouped by their eligibility to receive credits from IDA and loans from IBRD. 2 Page

13 conclusions. Annexes 1-4 provide additional information; and, for convenience, tables providing statistical data are all presented in Annex 5. II. Trends in RETF Disbursements 2.1 RETF disbursements increased from US$981 million in FY03 to US$2,811 million in FY09, for an average annualized disbursement rate of 19 percent (Annex 5, Table 1). A. Regional Disbursement Patterns 2.2 RETF disbursements in the Africa Region (AFR) rose from US$147 million in FY03 to US$984 million in FY09 (Annex 5, Table 2), more than doubling AFR s share in all RETF disbursements. The highest cumulative disbursement was in Ethiopia, which accounted for more than one-fourth of the Region s total (see Box 2.1), while Africa regional programs accounted for 13 percent. Box 2.1: RETF Portfolio in Ethiopia During FY03-09, Ethiopia had 28 RETFs in its portfolio, with cumulative disbursements of about US$902 million 28 percent of total RETF disbursements in AFR. Disbursements increased from US$8.0 million in FY03 to US$31.0 million in FY05 and then jumped to US$311.0 million in FY06. About 97 percent of the disbursements were from FY07 to FY09: in FY09, RETF grants in Ethiopia disbursed US$296.0 million. About 97 percent of the RETF grant disbursements were from MDTFs, three of which accounted for 75 percent of total RETF disbursements and 77 percent of total MDTF disbursements in the Region. All the MDTF programs supported by such donors as Canada, Japan, and Netherlands focused primarily on safety nets and protection of basic services. The highest-disbursing program was the Protection of Basic Services Project subprograms, which accounted for 43 percent of total RETF disbursements in Ethiopia. 2.3 Disbursement patterns in the other Regions were more mixed. The share of the Europe and Central Asia Region (ECA) declined to only 5 percent in FY09, largely because of the closure or imminent closure of post-conflict TFs for Bosnia and Herzegovina and Yugoslavia. A decline in the share of the East Asia and Pacific Region (EAP) (from 23 percent in FY03 to 13 percent in FY09) was mainly due to low disbursements in some large post-conflict TFs (such as the Timor-Leste TF), the MDTF for Aceh, Indonesia, and such other programs as the Vietnam free-standing TF (see Box 2.2), which are nearing completion. The share of the South Asia Region (SAR) remained steady at around 26 percent until FY08 and rose to 30 percent in FY09; the Afghanistan Reconstruction Trust Fund (ARTF) accounted for about 81 percent of SAR s disbursements from FY05 to FY09. Middle East and North Africa (MNA) region s disbursement share fluctuated from 11 percent in FY03 to 19 percent in FY05 to 6 percent in FY07 and 16 percent in FY09. This increase is mainly due to the Iraq TFs. Latin America and the Caribbean Regions (LCR) has the smallest share of RETFs on average 6 percent. This is understandable as the Region does not have many low income countries which are the major users of RETFs. 3 Page

14 Box 2. 2: RETF Portfolio in Vietnam During FY03-FY09, Vietnam had about 75 RETFs under implementation, with cumulative disbursements of about US$696 million. RETF disbursements rose from US$45 million in FY03 to US$103 million in FY09. About 13 percent of the total disbursements were from MDTFs. The largest-disbursing MDTF, Vietnam Primary Education for the Disadvantaged Children, accounted for 64 percent of the total MDTF disbursements. The next-largest MDTF disbursing program, GEF, accounted for 16 percent of total MDTFs in Vietnam. Among the single donors, 13 RETFs financed by the UK s Department of International Development accounted for about 32 percent of total single-donor RETF disbursements (US$604 million). The secondhighest disbursements (20 percent) were from Netherlands-funded programs. Programs funded by the EC (Belgium) and Denmark accounted for 12 and 13 percent, respectively. A notable feature of the RETF portfolio in Vietnam is the focus on poverty reduction programs: most of the donors have cofinanced a series of poverty reduction support credits. RETFs have also focused on the rural transport, education, and health sectors. B. Disbursements to Fragile and Conflict-Affected States 2.4 In recent years, the international development community has emphasized channeling increased amounts of financial resources to help conflict-affected countries and fragile states maintain any progress they have made and reach the Millennium Development Goals (MDGs), and to prevent the spill-over of political and social instability into neighboring countries. TFs, a major source of concessional financing for countries in this category, have mainly financed development activities, technical assistance, and advisory services for capacity building. This section examines the trends in RETF disbursements from FY03 to FY09 in these country types. 2.5 RETF grant disbursements in these conflict-affected and fragile states more than tripled from US$392 million in FY03 to US$1,280 million in FY09 (Figure 2.1) accounting for 42 percent of total RETF grant disbursements during that time, and an annualized growth in disbursements of 22 percent. The programs with the largest disbursements were the Afghanistan, West Bank and Gaza (WB&G), and Sudan MDTFs. 9 Afghanistan s trust funds have accounted for the lion s share about 51 percent of total disbursements in fragile states (Annex 5, Table 3). Without Afghanistan, the share of fragile states would be only 14 percent of total RETF disbursements. 9 WB&G is not a sovereign state and cannot apply for membership in the International Monetary Fund or the WBG institutions; therefore, it is not eligible for the sources of financing normally available to member countries. To overcome the legal and political obstacles, the Bank s Board of Governors approved a TF for WB&G for the area under the jurisdiction of the Palestinian Authority. The Bank began to reengage in Sudan in the early 2000 after an absence of nearly a decade. For Sudan, normal financial support from IDA is not possible because of the outstanding arrears Sudan has accumulated since Page

15 Figure 2.1: RETF Grant Disbursements in Fragile and Conflict-Affected States, FY03-FY09 (in US$M and as percent of total RETF disbursement) 1, , , FY03 FY04 FY05 FY06 FY07 FY08 FY09 0 RETF Disbursement in Fragile States Disbursements in Fragile States as % of Total RETF Disbursments C. Major Country Recipients 2.6 About half of the total RETF grant disbursements during FY03-FY09 were concentrated in six recipient states. Afghanistan accounted for the largest share (22 percent) of total RETF disbursements; the shares of other recipients were less than 10 percent each (Table 2.1). Table 2.1: Six Recipients with Highest Cumulative RETF Disbursements (FY03- FY09) Total Disbursement (US$M) Percent of Total RETF Recipient Disbursements Afghanistan 2, West Bank and Gaza 1,132 9 Ethiopia Vietnam Indonesia China Page

16 2.7 RETF disbursements in the fragile and conflict-affected/vulnerable states were also skewed toward a few recipients. Twelve countries in this category each disbursed more than US$50 million during FY03-FY09 (Annex 5, Table 3). Afghanistan and WB&G accounted for three-fourths of total RETF disbursements to fragile and conflict-affected countries, while others in this country category accounted for 5 percent or less. D. Disbursements of Recipient-Executed MDTFs 2.8 MDTFs have become an important part of the Bank s business as donors increasingly pool their resources to address post-conflict and post-disaster reconstruction activities, global/regional public goods (e.g., environment), and the achievement of the MDGs. Recipientexecuted MDTF disbursements increased from US$457 million in FY03 (47 percent of total RETF disbursements) to US$2,242 million (80 percent) in FY09 (Figure 2.2 and Annex 5, Table 4). The three largest-disbursing programs were the ARTF, GEF, and the Free-standing Cofinancing TF. Regionally, MDTFs in SAR disbursed, on average, about one-third of total recipient-executed MDTF grants from FY03 to FY09 (a dominance skewed by disbursements in Afghanistan), and those in AFR disbursed 28 percent. (Annex 5, Table 5 shows the changes in the regional shares when Afghanistan s RETFs are excluded.) Most of the MDTFs are Programmatic TFs. Figure 2.2: Total RETF and MDTF Disbursements (in US$M) by Fiscal Year and as Percent of Total RETFs 3,000 2,500 2,000 1,500 47% 54% 62% 73% 75% 80% 80% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 1, % % 10.0% - FY03 FY04 FY05 FY06 FY07 FY08 FY09 0.0% RE MDTF Disbursement. Total RETF Disbursement. % of RE MDTF Disbursement to total RETF Disbursement. 6 Page

17 E. Disbursements of Recipient-Executed Programmatic and Free-Standing TFs 2.9 Recipient-executed programmatic TFs 10 finance multiple grants to support a set of activities over several years, while free-standing TFs support a specific activity or set of activities in a specific country or region, or globally. In FY09 there were 195 active recipient-executed Programmatic TFs and 141 Free-standing TFs Of the total RETF disbursements during FY03-FY09, recipient-executed Programmatic TFs accounted for, on average, about 70 percent. The ARTF and GEF programs were the top-disbursing Programmatic TFs, accounting for 44 percent of total Programmatic TF disbursements (Annex 5, Table 6). The share of Free-standing TFs declined from FY03 to FY07, peaked in FY07 because of large disbursements from the free-standing cofinancing program that was introduced in FY06, and has remained steady since then. 11 The five highest-disbursing RETFs in each category are shown in Annex 5, Table 7. F. Disbursements of RETF Grants by Bank Operational Products 2.11 RETFs linked with Bank operations broadly reinforce the Bank s lending focus on sectors and themes at the country level. RETF grants may be linked to the following: IDA and IBRD lending operations for example, the project preparation and cofinancing grants of the Japan Policy and Human Resources Development Technical Assistance Program, and grants supported by the Japan Social Development Fund. Projects that are not included under IDA/IBRD lending operations for example, Rainforest Trust Fund, Ozone Phase-out TF, and WB&G TF. Developmental grants 12 are also included in this category. GEF operations and debt reduction programs. RETF grants that are not linked with any of the Bank s operational products are categorized as stand-alone recipient-executed projects for example, ARTF, MDTF for Aceh, MDTFs for North and South Sudan, and the Education for All Fast Track Initiative During FY03-FY09 the average disbursement shares of RETF grants linked with Bank-financed operations were comparable to those of the stand-alone projects, each accounting for about percent of total RETF disbursements. The share of stand-alone RETFs has steadily increased, while that of RETFs linked with Bank lending operations has fluctuated (see Table 2.2). The main reason for the declining disbursements of grants linked with Other recipient-executed projects is the closure or imminent closure of several large TFs 10 Programmatic TFs have a two-stage fund allocation process. In the first stage, the donor(s) agree on a broad framework to support a program over a number of years. In the second stage, donor(s) approve funds for specific activities. 11 The program included five grants, three for Ethiopia and one each for Bangladesh and the Democratic Republic of Congo. 12 Developmental grants are provided to external agencies such as Info Dev and NGOs, but for accounting purposes they are recorded as Bank-executed. 7 Page

18 those for the Federal Republic of Yugoslavia, Kosovo, East Timor, and WB&G which were financed from the Bank s net income and referred to as special financing. Table 2.2: RETF Grant Disbursements by Operational Product Category as Percentage of Total RETF Disbursements, FY03-FY09 RETF grants linked to: (i) Bank-financed (IBRD/IDA) lending (ii) Other recipient-executed project (iii) GEF and debt service operations RE grants linked to Stand-alone RE projects FY03 FY04 FY05 FY06 FY07 FY08 FY Total RETF disbursements (US$M) 981 1,151 1,474 1,450 2,098 2,584 2,811 G. Disbursements of RETF Grants by Usage 2.13 Disbursed RETFs were mainly used for technical assistance (TA) and cofinancing, which together constituted about 80 percent of the total RETF disbursements during FY TA usage was predominant (see Table 2.3). On average, about 38 percent of the total RETF disbursements during FY03-09 were for cofinancing grants to finance a range of activities: post-disaster rehabilitation activities (tsunami, avian flu), budget support, education, health, agriculture, environment, and so on. The highest-disbursing cofinancing grants were Ethiopia MDTF: Delivery of Basic Services; Bangladesh MDTF: Health, Nutrition and Population Project; and Ozone Phase-out for EAP. Between FY03 and FY09, the share of project preparation grants in total RETF disbursements declined by about 50 percent because the Japan Policy and Human Resources Development TA program, the largest-disbursing project preparation program, closed in FY08 after running for 19 years. Table 2.3: Share of RETF Disbursements by Usage (as percentage of total annual disbursements) TF Usage FY03 FY04 FY05 FY06 FY07 FY08 FY09 TA Cofinancing Project Preparation Partial Bank financing, Partial External financing a Other usage a These are special financing to states (e.g. WB&G) which are not members of the World Bank. 8 Page

19 III. RETFs Supporting GPPs 3.1 GPPs are an important instrument through which the Bank collaborates with donors on high-priority initiatives that complement its country and global activities. The TFs that support GPPs are mainly financed by multiple donors. This section looks at the features of BETFs and RETFs that support GPPs, with particular focus on their disbursement trends (Annex 3 has further details). 3.2 Disbursements from TFs supporting GPPs, as a share of total TF disbursements, almost doubled from 32 percent in FY03 to 59 percent in FY07 (Annex 5, Table 8). On average, about 44 percent of BETF disbursements and 22 percent of RETF disbursements supported GPPs. During this period the cumulative disbursements of RETFs supporting GPPs were US$1,567 million, and those of BETFs were US$563 million. 3.3 Disbursements from TFs supporting GPPs are concentrated among the larger programs (Annex 5, Table 9). In both FY03 and FY07, 61 percent of disbursements were from programs that disbursed more than US$5 million annually. In this disbursement range, the share of RETFs is significantly higher than other TF types, accounting for about 94 percent in FY03 and 81 percent in FY07. Between FY03 and FY07 disbursements increased significantly from programs in the US$5-50 million annual disbursement range, from US$66 million to US$155 million with a greater increase in disbursements from RETF-supported GPPs. 3.4 Of the 76 GPPs that were active in FY07, RETFs supported 26, which disbursed US$402 million an amount that had increased about 78 percent over the previous five years. Between FY03 and FY07 disbursements of RETFs supporting GPPs had increased in AFR and decreased in ECA (Annex 5, Table 10). Disbursements of RETFs supporting GPPs in fragile states have been relatively insignificant. 9 Page

20 IV. RETFs and Bank Lending 4.1 RETF disbursements, as a share of the combined IBRD, IDA, and RETF, grew steadily from FY03 to FY08. Although the absolute amount of RETF disbursements increased in FY09, the share fell by three percentage points mainly due to the significant increase in IBRD disbursements in the wake of the financial crisis. (Figure 4.1; Annex 5, Table 11). This section analyzes RETF disbursements by country groups and the relationships among the three funding sources. Figure 4.1: IBRD, IDA, and RETF Disbursements, FY03-09 US$ billion FY03 FY04 FY05 FY06 FY07 FY08 FY RETF disb as % of IBRD, IDA, RETF IBRD IDA RETF A. RETF Disbursements by Country Groups 4.2 The bulk of RETF disbursements are in IDA-only countries a share that has steadily increased from 50 percent in FY03 to 68 percent in FY09. In IBRD countries, the share of RETFs has remained between 15 and 18 percent; and for other country categories, the RETF disbursement shares declined (Annex 5, Table 12). 4.3 The region with the largest number of IDA-eligible countries AFR--accounted for 37 percent of the cumulative RETF disbursement in IDA countries during FY03-FY09. SAR accounted for 43 percent, with Afghanistan and Bangladesh accounting for 98 percent of the region s total and Afghanistan alone accounting for 85 percent. Three IDA-eligible countries Afghanistan, Ethiopia, and Vietnam accounted for 58 percent of total RETF disbursements in IDA countries. 4.4 In IBRD countries, about half of the RETF disbursements were in EAP. Among the IBRD-eligible countries, China, Indonesia, and Iraq disbursed the highest amounts during the FY03-FY09 period. The major disbursing programs in IBRD countries were the GEF, Ozone, and Iraq Trust Funds, which together accounted for about two-thirds of total RETF disbursement in this country type. 10 Page

21 B. Disbursements of RETFs, IDA Credits, and IBRD Loans 4.5 In IDA-eligible countries, disbursements from both IDA and RETFs were significantly higher in FY09 than in FY03. RETF disbursements more than tripled, from US$621 million to US$1.95 billion, and disbursements of IDA credits increased by 26 percent from US$7.29 billion to US$9.22 billion. (Annex 5, Table 13) 4.6 In IDA countries, regional disbursement patterns differed greatly. In AFR, IDA accounted for 47 percent of disbursements in both FY03 and FY09, and the share of RETF disbursements grew from 26 percent in FY03 to 42 percent in F Y09, nearly matching IDA disbursements. In SAR, the second-largest-disbursing region, IDA s disbursement share (27 percent in FY03, 30 percent in FY09) remained significantly below that of RETFs (42 percent in FY03, 43 percent in FY09). EAP accounted for the third-largest IDA disbursement (11 percent in FY03, increasing slightly in FY09), while the disbursement share of RETFs fell from 18 percent in FY03 to 8 percent in FY In the middle-income countries, an increased focus on environment, energy, and post-conflict situations brought about an overall increase of 67 percent in RETF disbursements, from US$156 million in FY03 to US$482 million in FY09. At the same time, IBRD loan disbursements increased by 56 percent, from US$11.92 billion in FY03 to US$18.56 billion in FY09 (Annex 5, Table 14) The highest RETF disbursement in IBRD countries was in EAP in both reference years. In the IBRD-eligible countries in the Middle East and North Africa Region (MNA), the share of RETF disbursements increased from only 3 percent in FY03 to 36 percent in FY09 a jump that was due mainly to the large disbursements from the Iraq TF. IBRD loan disbursements were concentrated in the Latin America and Caribbean Region (LCR) in both years, followed by EAP in FY03, and ECA in FY09. C. Disbursement Patterns of RETFs and IDA in Fragile and Conflict-Affected States 4.8 IDA and RETFs are increasingly important sources of financing for fragile and conflictaffected states. From FY03 to FY09, disbursements from IDA and RETFs in such states nearly doubled, from US$1,650 million to US$3,043 million. Even though IDA financing to these countries is somewhat constrained by the normal CPIA-based resource allocation system, IDA has been able to increase its levels of funds, albeit modestly, using a ring-fenced approach 13. An analysis of the disbursement trends reveals that in these challenging environments RETFs have substituted for IDA. Of the total amounts disbursed in these countries, IDA s FY03 share of 76 percent dropped to 58 percent in FY09, while the share of RETF disbursements almost doubled, from 24 percent to 42 percent. (Annex 5, Table 15) Even if the three states in this category that do not have access to IDA/IBRD financing (WB&G, Somalia, and Sudan) are taken out of the comparative analysis, the share of RETFs remains higher than IDA s. D. Disbursement Patterns of RETFs and IDA Credits and IBRD Loans by Sector 4.9 In general, RETF disbursements by sector complemented IDA s. The Public Administration and Law sector accounted for the bulk of IDA and RETF disbursements in both FY03 and FY09, and both IDA and RETF disbursements were high in the Health and Social 13 IDA Financial Assistance to Fragile States: Some Emerging Issues. World Bank (April 2009). 11 Page

22 Services and Education sectors. IDA s disbursement share in the Transportation sector has been comparatively higher than that of RETFs. (Annex 5, Table 16) 4.10 Unlike in IDA countries, in IBRD countries RETF disbursements have not been complementary to IBRD s. The RETF disbursement share has been the highest in the Industry and Trade (where disbursements from the GEF and Ozone Depletion programs have been high), Public Administration and Law, Energy and Mining, and Agriculture sectors. The IBRD disbursement shares have been the highest in Public Administration and Law, followed by the Finance and Transportation sectors (Annex 5, Table 17). V. RETFs and Country Performance 5.1 Whether RETF grants to recipient countries complement, substitute for, align with, or undermine IDA allocation discipline is an interesting question. While IDA allocations to countries are determined by the careful application of the Performance-Based Allocation (PBA) system, there is no comparable allocation mechanism for trust funds in general, or RETFs in particular (which often mirror IDA financing in many ways). The PBA formula gives considerable weight to CPIA scores and also includes per capita incomes, portfolio performance, population, and so on; but often the allocation of RETF grants does not seem to take country performance into account. The issue is briefly analyzed here, with the caveat that strong conclusions will require deeper and more thorough analysis. 5.2 A simple regression analysis was undertaken to examine whether there is any correlation between RETF commitments and country performance, using the Bank s annual CPIA ratings as the proxy for country performance. The analysis covered 74 low-income countries for the period FY03 to FY07 (Annex 4 provides details). For the full sample, the analysis shows that RETF grant commitments (in per capita terms) have a positive correlation with country performance, but the relationship is not strong. However, the relationship appears to be influenced by small countries, 14 such as Cape Verde and other island economies. Once the small countries are excluded from the sample, there is no longer a statistically significant correlation between RETF commitments and country performance. 5.3 Fragile states, which have low CPIA scores, have received a significant share of the RETF grants and thus could influence the results of the analysis. When both fragile states and small country observations are excluded from the sample, it is not surprising to see a strong correlation between RETF grant commitments per capita and country performance. At the same time, Armenia and Georgia appear to have strong influence on this relationship relative to other countries in the sample, pulling the trend line upward. 15 If Armenia and Georgia are excluded from the subsample, there is no correlation. 5.4 Overall, the results suggest that there may be some correlation between RETF grant commitments per capita and CPIA, but this is not robust. However, the question remains whether some ex-ante allocation criteria for determining country shares of RETFs would be possible. Would such a measure go directly against the spirit of flexibility and need-based 14 A small country is defined as one with a population of less than 1.5 million. 15 Georgia and Armenia can be referred to as the gap countries : with per capita GNI above the operational cut-off level for IDA credits, they still receive IDA credits on hardened terms as they are not yet IBRD-eligible. 12 Page

23 prioritization that is the hallmark of TFs? How could the Bank achieve a balance between equity and efficiency and country performance? These questions are worthy of further examination. VI. Conclusions 6.1 Overall, the growth of Bank-managed trust funds has been impressive. Disbursements of RETF grants grew 20 percent over FY03-FY09. Over the same period, the trend in the Regional shares of RETFs shifted away from SAR and EAP to AFR. Recipient concentration is evident: 50 percent of the disbursements were from six countries, of which two Afghanistan and WB&G are characterized as fragile states/entities. It would be interesting to study whether the focus on a few (non-fragile) recipients indicates need or donor preferences. 6.2 Disbursements of recipient-executed MDTFs doubled between FY03 and FY08, posing major challenges, particularly in fragile states, to achieving the desired results. An assessment by the Quality Assurance Group (QAG) of the Bank concluded that RETFs in countries with a low CPIA or without an existing lending program would have considerably better quality at entry if their design were more realistic The bulk of RETF disbursements are in IDA-only countries. The regional disbursement pattern of IDA funds is consistent with that for RETFs in these countries. However, disbursement patterns at the country level may reveal differences between RETFs and IDA. 6.4 The fact that, over time, the sectoral priorities for RETFs and IDA converge indicates that the RETFs are supporting the same sectors that IDA is financing. This concordance, in turn, presumably reflects country priorities as enunciated in Country Assistance Strategies (CASs). Recent efforts at mainstreaming TF activities in CASs show that TF resources are being considered as integral components of overall financing envelopes. These are increasingly being aligned with country priorities, outcomes, and results. 6.5 There may be some alignment between allocation of RETFs and country performance, but the results are not conclusive. The question remains whether trust fund allocations should be made on the basis of some need- or performance-based criteria. 6.6 The findings of this paper point toward some directions for future research and investigations. What is clear is that TFs have grown substantially to provide large volumes of financial resources which complements the Bank s country assistance programs, and also that RETF activities are a significant part of the Bank s operational portfolio. The Bank needs to manage this growing resource envelope and portfolio well. It needs to continue to apply to TF operations the same operational, fiduciary, social, and other safeguards that apply to regular Bank operations (IDA, IBRD, and IFC). RETFs, which mimic regular Bank lending operations in critical ways, are a special category for consideration. Yet, at the same time, the Bank needs to preserve the flexibility, nimbleness, and innovativeness that often categorize RETF activities. Thus the particular question that needs further consideration is whether RETFs should complement or substitute for Bank operations in specific cases. 16 Recipient Executed Trust Funds a QAG Assessment (March 2008). 13 Page