Low-Income Countries

International Financial Institutions Met Many Goals in Response to Financial, Food, and Fuel Crises, but Impact on Spending Difficult to Establish Gao ID: GAO-11-832 September 28, 2011

The 40 poorest countries in the world, known as low-income countries (LICs), have been negatively impacted by successive food, fuel, and financial crises since 2007. In response, international financial institutions (IFI), including the World Bank and International Monetary Fund (IMF), have taken actions to increase financial assistance for affected countries. Between 2008 and 2010, Congress appropriated $3.3 billion to the World Bank's International Development Association, which funds development programs in LICs. Congress also authorized the U.S. representative at the IMF to vote to approve the sale of some of the IMF's gold to increase lending to LICs. LICs' ability to repay debt remains important as financing levels rise and decisions are made about the mix of loans and grants they receive. GAO was asked to examine (1) the economic impact of the crises on LICs, (2) IFIs' responses and reported results, and (3) IFIs' assessment of the impact of the crises on LICs' ability to repay their debt. GAO analyzed documents and information from the World Bank and the IMF, including data on macroeconomic indicators, financial commitments, and debt analyses. GAO interviewed staff from the World Bank, IMF, and U.S. Treasury. GAO selected three African countries for more thorough analysis, a sample that is meant to be illustrative, not representative.

In LICs, the recent food, fuel, and financial crises resulted in slower economic growth, higher deficits, and higher inflation, but the macroeconomic impacts were less than experienced by the advanced economies. The crises also slowed foreign direct investment in LICs, which had been growing steadily since 2000. During the crises period, LICs' average economic growth slowed from 7.1 percent in 2007 to 5.3 percent in 2009. IFIs have reported that lower growth rates caused by the crises could lead to increases in poverty in LICs, and our previous work shows that many LICs were experiencing protracted food emergencies and had severe and widespread malnourishment even prior to the onset of the crises. During the crises, food and fuel prices rose significantly, then declined, and have risen again in 2011 to levels experienced during the crises. In response to the crises, IFIs increased funding and disbursed some funds more quickly to LICs, but the impact of these actions on LIC government spending has been difficult to establish. Between 2008 and 2010, the World Bank committed $18.1 billion through regular lending and five crisis response initiatives, an increase of 39 percent from the pre-crises period. Total first year disbursements also increased by 12.7 percent. Three of four of the initiatives designed to increase the speed of disbursements met their goal. However, the proportion of committed funds that have been disbursed in the first year following project approval declined, as compared to the pre-crises period. Disbursement rates depend on several factors, including recipient country capacity, need, and governance; and the type of lending. The World Bank's International Finance Corporation responded to the crises through investments, trade initiatives, and enhanced coordination with donors, but its response was limited by the availability of resources and recipient countries' limited ability to implement programs quickly. The IMF boosted lending to LICs more than sixfold to $4.9 billion, which governments could use to bolster their reserves or make international payments. While most LIC governments' spending increased during the crises, we found that the impact of World Bank and IMF actions on spending has been difficult to establish. According to IFIs' analysis, the crises did not significantly impair LICs' ability to repay their future debt, and thus did not necessitate an increase in their access to grants, which do not have to be repaid, relative to loans. The reliability of this analysis depends on the realism of IFIs' projections, which include quick economic recovery, implementation of policy reforms, and low inflation. According to IFIs' projections, the ability of six LICs to repay their debt improved during the crises, and thus they received more loans instead of grants. However, the IMF subsequently reported renewed risks to the global economic recovery, meaning that projections for future export growth, government revenue, and inflation might be too optimistic. This report contains no recommendations. The World Bank, IMF, and U.S Treasury generally agreed with our findings but identified areas to provide greater context.



GAO-11-832, Low-Income Countries: International Financial Institutions Met Many Goals in Response to Financial, Food, and Fuel Crises, but Impact on Spending Difficult to Establish This is the accessible text file for GAO report number GAO-11-832 entitled 'Low-Income Countries: International Financial Institutions Met Many Goals in Response to Financial, Food, and Fuel Crises, but Impact on Spending Difficult to Establish' which was released on October 26, 2011. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. United States Government Accountability Office: GAO: Report to Congressional Requesters: September 2011: Low-Income Countries: International Financial Institutions Met Many Goals in Response to Financial, Food, and Fuel Crises, but Impact on Spending Difficult to Establish: GAO-11-832: GAO Highlights: Highlights of GAO-11-832, a report to congressional requesters. Why GAO Did This Study: The 40 poorest countries in the world, known as low-income countries (LICs), have been negatively impacted by successive food, fuel, and financial crises since 2007. In response, international financial institutions (IFI), including the World Bank and International Monetary Fund (IMF), have taken actions to increase financial assistance for affected countries. Between 2008 and 2010, Congress appropriated $3.3 billion to the World Bank‘s International Development Association, which funds development programs in LICs. Congress also authorized the U.S. representative at the IMF to vote to approve the sale of some of the IMF‘s gold to increase lending to LICs. LICs‘ ability to repay debt remains important as financing levels rise and decisions are made about the mix of loans and grants they receive. GAO was asked to examine (1) the economic impact of the crises on LICs, (2) IFIs‘ responses and reported results, and (3) IFIs‘ assessment of the impact of the crises on LICs‘ ability to repay their debt. GAO analyzed documents and information from the World Bank and the IMF, including data on macroeconomic indicators, financial commitments, and debt analyses. GAO interviewed staff from the World Bank, IMF, and U.S. Treasury. GAO selected three African countries for more thorough analysis, a sample that is meant to be illustrative, not representative. This report contains no recommendations. The World Bank, IMF, and U.S Treasury generally agreed with our findings but identified areas to provide greater context. What GAO Found: In LICs, the recent food, fuel, and financial crises resulted in slower economic growth, higher deficits, and higher inflation, but the macroeconomic impacts were less than experienced by the advanced economies. The crises also slowed foreign direct investment in LICs, which had been growing steadily since 2000. During the crises period, LICs' average economic growth slowed from 7.1 percent in 2007 to 5.3 percent in 2009. IFIs have reported that lower growth rates caused by the crises could lead to increases in poverty in LICs, and our previous work shows that many LICs were experiencing protracted food emergencies and had severe and widespread malnourishment even prior to the onset of the crises. During the crises, food and fuel prices rose significantly, then declined, and have risen again in 2011 to levels experienced during the crises. In response to the crises, IFIs increased funding and disbursed some funds more quickly to LICs, but the impact of these actions on LIC government spending has been difficult to establish. Between 2008 and 2010, the World Bank committed $18.1 billion through regular lending and five crisis response initiatives, an increase of 39 percent from the pre-crises period. Total first year disbursements also increased by 12.7 percent. Three of four of the initiatives designed to increase the speed of disbursements met their goal. However, the proportion of committed funds that have been disbursed in the first year following project approval declined, as compared to the pre-crises period. Disbursement rates depend on several factors, including recipient country capacity, need, and governance; and the type of lending. The World Bank‘s International Finance Corporation responded to the crises through investments, trade initiatives, and enhanced coordination with donors, but its response was limited by the availability of resources and recipient countries‘ limited ability to implement programs quickly. The IMF boosted lending to LICs more than sixfold to $4.9 billion, which governments could use to bolster their reserves or make international payments. While most LIC governments‘ spending increased during the crises, we found that the impact of World Bank and IMF actions on spending has been difficult to establish. According to IFIs‘ analysis, the crises did not significantly impair LICs‘ ability to repay their future debt, and thus did not necessitate an increase in their access to grants, which do not have to be repaid, relative to loans. The reliability of this analysis depends on the realism of IFIs‘ projections, which include quick economic recovery, implementation of policy reforms, and low inflation. According to IFIs‘ projections, the ability of six LICs to repay their debt improved during the crises, and thus they received more loans instead of grants. However, the IMF subsequently reported renewed risks to the global economic recovery, meaning that projections for future export growth, government revenue, and inflation might be too optimistic. In addition, for the three countries we reviewed, macroeconomic projections did not adequately take into account country-specific vulnerabilities, such as the failure to implement reforms and make planned investments. However, given that the IFIs update projections on a regular basis, any excessive optimism should become evident over time, and some lenders could then increase the amount of grants they provide which would help mitigate potential debt problems. View [hyperlink, http://www.gao.gov/products/GAO-11-832] or key components. For more information, contact Thomas Melito at (202) 512- 9601 or melitot@gao.gov. [End of section] Contents: Letter: Background: Crises Negatively Impacted LICs, but Less Than Experienced by Advanced Economies: IFIs Increased Funding and Met Many Disbursement Goals, but the Impact of Their Actions on LIC Government Spending Has Been Difficult to Establish: IFIs Did Not Lower Debt Distress Ratings Due to the Crises, but If Underlying Projections Prove Too Optimistic Then Grants Could Be Increased: Concluding Observations: Agency Comments and Our Evaluation: Appendix I: Objectives, Scope, and Methodology: Appendix II: Economic, Demographic, and Financial Assistance Data for LICs: Appendix III: World Bank Commitments and Disbursements to LICs through Crisis Response Initiatives: Appendix IV: Comments from the U.S. Department of the Treasury: Appendix V: Comments from the World Bank: Appendix VI: Comments from the International Monetary Fund: Appendix VII: GAO Contact and Staff Acknowledgments: Figures: Figure 1: Selected Crises-Related Events and Actions: Figure 2: Economic, Demographic, and Financial Assistance Data for 40 LICs: Figure 3: Real GDP Weighted Average Growth Rate for 38 LICs, 2004 through 2010: Figure 4: Consumer Price Inflation and Real GDP Growth Rate for 38 LICs, 2004 through 2010: Figure 5: World Food Prices, January 2001 through May 2011: Figure 6: World Crude Oil Prices in U.S. Dollars per Barrel, January 2007 through June 2011: Figure 7: Current Account Deficit as a Percent of GDP for 38 LICs, 2004 through 2010: Figure 8: Foreign Direct Investment Net Inflows for 39 LICs in Current U.S. Dollars, 2000 through 2009: Figure 9: First Year Disbursement Rates of Initiative Lending and Non- Initiative Lending, Ranked from Fastest to Slowest: Figure 10: Percentage Point Change in First Year Disbursement Rates from the Pre-Crises Period to the Crises Response Period: Figure 11: Performance Ratings and Associated Debt Burden Threshold Indicators: Figure 12: Risk of Debt Distress and Performance Ratings for 37 LICs, as of December 2010: Figure 13: Changes in 10 Countries' Debt Distress Ratings and Financing Terms, 2007 through 2010: Figure 14: Distribution of World Bank Commitments to 38 LICs, by Country, 2008-2010: Figure 15: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds through the Global Food Crisis Response Program, by Country: Figure 16: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds through the Fast Track Facility, by Country: Figure 17: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds through the Rapid Social Response Program, by Country: Figure 18: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds, Pilot Crisis Response Window, by Country: Abbreviations: CPIA: Country Policy and Institutional Assessment: DSA: debt sustainability analysis: G-20: Group of Twenty: GDP: gross domestic product: IDA: International Development Association: IFI: international financial institution: IMF: International Monetary Fund: IFC: International Finance Corporation: LIC: low-income country: MDB: multilateral development bank: MDG: Millennium Development Goal: ODA: Official Development Assistance: PPP: purchasing power parity: SDR: Special Drawing Right: UN: United Nations: USAID: U.S. Agency for International Development: U.S. Treasury: U.S. Department of the Treasury: [End of section] United States Government Accountability Office: Washington, DC 20548: September 28, 2011: The Honorable Barney Frank: Ranking Member: Committee on Financial Services: House of Representatives: The Honorable Gregory Meeks: House of Representatives: [End of section] The 40 poorest countries in the world, known as low-income countries (LICs), have been negatively impacted by successive and overlapping food, fuel, and financial crises since 2007.[Footnote 1] International financial institutions (IFIs), including the World Bank and the International Monetary (IMF), report that these crises have had a severe impact on poverty.[Footnote 2] In addition, the World Bank reports that the crises led to significant increases in hunger and infant mortality, impacting millions of people in developing countries. In response to the crises, IFIs have taken actions to increase financial assistance for affected countries. Between 2008 and 2010, Congress appropriated $3.3 billion to the World Bank's International Development Association (IDA), which funds health, education, infrastructure, agriculture, economic, and institutional development programs in about 80 developing countries, including LICs.[Footnote 3] During this period, IDA launched crises response programs in LICs to protect the poor and vulnerable, maintain and increase social and infrastructure spending, and--in conjunction with the World Bank's International Finance Corporation (IFC)--support and promote private sector activity. In 2009, Congress authorized the Secretary of the Treasury to instruct the U.S. Executive Director to the IMF to vote to approve the sale of a limited amount of the IMF's gold to, among other things, provide additional lending to LICs. [Footnote 4] The IMF said it would respond to the crises in part by increasing below-market-rate financing to LICs with crises-related balance-of-payments problems.[Footnote 5] Since the 1990s, many LICs have received a significant amount of debt relief from the IMF, the World Bank, other multilateral creditors, country creditors, and commercial creditors to lower debt to levels considered "sustainable," meaning the country can make its future debt payments on time and without rescheduling. Recognizing that debt relief recipients remain vulnerable to future debt problems, caused by shocks, such as food and fuel price increases or adverse weather events, the World Bank and IMF jointly conduct a debt sustainability analysis (DSA) to assess how a country's current level of debt and prospective new borrowing affect its ability to service its future debt. An important output of the DSA is a country's debt distress rating, which is used to determine the mix of grants and loans provided by IDA to countries. For example, countries classified at a high risk of debt distress receive 100 percent grants from IDA, which do not have to be repaid, while countries at moderate risk receive a blend of 50 percent grants and 50 percent loans, which are provided at below-market rates. You asked us to assess the IFIs' response to LICs' needs during the crises. We examined (1) the economic impact of the crises on LICs, (2) IFIs' responses and reported results, and (3) IFIs' assessment of the impact of the crises on LICs' ability to repay their debt. To address our objectives, we analyzed documents and data from the U.S. Department of the Treasury (U.S. Treasury), the World Bank, and the IMF. We interviewed officials from these institutions, as well as from the Department of State, and the U.S. Agency for International Development (USAID). We compiled data for macroeconomic indicators from IMF, World Bank, and United Nations (UN) databases and compared each LIC's macroeconomic performance during the crises period of 2007 through 2009 to the pre-crisis period of 2004 through 2006.[Footnote 6] We selected three countries in Africa for more thorough analysis: Burundi, Ethiopia, and Tanzania. We selected these countries based on several criteria, including the number of IFI projects and amount of IFI financial support. Our selection of countries is nongeneralizable and meant to be illustrative, not representative. We focused on the amount and speed of financial assistance and the extent to which IFI actions caused increases in recipient governments' spending, but we recognize that there are other less quantifiable considerations for assessing the impacts and effectiveness of IFI financial assistance. We analyzed World Bank data on financial commitments and disbursements made to LICs between 2005 and 2007, the pre-crisis period, and between 2008 and 2010, the crises response period, to evaluate whether the World Bank's crisis response was consistent with the institutions' stated goals. We used information on IMF program and funding levels to determine the nature and size of the IMF's response to the crises and assessed the sensitivity of the results of an IMF analysis of government spending in LICs in 2009 using data from the IMF and the World Governance Indicators. We analyzed joint IMF-World Bank debt analyses, IMF program documents, and World Bank Country Assistance Strategies to assess IMF-World Bank efforts to monitor debt sustainability for the three case study countries. Our assessment of a country's debt distress rating, or the degree to which a country faces repayment difficulties, included an examination of the debt analyses' underlying macroeconomic projections and assumptions. We conducted this performance audit from September 2010 to September 2011 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. (See appendix I for a more detailed discussion of our scope and methodology.) Background: Since 2007, a series of overlapping food, fuel, and financial crises have negatively affected the world economy, prompting the IFIs to respond. (See figure 1.) Figure 1: Selected Crises-Related Events and Actions: [Refer to PDF for image: timeline] Subprime mortgage crisis in the United States: 2007, Q1, Q2. Financial crisis in advanced economies: 2007 Q2 - 2008 Q3. Global economic crisis: 2008 Q3 - 2009 Q2. Gradual recovery: 2009 Q3 - 2010 Q4. Fuel price increase: 2007 Q1 - 2008 Q2; Fuel price decrease: 2008 Q2 - 2009 Q2; Fuel price increase: 2009 Q3 - 2010 Q4. Gradual food price increase: 2007 Q1 - 2007 Q4; Food crisis: 2007 Q4 - 2008 Q2; Gradual food price decrease: 2008 Q2 - 2009 Q2; Food price increase: 2009 Q3 - 2010 Q4. 2007 Q4: Beginning of food export restrictions. June 2008: World Bank announces food crisis response; Food price peak. December 2008: World Bank announces financial crisis response. March 2009: World Bank announces crises response framework. July 2009: The IMF announces planned increase in financial assistance for LICs. December 2009: World Bank announces additional resources for LICs. Source: GAO based on World Bank, IMF, Energy Information Agency Administration, and UN Food and Agriculture Organization information. [End of figure] Between 2007 and 2009, a financial crisis in the U.S. subprime mortgage market gradually extended to other developed countries and became a global financial crisis, which in turn generated a global economic crisis impacting developed and developing countries-- including LICs--with varying degrees of intensity. Prior to, and partly in parallel with the financial crisis, global food and fuel prices increased sharply between mid-2007 and mid-2008. In some cases this exacerbated the impact of the financial crisis and some governments banned food exports. Although international food and fuel prices dropped precipitously as the global financial crisis unfolded in late 2008, they have resurged to record highs again in 2011, igniting concerns about a repeat of the 2008 food crisis and its negative impact on poor people in developing countries, particularly LICs. IFIs and global leaders have launched a coordinated international response to the crises. In March 2009, the World Bank announced a comprehensive crises response framework that could channel additional donor contributions to poor countries to bolster ongoing World Bank activities. To help ensure that IFIs would have sufficient resources to respond to the crises, in April 2009 the Group of Twenty (G-20) world leaders committed to measures designed to increase IFI resources available to LICs, including through voluntary bilateral contributions to the World Bank's crises response framework.[Footnote 7] The G-20 also endorsed the IMF's intention to increase financing for LICs, including from resources derived from proceeds of IMF gold sales. [Footnote 8] LICs, the majority of which are in sub-Saharan Africa, comprised a population of about 810 million people at the outset of the crises in 2007. Most LICs depend to some extent upon imports of food and fuel, and about half are classified as fragile states challenged by weak capacity, poor governance, political instability, ongoing violence or the legacy of past conflict, as shown in figure 2. According to IFIs, these factors can render countries vulnerable to crises driven by fluctuations in international food and fuel prices. Figure 2: Economic, Demographic, and Financial Assistance Data for 40 LICs: [Refer to PDF for image: interactive map of the world] Instructions: Online, hover over a country name in the table for more information. For print version, see appendix II, page 53. Country: Afghanistan; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Bangladesh; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Benin; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Burkina Faso; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Burma (Myanmar); Net food importer[A]: No; Net energy consumer[B]: No; Fragile[C]: Yes. Country: Burundi; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Cambodia; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Central African Republic; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Chad; Net food importer[A]: Yes; Net energy consumer[B]: No; Fragile[C]: Yes. Country: Comoros; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Democratic Republic of Congo; Net food importer[A]: Yes; Net energy consumer[B]: No; Fragile[C]: Yes. Country: Eritrea; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Ethiopia; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Gambia, The; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Ghana; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Guinea; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Guinea-Bissau; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Haiti; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Kenya; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Kyrgyzstan; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Laos; Net food importer[A]: Yes; Net energy consumer[B]: No; Fragile[C]: No. Country: Liberia; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Madagascar; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Malawi; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Mali; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Mauritania; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Mozambique; Net food importer[A]: Yes; Net energy consumer[B]: No; Fragile[C]: No. Country: Nepal; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Niger; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: North Korea; Net food importer[A]: Yes; Net energy consumer[B]: No; Fragile[C]: No. Country: Rwanda; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Sierra Leone; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Solomon Islands; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Somalia; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Tajikistan; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Tanzania; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Togo; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Country: Uganda; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Zambia; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: No. Country: Zimbabwe; Net food importer[A]: Yes; Net energy consumer[B]: Yes; Fragile[C]: Yes. Sources: GAO analysis of World Bank UN and US Department of Energy data; Map Resources (map). [A] In food-importing countries, imports of basic foodstuffs outweighed exports over the past 3 years, according to the UN Food and Agriculture Organization. [B] We define net energy consumers as those countries for which net energy production was less than consumption in 2009. [C] The World Bank uses the term "fragile" to refer to countries with particularly weak policies and institutions, as well as those with the presence of a UN or regional peace-keeping or peace-building mission during the past 3 years. [End of figure] The World Bank and IMF provide financial and technical assistance to member countries. Two World Bank institutions--IDA and IFC--assist LICs. IDA, the primary World Bank financer to LICs, provides no- interest loans and grants to eligible countries that have limited or no access to international credit markets.[Footnote 9] IDA funds long- term programs in agriculture, infrastructure, and social services such as health and education, and provides technical assistance for programs in economic and institutional development to strengthen country policies and institutional capacity. Commitments to these programs are disbursed at different rates depending on a number of factors, including recipient country capacity and whether the project is an investment lending project or a development policy lending project.[Footnote 10] The rate of disbursements is important because committed funds cannot be used by recipient countries until the funds are disbursed. IFC provides investments and advisory services to build the private sector in developing countries, including LICs.[Footnote 11] The IMF provides economic surveillance, lending, and technical assistance to its member countries. IMF surveillance involves the monitoring of economic and financial developments and the provision of policy advice. The primary purpose of IMF lending is to assist countries facing balance-of-payments difficulties, and IMF loans to LICs are intended to help foster economic growth and reduce poverty. IMF lending is conditional upon borrowing countries' implementation of policies.[Footnote 12] To help countries manage their economies, the IMF provides guidance and training on how to strengthen institutions and design appropriate macroeconomic, financial, and structural policies. Since 1996 the World Bank and IMF have participated in bilateral and multilateral efforts to relieve the debt burdens of poor countries to help them achieve long-term economic growth and debt sustainability, meaning they can make their future debt payments on time without rescheduling. To assess how a country's current level of debt and prospective new borrowing affect its ability to service its debt in the future, the World Bank and IMF jointly conduct a DSA. DSAs include an analysis of a country's projected debt burden over the next 20 years and its vulnerability to shocks. In 2009, we reported that the World Bank and IMF had improved their DSAs, including by considering the strength of a country's policies and institutions, and that the DSAs identified numerous ambitious actions countries should take in order to avoid future unsustainable debt levels.[Footnote 13] Crises Negatively Impacted LICs, but Less Than Experienced by Advanced Economies: The food, fuel, and financial crises negatively impacted LIC economies, but the slowdown in growth was less than experienced by advanced economies. Our analysis shows that during the crises period of 2007 through 2009, key economic indicators slowed or declined for 38 LICs compared to the pre-crises period of 2004 through 2006. [Footnote 14] Economic Growth Slowed and Inflation Increased: While the average annual growth rate in real gross domestic product (GDP),[Footnote 15] or national income, for the 38 LICs remained positive during the crisis period, it declined by an average of about 1 percentage point, dropping from an average of 7.1 percent during the pre-crises period to an average of 6.2 percent during the crises period. The largest decline occurred between 2007 and 2009 when real GDP growth fell nearly 2 percentage points, from 7.1 percent to 5.3 percent. (See figure 3.) Nine countries experienced an actual decline in their real GDP in 2008 or 2009: Cambodia, Chad, Eritrea, Guinea, Madagascar, Mauritania, Niger, Solomon Islands, and Zimbabwe. Figure 3: Real GDP Weighted Average Growth Rate for 38 LICs, 2004 through 2010: [Refer to PDF for image: line graph] Crisis period: mid-2006 through mid-2009. Year: 2004; Percent change in real GDP: 7.04. Year: 2005; Percent change in real GDP: 7.2%. Year: 2006; Percent change in real GDP: 7.04. Year: 2007; Percent change in real GDP: 7.11. Year: 2008; Percent change in real GDP: 6.09. Year: 2009; Percent change in real GDP: 5.29. Year: 2010; Percent change in real GDP: 5.62. Source: GAO analysis of World Economic Outlook data as of April 2011. Note: The weighted average is an aggregate reflecting the relative size of countries in terms of their share in total GDP of the entire group. The derivation of these weights requires that GDP in national currency terms be converted to a common currency (in practice, the U.S. dollar). Since 1993, exchange rates based on purchasing power parities (PPP) have been used for this purpose. PPP-based GDP takes differences in price levels across countries into account to ensure that the GDP weights reflect each country's share in real output. [End of figure] The slowdown in the growth rate of real GDP in LICs was milder than the downturn in real GDP growth experienced by advanced economies during the crises, which declined from 2.7 percent in 2007 to -3.4 percent in 2009. According to the IMF, the LICs' period of growth prior to the crises provided a cushion, helping many countries weather the food and fuel price increases between 2007 and 2008 and the global financial crisis. However, IFIs have reported that lower growth rates caused by the crises could lead to increases in poverty in LICs. Moreover, the slowdown in real GDP growth occurred while LICs' inflation was rising. The average annual inflation rate for 38 LICs increased from 8.6 percent during the pre-crises period to an average of 11.6 percent during the crises, peaking at nearly 14 percent in 2008. (See figure 4.) Figure 4: Consumer Price Inflation and Real GDP Growth Rate for 38 LICs, 2004 through 2010: [Refer to PDF for image: line graph] Crisis period: mid-2006 through mid-2009. Year: 2004; CPI inflation: 7.40%; Real GDP growth rate: 7.04%. Year: 2005; CPI inflation: 9.07%; Real GDP growth rate: 7.23%. Year: 2006; CPI inflation: 9.27%; Real GDP growth rate: 7.04%. Year: 2007; CPI inflation: 10.23%; Real GDP growth rate: 7.12%. Year: 2008; CPI inflation: 13.88%; Real GDP growth rate: 6.09%. Year: 2009; CPI inflation: 10.61%; Real GDP growth rate: 5.29%. Year: 2010; CPI inflation: 7.36%; Real GDP growth rate: 5.62%. Source: GAO analysis of World Economic Outlook data as of April 2011. Note: The weighted average is an aggregate reflecting the relative size of countries in terms of their share in total GDP of the entire group. The derivation of these weights requires that GDP in national currency terms be converted to a common currency (in practice, the U.S. dollar). Since 1993, exchange rates based on purchasing power parities (PPP) have been used for this purpose. PPP-based GDP takes differences in price levels across countries into account to ensures that the GDP weights reflect each country's share in real output. [End of figure] Higher food and fuel prices contributed to rising inflation. World food prices were stable from 2001 to the beginning of 2007, and then climbed steeply during 2007 and 2008.[Footnote 16] After a brief downturn in the latter part of 2008, world food prices began rising again, resulting in a net increase in prices of about 74 percent between January 2007 and May 2011. (See figure 5.) Figure 5: World Food Prices, January 2001 through May 2011: [Refer to PDF for image: line graph] Crisis period: 1/2007 through 1/2010. Month/Year: 1/2001; Food price index: 97.6. Month/Year: 2/2001; Food price index: 98.5. Month/Year: 3/2001; Food price index: 99.9. Month/Year: 4/2001; Food price index: 98.2. Month/Year: 5/2001; Food price index: 99.7. Month/Year: 6/2001; Food price index: 98.5. Month/Year: 7/2001; Food price index: 101.2. Month/Year: 8/2001; Food price index: 101.1. Month/Year: 9/2001; Food price index: 99.4. Month/Year: 10/2001; Food price index: 97.4. Month/Year: 11/2001; Food price index: 98.3. Month/Year: 12/2001; Food price index: 97.9. Month/Year: 1/2002; Food price index: 96.9. Month/Year: 2/2002; Food price index: 94.9. Month/Year: 3/2002; Food price index: 95. Month/Year: 4/2002; Food price index: 93.3. Month/Year: 5/2002; Food price index: 91.6. Month/Year: 6/2002; Food price index: 92.5. Month/Year: 7/2002; Food price index: 94.7. Month/Year: 8/2002; Food price index: 96.3. Month/Year: 9/2002; Food price index: 99.9. Month/Year: 10/2002; Food price index: 100.4. Month/Year: 11/2002; Food price index: 102.3. Month/Year: 12/2002; Food price index: 101.7. Month/Year: 1/2003; Food price index: 95.5. Month/Year: 2/2003; Food price index: 97.1. Month/Year: 3/2003; Food price index: 95.3. Month/Year: 4/2003; Food price index: 94.5. Month/Year: 5/2003; Food price index: 95.3. Month/Year: 6/2003; Food price index: 95.6. Month/Year: 7/2003; Food price index: 94.7. Month/Year: 8/2003; Food price index: 96.2. Month/Year: 9/2003; Food price index: 98.2. Month/Year: 10/2003; Food price index: 100.8. Month/Year: 11/2003; Food price index: 103.6. Month/Year: 12/2003; Food price index: 105.3. Month/Year: 1/2004; Food price index: 101.4. Month/Year: 2/2004; Food price index: 102.5. Month/Year: 3/2004; Food price index: 105.9. Month/Year: 4/2004; Food price index: 106. Month/Year: 5/2004; Food price index: 104.6. Month/Year: 6/2004; Food price index: 106.6. Month/Year: 7/2004; Food price index: 105.7. Month/Year: 8/2004; Food price index: 105.1. Month/Year: 9/2004; Food price index: 105.5. Month/Year: 10/2004; Food price index: 104.8. Month/Year: 11/2004; Food price index: 106.3. Month/Year: 12/2004; Food price index: 107 . Month/Year: 1/2005; Food price index: 107.4. Month/Year: 2/2005; Food price index: 106.7. Month/Year: 3/2005; Food price index: 109.4. Month/Year: 4/2005; Food price index: 107.3. Month/Year: 5/2005; Food price index: 108.4. Month/Year: 6/2005; Food price index: 109.5. Month/Year: 7/2005; Food price index: 109.4. Month/Year: 8/2005; Food price index: 109.3. Month/Year: 9/2005; Food price index: 111.1. Month/Year: 10/2005; Food price index: 112.6. Month/Year: 11/2005; Food price index: 111.4. Month/Year: 12/2005; Food price index: 113.6. Month/Year: 1/2006; Food price index: 111.5. Month/Year: 2/2006; Food price index: 115.6. Month/Year: 3/2006; Food price index: 113.3. Month/Year: 4/2006; Food price index: 114.8. Month/Year: 5/2006; Food price index: 115.5. Month/Year: 6/2006; Food price index: 114.8. Month/Year: 7/2006; Food price index: 117.5. Month/Year: 8/2006; Food price index: 115.9. Month/Year: 9/2006; Food price index: 115.1. Month/Year: 10/2006; Food price index: 118.1. Month/Year: 11/2006; Food price index: 122. Month/Year: 12/2006; Food price index: 123.7. Month/Year: 1/2007; Food price index: 117.7. Month/Year: 2/2007; Food price index: 120. Month/Year: 3/2007; Food price index: 120.8. Month/Year: 4/2007; Food price index: 123.6. Month/Year: 5/2007; Food price index: 127.3. Month/Year: 6/2007; Food price index: 135.4. Month/Year: 7/2007; Food price index: 140.8. Month/Year: 8/2007; Food price index: 146.3. Month/Year: 9/2007; Food price index: 154.1. Month/Year: 10/2007; Food price index: 156.7. Month/Year: 11/2007; Food price index: 162.8. Month/Year: 12/2007; Food price index: 167.7. Month/Year: 1/2008; Food price index: 164.5. Month/Year: 2/2008; Food price index: 177.3. Month/Year: 3/2008; Food price index: 179.7. Month/Year: 4/2008; Food price index: 178.9. Month/Year: 5/2008; Food price index: 179.9. Month/Year: 6/2008; Food price index: 184.7. Month/Year: 7/2008; Food price index: 181.4. Month/Year: 8/2008; Food price index: 171.9. Month/Year: 9/2008; Food price index: 161.9. Month/Year: 10/2008; Food price index: 142. Month/Year: 11/2008; Food price index: 129.5. Month/Year: 12/2008; Food price index: 122. Month/Year: 1/2009; Food price index: 125.7. Month/Year: 2/2009; Food price index: 121.4. Month/Year: 3/2009; Food price index: 123. Month/Year: 4/2009; Food price index: 126.8. Month/Year: 5/2009; Food price index: 135.4. Month/Year: 6/2009; Food price index: 135.9. Month/Year: 7/2009; Food price index: 132.6. Month/Year: 8/2009; Food price index: 137.1. Month/Year: 9/2009; Food price index: 137.4. Month/Year: 10/2009; Food price index: 140. Month/Year: 11/2009; Food price index: 150.3. Month/Year: 12/2009; Food price index: 153. Month/Year: 1/2010; Food price index: 153.6. Month/Year: 2/2010; Food price index: 150.3. Month/Year: 3/2010; Food price index: 143.8. Month/Year: 4/2010; Food price index: 145.2. Month/Year: 5/2010; Food price index: 144.8. Month/Year: 6/2010; Food price index: 143.6. Month/Year: 7/2010; Food price index: 147.3. Month/Year: 8/2010; Food price index: 156.1. Month/Year: 9/2010 Food price index: 165.7. Month/Year: 10/2010; Food price index: 174.9. Month/Year: 11/2010; Food price index: 181.6. Month/Year: 12/2010; Food price index: 190.6. Month/Year: 1/2011; Food price index: 203.3. Month/Year: 2/2011; Food price index: 209.1. Month/Year: 3/2011; Food price index: 203.8. Month/Year: 4/2011; Food price index: 206.5. Month/Year: 5/2011; Food price index: 204.4. Source: UN Food and Agriculture Organization of Monthly World Food Price Index. Note: The monthly index includes meat, dairy, cereals, oil, and sugar. [End of figure] Similarly, world crude oil prices rose sharply during 2007 and 2008 and, after receding through early 2009, rose again through May 2011, resulting in a net price increase of over 99 percent when compared to January 2007. (See figure 6.) Figure 6: World Crude Oil Prices in U.S. Dollars per Barrel, January 2007 through June 2011: [Refer to PDF for image: line graph] Crisis period: 2/2007 through 1/2010. Month/Year: 1/2007; Dollars per barrel: $54.63. Month/Year: 2/2007; Dollars per barrel: $52.11. Month/Year: 3/2007; Dollars per barrel: $57.83. Month/Year: 4/2007; Dollars per barrel: $64.93. Month/Year: 5/2007; Dollars per barrel: $63.4. Month/Year: 6/2007; Dollars per barrel: $66.79. Month/Year: 7/2007; Dollars per barrel: $69.91. Month/Year: 8/2007; Dollars per barrel: $73.81. Month/Year: 9/2007; Dollars per barrel: $71.42. Month/Year: 10/2007; Dollars per barrel: $75.57. Month/Year: 11/2007; Dollars per barrel: $89.64. Month/Year: 12/2007; Dollars per barrel: $85.91. Month/Year: 1/2008; Dollars per barrel: $92.93. Month/Year: 2/2008; Dollars per barrel: $88.71. Month/Year: 3/2008; Dollars per barrel: $98.01. Month/Year: 4/2008; Dollars per barrel: $98.39. Month/Year: 5/2008; Dollars per barrel: $115.11. Month/Year: 6/2008; Dollars per barrel: $121.36. Month/Year: 7/2008; Dollars per barrel: $137.11. Month/Year: 8/2008; Dollars per barrel: $121.29. Month/Year: 9/2008; Dollars per barrel: $106.41. Month/Year: 10/2008; Dollars per barrel: $93.38. Month/Year: 11/2008; Dollars per barrel: $58.66. Month/Year: 12/2008; Dollars per barrel: $43.12. Month/Year: 1/2009; Dollars per barrel: $34.57. Month/Year: 2/2009; Dollars per barrel: $42.07. Month/Year: 3/2009; Dollars per barrel: $42.85. Month/Year: 4/2009; Dollars per barrel: $50.27. Month/Year: 5/2009; Dollars per barrel: $52.92. Month/Year: 6/2009; Dollars per barrel: $65.62. Month/Year: 7/2009; Dollars per barrel: $68.47. Month/Year: 8/2009; Dollars per barrel: $67.5. Month/Year: 9/2009; Dollars per barrel: $68.07. Month/Year: 10/2009; Dollars per barrel: $65. Month/Year: 11/2009; Dollars per barrel: $76.34. Month/Year: 12/2009; Dollars per barrel: $73.89. Month/Year: 1/2010; Dollars per barrel: $78.89. Month/Year: 2/2010; Dollars per barrel: $70.42. Month/Year: 3/2010; Dollars per barrel: $78.17. Month/Year: 4/2010; Dollars per barrel: $82.57. Month/Year: 5/2010; Dollars per barrel: $76.56. Month/Year: 6/2010; Dollars per barrel: $73.85. Month/Year: 7/2010; Dollars per barrel: $73.49. Month/Year: 8/2010; Dollars per barrel: $73.81. Month/Year: 9/2010; Dollars per barrel: $76.11. Month/Year: 10/2010; Dollars per barrel: $79.35. Month/Year: 11/2010; Dollars per barrel: $82.82. Month/Year: 12/2010; Dollars per barrel: $88.34. Month/Year: 1/2011; Dollars per barrel: $92.6. Month/Year: 2/2011; Dollars per barrel: $97.78. Month/Year: 3/2011; Dollars per barrel: $108.03. Month/Year: 4/2011; Dollars per barrel: $118.38. Month/Year: 5/2011; Dollars per barrel: $108.32. Month/Year: 6/2011; Dollars per barrel: $113.55. Source: GAO analysis of U.S. Department of Energy, Energy Information Agency. [End of figure] While the human and social development impacts of higher prices vary by country, the resurgence in prices has triggered renewed concern. Our previous work shows that many LICs were experiencing protracted food emergencies and had severe and widespread malnourishment even prior to the onset of the crises.[Footnote 17] In April 2011, the IFIs warned that the resurgence of higher food prices was increasing the cost of food imports in LICs, aggravating existing balance-of-payments problems and putting pressure on government budgets. Moreover, in July 2011, the UN World Food Program declared a food crisis in eastern Africa due mainly to low domestic harvests resulting from consecutive droughts.[Footnote 18] Trade and Fiscal Deficits Worsened: The average current account (trade) deficit-to-GDP ratio for the 38 LICs increased from 3.6 percent in 2007, to 5.4 percent in 2008, and decreased to 4.2 percent in 2009, as shown in figure 7.[Footnote 19] Twenty-eight countries experienced a widening of their current account deficit-to-GDP ratio during the crises period compared to the pre- crises period. Figure 7: Current Account Deficit as a Percent of GDP for 38 LICs, 2004 through 2010: [Refer to PDF for image: line graph] Crisis period: mid-2006 through mid-2009. Year: 2004; Percent of GDP: -2.80. Year: 2005; Percent of GDP: -3.44. Year: 2006; Percent of GDP: -2.74. Year: 2007; Percent of GDP: -3.58. Year: 2008; Percent of GDP: -5.42. Year: 2009; Percent of GDP: -4.16. Year: 2010; Percent of GDP: -4.66. Source: GAO analysis of World Economic Outlook data as of April, 2011. Note: The weighted average is an aggregate reflecting the relative size of countries in terms of their share in total GDP of the entire group. The derivation of these weights requires that GDP in national currency terms be converted to a common currency (in practice, the U.S. dollar). Since 1993, exchange rates based on purchasing power parities (PPP) have been used for this purpose. PPP-based GDP takes differences in price levels across countries into account to ensures that the GDP weights reflect each country's share in real output. PPP GDP weighted average for 38 LICs. [End of figure] Moreover, the fiscal deficit-to-GDP ratio increased from an average of 2.6 percent in 2007 to 3.7 percent in 2009 for 37 LICs, attributable more to rising expenditures than declining revenues. The fiscal deficit-to-GDP ratio for the LICs averaged 1.8 percent between 2004 and 2006 and nearly doubled to 3.2 percent between 2007 and 2009. The IMF reported that most LICs adopted a countercyclical fiscal response, such as preserving or expanding spending to support the economy and protect the poor.[Footnote 20] The growth rate of real primary expenditures[Footnote 21] accelerated, leading to a widening of the fiscal deficit-to-GDP ratio. The IMF also reported that LICs could increase spending, in part, because they had established sufficiently strong fiscal positions before the crises began. Foreign Direct Investment Declined in 2009: Net foreign direct investment inflows for 39 LICs declined by 17 percent during 2009, ending a generally steady increase since 2000, as shown in figure 8.[Footnote 22] Twenty-three of the 39 countries, or nearly 60 percent, had lower net foreign direct investment inflows in 2009 compared to 2007. Figure 18: Foreign Direct Investment Net Inflows for 39 LICs in Current U.S. Dollars, 2000 through 2009: [Refer to PDF for image: line graph] Crisis period: mid-2006 through mid-2009. Millions of current U.S. dollars: Year: 2000; Foreign direct investment, net inflows: $2,759. Year: 2001; Foreign direct investment, net inflows: $2,842. Year: 2002; Foreign direct investment, net inflows: $3,592. Year: 2003; Foreign direct investment, net inflows: $4,740. Year: 2004; Foreign direct investment, net inflows: $5,440. Year: 2005; Foreign direct investment, net inflows: $5,428. Year: 2006; Foreign direct investment, net inflows: $7,157. Year: 2007; Foreign direct investment, net inflows: $12,812. Year: 2008; Foreign direct investment, net inflows: $14,526. Year: 2009; Foreign direct investment, net inflows: $12,933. Source: GAO analysis of World Development Indicators data. Note: Data not available for 2010. Net inflows is net foreign direct investment in the reporting economy from foreign sources less net foreign direct investment by the reporting economy to the rest of the world. [End of figure] IFIs Increased Funding and Met Many Disbursement Goals, but the Impact of Their Actions on LIC Government Spending Has Been Difficult to Establish: In response to the crises, we found that IDA met its goal to increase the amount of financial assistance and partly met its goal to increase the speed of disbursements to LICs, but the impact of IDA's actions on LIC government spending has been difficult to establish. The IFC increased assistance to LICs but its response was limited by capacity constraints and is difficult to measure. The IMF significantly boosted financial assistance to LICs, but its contribution to LIC government spending increases during the crises has been difficult to establish. IDA Increased Financial Assistance and Met Many Disbursement Goals, but Impact on Government Spending Has Been Difficult to Establish: IDA met its goal to increase the amount of financial assistance and partly met its goal to increase the speed of disbursements to LICs. [Footnote 23] In addition, IDA reported that its crises response initiatives supported LIC government spending but we found that the impact has been difficult to establish. The World Bank Used Regular Lending and Established New Initiatives to Respond to the Crises in LICs: The World Bank responded to crises in LICs through regular IDA lending and by establishing initiatives. The World Bank committed a total of $18.1 billion in IDA funds to LICs during the crises response period between 2008 and 2010 through both regular IDA lending and crisis response initiatives. This represented an increase in new commitments of approximately $5 billion, or 39 percent, as compared to commitments made between 2005 and 2007.[Footnote 24] These resources were part of a fixed 3-year allocation, replenished in 2008 prior to the onset of the global economic crisis.[Footnote 25] This allocation represented an increase of 12.8 percent as compared to the 2005-2008 IDA allocation and was intended in part to help countries achieve the Millennium Development Goals (MDGs).[Footnote 26] To respond to the crises within the context of this fixed set of resources, IDA could choose to shift its priorities toward crisis response, accelerate disbursements of existing IDA funds, or provide additional funds from donors or internal resources. To complement regular IDA lending during the crises, the World Bank established a crisis response framework comprised of five initiatives that committed $12.2 billion in financial assistance to LICs between 2008 and 2010. These commitments included $10.8 billion in existing IDA funds and $1.4 billion in new financial assistance. The five initiatives are: * the Global Food Crisis Response Program (food program), established in May 2008 to help countries reduce the impact of high food prices on the poor by providing rapid financial assistance, policy advice, and social protection services[Footnote 27] such as food stamps and school feeding programs for the most vulnerable;[Footnote 28] * the IDA Fast Track Facility, established in December 2008 to help countries offset the impacts of the financial crisis on governments' budget expenditures, including social and infrastructure programs; [Footnote 29] * the Rapid Social Response Program (social protection program), established in April 2009 to help countries mitigate the impacts of crises by promoting social protection programs through rapid financing of immediate interventions in safety nets[Footnote 30] and other areas, and by improving capacity needed to establish and implement effective safety net systems;[Footnote 31] * the Infrastructure Recovery and Assets Platform (infrastructure program), established in March 2009 to help countries mitigate the impacts of the crises by supporting critical infrastructure investments and new project development and implementation, and; [Footnote 32] * the Pilot Crisis Response Window, established in November 2009 to help reduce the need for countries to make tradeoffs between financing crises response efforts or long-term development programs by providing new financing that was additional to countries' existing IDA funds. [Footnote 33] IDA Increased the Amount of Financial Assistance to LICs by $1.4 Billion: In response to the crises, IDA committed $1.4 billion in new financial assistance to 36 LICs between 2008 and 2010.[Footnote 34] Four of the five initiatives--the food, social protection, and infrastructure programs and the Pilot Crisis Response Window--aimed to increase the amount of financial assistance available to LICs using additional donor contributions and internal World Bank resources.[Footnote 35] We found that IDA committed $1.1 billion in new financial assistance to LICs through the Pilot Crisis Response Window and $288 million in new grant assistance to LICs through the food and social protection programs. The World Bank reported that governments' requests for social protection grants to establish and enhance safety net systems provided by the social protection program significantly exceeded the availability of new resources. By April 2011, LICs had submitted 133 project proposals totaling $161 million against available funding of about $58.5 million. These projects were to establish or enhance social protection activities and safety net systems benefiting the poorest, as well as to improve the data and institutional capacity necessary for effective implementation.[Footnote 36] The World Bank further reported that it has established a permanent Crisis Response Window effective July 2011 that could be used to continue to fund activities supporting both crises response and preparedness in IDA- eligible countries, including LICs.[Footnote 37] Finally, donors did not provide additional funding to the infrastructure initiative, which was originally designed to provide up to $3 billion to help offset the impact of soaring energy prices. The World Bank Intended to Increase the Speed of Financial Assistance to LICs during the Crises; Results Are Mixed: Four of the five IDA initiatives--the Fast Track Facility, food program, social protection program, and Pilot Crisis Response Window-- were designed to increase the disbursement speed of commitments made from existing IDA funds. While disbursement rates are a useful metric for capturing the World Bank's response to the immediate needs of recipient countries through the crises response initiatives, we recognize that there are other less quantifiable considerations for assessing the impacts and effectiveness of development assistance. These initiatives were also designed to increase the speed of project preparation and processing, which occurs prior to project approval. According to the World Bank, preparation time was reduced during the crisis for both investment lending and development policy lending programs. To determine whether the disbursement speed of commitments made through the crises response initiatives had increased, we compared the first year disbursement rates for each initiative to the first year disbursement rate of projects approved from 2008 through 2010 that did not fall under any initiative. The World Bank uses a different methodology to report disbursement rates. We did not use the World Bank's standard disbursement rate methodology because our analysis sought to isolate those activities which were explicitly undertaken in response to the crises.[Footnote 38] These initiatives committed approximately $3.9 billion to 32 LICs from existing IDA funds. Three of four initiatives increased the speed of disbursements. The infrastructure program, which committed $6.9 billion in existing IDA funds in addition to the $3.9 billion, did not have a goal to increase the speed of disbursements. More specifically: * The first year disbursement rate was 69.1 percent for the Fast Track Facility, compared to a first year disbursement rate of 33.5 percent for projects that were not funded through an initiative. * The first year disbursement rate was 64.5 percent for the food program, compared to a first year disbursement rate of 33.5 percent for projects that were not funded through an initiative. However, we found that almost half of the commitments made through existing IDA funds, about $405 million, went to three projects in Ethiopia and Bangladesh. When these projects are excluded from the analysis, the disbursement rate for this initiative declines to 39 percent. * The first year disbursement rate for the social protection program was 34.1 percent, slightly higher than the first year disbursement rate of 33.5 percent for projects that were not funded through an initiative. However, the social protection program did not increase the speed of disbursements when compared to social protection projects approved during the pre-crises period, which had a first year disbursement rate of 47.8 percent.[Footnote 39] According to the World Bank, increasing the speed of disbursements for social protection programs in LICs has been challenging due to a lack of existing social protection programs and recipient countries' capacity to effectively implement them.[Footnote 40] To address this challenge and facilitate crises preparedness in LICs, the World Bank intends to continue to finance the development of social protection programs. In early 2011 donors emphasized the importance of a continued focus on capacity building and improved data collection in LICs, to help overcome these constraints. * The Pilot Crisis Response Window did not increase the disbursement speed of commitments, with a first year disbursement rate of 27.5 percent, compared to a first year disbursement rate of 33.5 percent for projects that were not funded through an initiative (see figure 9). Figure 9: First Year Disbursement Rates of Initiative Lending and Non- Initiative Lending, Ranked from Fastest to Slowest: [Refer to PDF for image: vertical bar graph] Crisis response initiative: Fast track facility; Percentage of funds disbursed: 69.1%. Crisis response initiative: Food program; Percentage of funds disbursed: 64.5%. Crisis response initiative: Social protection program; Percentage of funds disbursed: 34.1%. Non-initiative lending: 33.5%. Crisis response initiative: Pilot crisis response window; Percentage of funds disbursed: 27.4%. Source: GAO analysis of World Bank data. Note: We determined the first year disbursement rates by calculating the total disbursements for each project for the first four quarters following project approval, then taking the ratio of total disbursements to the total committed funds. [End of figure] In 2009, the Development Committee, an advisory group to the World Bank and IMF, also urged the acceleration of the delivery of financial assistance to recipient countries. However, we found that the World Bank did not accelerate disbursements for both investment lending projects or development policy lending projects for the group of LICs on average as a whole even though a majority of countries received disbursements faster during the crises response period. Specifically, the average first year disbursement rate to LICs was 16.8 percent for all investment lending projects approved between 2008 through 2010, as compared to an average first year disbursement rate of 17.3 percent for all investment lending projects approved between 2005 through 2007. Similarly, the average first year disbursement rate to LICs was 90.7 percent for all development policy lending projects approved between 2008 through 2010, as compared to an average first year disbursement rate of 96.9 percent for all development policy lending projects approved between 2005 through 2007. Overall, the average first year disbursement rate to LICs was 31.1 percent for all projects approved between 2008 and 2010, as compared to an average first year disbursement rate of 39.3 percent for all projects approved between 2005 and 2007, a difference of about 8 percentage points. According to U.S. Treasury, this decline in part reflects the World Bank's need to ensure that recipient country capacity and governance controls were sufficiently robust to absorb the additional resources provided during the crisis period. However, at the individual country level, total commitments in 22 of 36 LICs, including commitments made to both investment lending projects and development policy lending projects, were disbursed faster during the crises period than during the pre-crises period. [Footnote 41] Disbursement rates, which vary over time, depend on a number of factors, including recipient country capacity, need, and governance, and the type of lending. For example, commitments to Burundi, a fragile state with limited capacity, increased by 114 percent while disbursements increased by 46 percent, which results in a lower disbursement rate during the crises response period. (See figure10.) LIC governments reported mixed experiences relating to the timeliness of the World Bank's response to crises. Some governments said they received financial support very rapidly, while others noted that World Bank support had been sluggish[Footnote 42]. For a more detailed analysis of World Bank commitments and disbursement rates to individual countries, see appendix III. Figure 10: Percentage Point Change in First Year Disbursement Rates from the Pre-Crises Period to the Crises Response Period: [Refer to PDF for image: horizontal bar graph] Disbursed slower during the crises period than pre-crises: Country: Niger; Change in disbursement rate: -59.38%. Country: Central African; Change in disbursement rate: -54.70%. Country: Liberia; Change in disbursement rate: -51.43%. Country: Bangladesh; Change in disbursement rate: -33.36%. Country: Ghana; Change in disbursement rate: -20.22%. Country: Uganda; Change in disbursement rate: -19.42%. Country: Democratic Republic of Congo; Change in disbursement rate: -17.3864 Country: Cambodia; Change in disbursement rate: -13.99%. Country: Burundi; Change in disbursement rate: -11.77%. Country: Afghanistan; Change in disbursement rate: -10.83%. Country: Madagascar; Change in disbursement rate: -6.47%. Country: Tanzania; Change in disbursement rate: -5.43%. Country: Rwanda; Change in disbursement rate: -0.53%. Country: Haiti; Change in disbursement rate: -0.05%. Disbursed faster during the crises period than pre-crises: Country: Togo; Change in disbursement rate: 75.13%. Country: Guinea; Change in disbursement rate: 61.66%. Country: Guinea-Bissau; Change in disbursement rate: 54.95%. Country: Tajikistan; Change in disbursement rate: 46.00%. Country: Gambia, The; Change in disbursement rate: 24.51%. Country: Sierra Leone; Change in disbursement rate: 19.46%. Country: Kyrgyz Republic; Change in disbursement rate: 19.27%. Country: Mali; Change in disbursement rate: 16.94%. Country: Chad; Change in disbursement rate: 16.11%. Country: Ethiopia; Change in disbursement rate: 11.42%. Country: Burkina Faso; Change in disbursement rate: 10.08%. Country: Malawi; Change in disbursement rate: 9.70%. Country: Laos; Change in disbursement rate: 8.88%. Country: Mozambique; Change in disbursement rate: 6.92%. Country: Mauritania; Change in disbursement rate: 5.73%. Country: Nepal; Change in disbursement rate: 3.97%. Country: Zambia; Change in disbursement rate: 3.94%. Country: Kenya; Change in disbursement rate: 2.77%. Country: Eritrea; Change in disbursement rate: 2.76%. Country: Comoros; Change in disbursement rate: 1.55%. Country: Benin; Change in disbursement rate: 0.43%. Country: Solomon Islands; Change in disbursement rate: 0.39%. Source: GAO analysis of World Bank project data. Note: Pre-crises period is between 2005 and 2007; crises period is between 2008 and 2010. Disbursement rates, which vary over time, depend on a number of factors, including recipient country capacity, need, and governance, and the type of lending. For example, commitments to Burundi, a fragile state with limited capacity, increased by 114 percent while disbursements increased by 46 percent, which results in a lower disbursement rate during the crises response period. [End of figure] The World Bank reported that speed was facilitated by the Bank's new rapid response policy and increased use of development policy lending where circumstances permitted. In addition, World Bank and U.S. Treasury officials reported that the restructuring of existing lending portfolios facilitated an expedient response in some countries. For Two Initiatives, the Impact of World Bank Actions on Government Spending Has Been Difficult to Establish: Two initiatives were designed to support domestic spending in recipient countries during the crises, in areas including social services, education, and infrastructure. In 2010, the World Bank reported that both initiatives--the Fast Track Facility and the infrastructure program--met this goal. For example, the World Bank reported that the Fast Track Facility operation in the Democratic Republic of Congo prevented the government from having to cut essential social spending or resort to inflationary spending. Similarly, the World Bank reported that the infrastructure program supported domestic spending and helped to mitigate the direct impacts of the crisis. However, as we previously reported, IFIs do not independently track developing countries' poverty-reducing expenditures and instead rely upon developing countries' governments to provide such data, even though the accuracy of these data and country capacity to provide this information is questionable.[Footnote 43] Additionally, for the infrastructure program, the World Bank developed a rapid diagnostic tool to identify at-risk countries and provide a detailed assessment of crises impacts and associated country infrastructure spending needs, but conducted the diagnostic in only one LIC, Bangladesh.[Footnote 44] Therefore, we found that the degree to which World Bank actions impacted government spending has been difficult to establish. IFC Increased Assistance to LICs, but Its Response Was Limited by Capacity Constraints: The IFC responded to the food and fuel crises through lending in the agriculture and energy sectors and responded to the financial crisis through existing and new initiatives and by enhancing coordination with donors, but its response was limited by capacity constraints. Between 2008 and 2010, IFC increased its new lending commitments in LICs while new IFC commitments overall declined and foreign direct investment in LICs declined by 17 percent between 2008 and 2009. As a result, IFC's investments during the crises in LICs increased as a percentage of net foreign direct investment. Annual IFC commitments in LICs in fiscal years 2009 and 2010 exceeded $900 million each year, while commitments in fiscal year 2008 were about $460 million. IFC also committed a total of $1.1 billion to LICs through two crisis response initiatives,[Footnote 45] the pre-existing Global Trade Finance Program and the new Global Trade Liquidity Program, which supported $3.4 billion in trade through credit guarantees and risk sharing.[Footnote 46] According to officials, IFC also developed new approaches for coordinating with other multilateral institutions in LICs at the regional level. The Joint Action Plan for Africa, established in 2009, for example, developed a method for IFC to collaborate more closely with other lenders in support of development activities in Africa. Similarly, officials said that agreements with donors, made to enhance the response to the financial crisis, will allow IFC to quickly coordinate with donors in response to a future crisis. According to IFC officials, IFC's response was limited by internal and external constraints. Internally, under its Articles of Agreement IFC must undertake its financing on terms and conditions which it considers appropriate, taking into account, among other things, the terms and conditions normally obtained by private investors for similar financing. In addition, IFC has relatively limited resources as compared to other IFIs.[Footnote 47] Officials told us that because of these constraints, much of its crisis response relied on donor governments to provide additional funds. In some cases, this dependence negatively affected the speed of IFC's response because IFC could not respond until donor governments fulfilled their commitments. IFC also faced external capacity constraints in recipient countries. For example, officials explained that in Ethiopia, foreign investors, including IFC, are often subject to additional scrutiny by the government, which has limited IFC's ability to do business there. Overall, IFC officials said that the actions they took to respond to the crises sent a positive signal to the market, but officials noted that this is difficult to measure and did not provide quantitative evidence of this effect. IMF Significantly Boosted Financial Assistance to LICs, but Its Contribution to Government Spending Increases Has Been Difficult to Establish: During the crises response period between 2008 and 2010, the IMF response included committing approximately $4.9 billion in new lending to 28 LICs, temporarily lowering interest rates on its loans, and doubling the limit individual countries could borrow against.[Footnote 48] In addition, the IMF provided $250 billion to support all of its members, including LICs, which collectively received the equivalent of $5.8 billion.[Footnote 49] Governments could use these funds to boost international reserves, cushion against shocks, or meet balance-of- payment needs. Moreover, the IMF changed its lending instruments to address crisis impacts, aiming to make them more flexible and tailored to specific country needs. For example, according to the IMF, the newly created Rapid Credit Facility provides low-access, rapid, and below-market-rate financial assistance to LICs facing an urgent balance-of-payments need, without requiring program-based conditions. According to IMF officials, these efforts were supplemented by technical support and surveillance activities, which also played a role in assisting LICs through the crises. In addition, the IMF reported that its policy advice and programs in LICs were supportive of a countercyclical policy response and higher government spending during the crisis. For example, according to the IMF, government spending increased in almost 90 percent of LICs with IMF programs in 2009.[Footnote 50] We found that IMF loans to LICs increased more than sixfold from approximately $748 million between 2005 and 2007 to about $4.9 billion between 2008 and 2010. In the three countries we reviewed--Burundi, Ethiopia, and Tanzania--the IMF reported that country-specific program goals were achieved. However, circumstances may change quickly, and in one case, inflation resurged soon after the program ended. While conclusions from this sample are not generalizable to all LICs, these examples illustrate how IMF-supported programs operated in these three countries. * In July 2008, Burundi started a 3-year $76 million arrangement with goals to support poverty reduction and macroeconomic stability. Approximately 3 years later, a June 2011 IMF review stated that performance under the program has been broadly satisfactory, despite the impact of the food and fuel shocks. At the same time, however, the IMF also lowered Burundi's 2011 economic growth projection to 4.2 percent, due in part to the expectation that higher food and fuel prices will continue. * Ethiopia requested a $240 million arrangement in August 2009 to help steer the economy through the global financial crisis, with the goals of reducing inflation and building international reserves. In October 2010, an IMF review concluded that the program was on track and that government policies to reduce inflation and increase reserves had been successfully implemented. According to IMF officials, the program ended in November 2010 and inflation rose to about 30 percent in May 2011, mainly because the government did not implement agreed-to reforms. * Tanzania began a $328 million 12-month arrangement in May 2009 with the goal of mitigating the adverse impact of the global financial crisis and addressing a projected deterioration in balance of payments stemming from a decline in exports and foreign direct investment. The IMF's subsequent review determined that country program goals were met, and the program was concluded on schedule a year after inception. The IMF also noted that LICs supported by an IMF program increased government spending during the crises, including on health and education services, more than countries that were not supported by an IMF program, and implied that this was attributable to the IMF programs "as Fund financing reduced liquidity constraints and helped catalyze donors' support."[Footnote 51] However, we found that this causal link has been difficult to establish because the comparison groups differ in important ways. In order to conclude that differences in government spending are driven by IMF programs, the groups of countries being compared need to be as similar as possible. Our analysis of the data underlying the IMF's assertion found that non- program LICs consistently differed from program LICs across certain measures of institutional quality and macroeconomic policy. Furthermore, the finding that program LICs increased spending more than non-program LICs is highly sensitive to the inclusion of a few countries in the group of non-program LICs,[Footnote 52] which either did not need a program or could not obtain one. Non-program LICs had lower scores on a variety of measures of institutional quality, such as political stability, government effectiveness, and rule of law.[Footnote 53] Countries with institutional weaknesses in these categories may overlap with the fragile states, as shown in figure 2. In addition, non-program LICs had higher inflation rates and larger budget deficits prior to the crises than LICs with IMF-supported programs, which may indicate that non-program countries had less capacity to use fiscal and monetary policy to respond to the crises. We conducted several analyses to determine if the IMF's results were sensitive to small changes in the underlying sample of countries designed to make the set of non-program LICs more like LICs with IMF programs.[Footnote 54] In one analysis, we omitted the countries with the two lowest scores on political stability in 2009 (Sudan and Yemen) from the set of non-program LICs, thereby making the sample more similar to program LICs. Based on the new sample, program LICs no longer increased spending more than non-program LICs. In another analysis, we omitted the countries with the two largest budget deficits prior to the crisis (Eritrea and Guyana) from the set of non- program LICs and similarly found that program LICs no longer increased spending more than non-program LICs. Importantly, it does not necessarily follow from these sensitivity analyses that IMF-supported programs were ineffective at increasing spending. A reasonable estimate of what might have happened in the absence of an IMF- supported program is necessary to assess the impact of programs on spending, which the IMF analysis implicitly assumes is the experience of non-program LICs. A more rigorous analytical approach would be needed to conclude that IMF-supported programs resulted in increased government spending during the crises. Analytical approaches that systematically account for differences between program and non-program countries would be necessary to credibly conclude whether or not an IMF-supported program led to greater public spending.[Footnote 55] IFIs Did Not Lower Debt Distress Ratings Due to the Crises, but If Underlying Projections Prove Too Optimistic Then Grants Could Be Increased: The IMF and World Bank prepare annual debt distress ratings, which assess countries' ability to repay their debt.[Footnote 56] The IMF and World Bank did not lower any LICs' debt distress rating as a result of the food, fuel, and financial crises. However, we found that some of the underlying macroeconomic projections might prove too optimistic based on current risks to the global economic recovery and rising commodity prices, as well as on our review of the debt sustainability analysis (DSAs) for three countries. If these projections ultimately prove too optimistic and countries' ability to repay their debt declines significantly, some multilateral institutions could subsequently choose to provide more grants than loans to help lower the risk of debt problems reemerging. Our review of these DSAs is nongeneralizable and meant to be illustrative, not representative. Debt Distress Ratings Assess Countries' Ability to Repay Debt: The debt distress rating is the IMF and World Bank's assessment of the risk that a country will not be able to repay its future debt. In assessing risk and determining a sustainable debt level, the DSA considers the strength of the country's policies and institutions based on the World Bank's Country Policy and Institutional Assessment index (CPIA). The index classifies LICs as weak, medium, or strong performers, with debt burden thresholds associated with each performance category, as shown in figure 11.[Footnote 57] For example, Ethiopia is in the "medium" performer category, which means that its performance will be judged against the debt burden threshold indicators for that category. The threshold indicator for the present value debt-to-export ratio is 150 percent.[Footnote 58] Exports are an important source of funding for repaying debt. The IMF and World Bank have determined that debt-to-export levels in excess of 150 percent put LICs ability to repay debt at risk. According to Ethiopia's 2010 DSA, Ethiopia's debt-to-export ratio is projected to reach a high of 133 percent in 2011 and then decline. Burundi, which is classified as a "weak" performer, faces a lower, more constraining threshold of 100 percent. According to its 2010 DSA, Burundi's debt-to-export ratio exceeded this limit throughout the projection period by a wide margin. For example, from 2011 through 2013, that ratio was projected to be at or above 200 percent. The IMF and World Bank use the extent and duration of the threshold breaches to determine the country's debt distress risk rating, as discussed below. Figure 11: Performance Ratings and Associated Debt Burden Threshold Indicators: [Refer to PDF for image: illustrated table] Performance Rating: Strong Performer (CPIA greater than or equal to 3.75); Present value of debt as a percentage of: Exports: 200%; GDP: 50%; Revenue: 300%; Present value of debt service as a percentage of: Exports: 25%; Revenue: 35%. Performance Rating: Medium Performer (CPIA greater than 3.25; less than 3.75); Present value of debt as a percentage of: Exports: 150%; GDP: 40%; Revenue: 250%; Present value of debt service as a percentage of: Exports: 20%; Revenue: 30%. Performance Rating: Weak Performer (CPIA less than or equal to 3.25); Present value of debt as a percentage of: Exports: 100%; GDP: 30%; Revenue: 200%; Present value of debt service as a percentage of: Exports: 15%; Revenue: 25%. Source: GAO analysis of World Bank and IMF documents. Note: According to the IMF, since loans to LICs vary considerably in their interest rates and length of repayment, the framework focuses on the present value of debt obligations to ensure comparability over time and across countries. [End of figure] The assessment of the country's risk of debt distress--meaning the country cannot service its debt without resorting to exceptional finance (such as debt relief) or a major correction in balancing its income and expenditures--depends on how the country's debt indicators compare with these debt burden threshold indicators under the DSA's "baseline" scenario, as well as under alternative scenarios and stress tests.[Footnote 59] The baseline is the main macroeconomic scenario which describes the evolution of the debt and the macroeconomic variables based on realistic assumptions and projections of key macroeconomic variables such as GDP, inflation, exports, imports, and government revenues. Countries are classified into four categories-- low, moderate, high risk, or in debt distress--according to their likelihood of debt distress, based on the extent and duration of breaches in their threshold indicators.[Footnote 60] (See figure 12.) Debt burden thresholds are not rigid ceilings, and, according to the IMF and World Bank, the debt distress rating seeks to strike a balance between a mechanistic use of the categories and a judgmental approach. Countries classified as "in debt distress" or "high risk of debt distress" receive 100 percent grant financing from IDA, while countries at moderate risk receive 50 percent grants and 50 percent concessional loans, and countries at low risk continue to receive 100 percent concessional loan financing. As shown in figure 12, 13 of the LICs are "in" or at "high" risk of debt distress and 24 are either at "moderate" or "low" risk. Figure 12: Risk of Debt Distress and Performance Ratings for 37 LICs, as of December 2010: [Refer to PDF for image: illustrated table] Performer Rating: Weak Performer; Risk of Debt Distress and Financing Type: In Debt Distress (All grants): Comoros, Eritrea, Guinea, Zimbabwe; High (All grants): Afghanistan, Burundi, Democratic Republic of Congo, Guinea-Bissau, Haiti, Laos, Tajikistan, The Gambia; Moderate (50 percent grants and 50 percent concessional loans): Cambodia, Central African Republic, Chad, Sierra Leone, Solomon Islands, Togo; Low (All concessional loans): Liberia. Performer Rating: Medium Performer; Risk of Debt Distress and Financing Type: In Debt Distress (All grants): None; High (All grants): Burkina Faso; Moderate (50 percent grants and 50 percent concessional loans): Benin, Kyrgyzstan, Malawi, Mauritania, Nepal, Rwanda; Low (All concessional loans): Bangladesh, Ethiopia, Kenya, Madagascar, Mali, Mozambique, Niger, Zambia. Performer Rating: Strong Performer; Risk of Debt Distress and Financing Type: In Debt Distress (All grants): None; High (All grants): None; Moderate (50 percent grants and 50 percent concessional loans): Ghana; Low (All concessional loans): Tanzania, Uganda. Source: GAO analysis of World Bank and IMF DSAs. Notes: Thirty-seven of the 40 LICs are included. Burma, North Korea, and Somalia are not included because they do not have public DSAs. [End of figure] The DSA uses a 3-year moving average CPIA score in determining a country's policy performance in order to reduce variations in the risk of debt distress rating stemming from small annual fluctuations in the CPIA that do not represent a material change in countries' capacity to service their debt. If following the release of the new annual CPIA score, the updated 3-year moving average CPIA rating breaches the applicable CPIA boundary, the country's performance category would change only if the size of the breach exceeds 0.05; if below 0.05, the country's performance category would change only if the breach is sustained for 2 consecutive years. IMF and World Bank Did Not Lower Debt Distress Ratings Due to Crises: Although the crises adversely impacted LICs' economies, the IMF and World Bank did not change any country's debt distress rating as a result of the crises, indicating that they did not expect the crises to adversely impact LIC economies enough to significantly impair their ability to repay debt. The IMF forecasted a rebound in LIC growth in line with the forecast of a quick recovery for the global economy. The global recovery, which the IMF subsequently reported is subject to risks, is expected to boost demand for LIC exports and improve access to foreign capital, both of which are expected to facilitate private sector growth. According to the IMF, the LICs' period of growth prior to the crises provided a cushion, helping countries weather the food and fuel price increases between 2007 and 2008 and the global financial crises. As a result, LICs were able to implement countercyclical policies, such as preserving or expanding spending to support the economy and protect the poor, and expanding public investment. For reasons other than the crises, the IMF and World Bank changed 10 LICs' debt distress ratings from 2007 through 2010. (See figure 13.) Nine ratings improved for the following reasons: * Six countries received full and irrevocable debt relief from government and multilateral creditors.[Footnote 61] * Two countries, Chad and Niger, had higher projected GDP growth stemming from growth in mineral sectors. Chad achieved higher GDP growth due to oil sector growth. Niger is implementing large uranium and oil projects, which are expected to boost exports and government revenues significantly. * Ethiopia's rating changed due to the inclusion of workers' remittances as an important source of debt service financing and resilience of the Ethiopian economy to the global economic crisis. Only one rating worsened. Burkina Faso's debt distress rating changed from moderate in 2007 to high in 2008 because the country was reclassified from a strong to medium performer and therefore exceeded the new, lower debt burden indicators. The implication of a change in a country's debt distress rating is that a country whose rating improves will generally receive a larger proportion of concessional loans, and if it worsens, it will receive a greater proportion of grants, as shown in figure 13.[Footnote 62] Figure 13: Changes in 10 Countries' Debt Distress Ratings and Financing Terms, 2007 through 2010: [Refer to PDF for image: illustrated table] Country: Burkina Faso; 2007: Moderate risk of debt distress; 2008: High risk of debt distress; 2009: High risk of debt distress; 2010: High risk of debt distress. Country: Burundi; 2007: N/A (years when the country did not have a DSA); 2008: In debt distress; 2009: High risk of debt distress[A]; 2010: High risk of debt distress. Country: Central African Republic; 2007: In debt distress; 2008: In debt distress; 2009: Moderate risk of debt distress[A]; 2010: Moderate risk of debt distress. Country: Chad; 2007: High risk of debt distress; 2008: N/A (years when the country did not have a DSA); 2009: Moderate risk of debt distress; 2010: Moderate risk of debt distress. Country: Congo, Democratic Republic; 2007: In debt distress; 2008: N/A (years when the country did not have a DSA); 2009: In debt distress[A]; 2010: High risk of debt distress. Country: Ethiopia; 2007: Moderate risk of debt distress; 2008: Moderate risk of debt distress; 2009: Moderate risk of debt distress; 2010: Low risk of debt distress. Country: Guinea Bissau; 2007: In debt distress; 2008: N/A (years when the country did not have a DSA); 2009: In debt distress; 2010: High risk of debt distress[A]. Country: Liberia; 2007: N/A (years when the country did not have a DSA); 2008: In debt distress; 2009: In debt distress; 2010: Low risk of debt distress[A]. Country: Niger; 2007: Moderate risk of debt distress; 2008: N/A (years when the country did not have a DSA); 2009: Moderate risk of debt distress; 2010: Low risk of debt distress. Country: Togo; 2007: In debt distress; 2008: In debt distress; 2009: In debt distress; 2010: Moderate risk of debt distress[A]. [A] Indicates that a country's rating change coincided with the country completing requirements for full and irrevocable debt relief under the Heavily Indebted Poor Countries Initiative. Source: GAO analysis of World Bank and IMF DSAs. [End of figure] For LICs We Reviewed, Ratings Depend on Projections That Might Prove Too Optimistic: According to the IMF and World Bank, the DSA's quality depends to a large extent on the realism of the projections under the baseline scenario. As explained in their policy paper, realistic means that the scenario takes account of a country's growth potential as well as its capacity constraints, including the risk that governments do not implement desired policy reforms.[Footnote 63] Further, historical averages for the key macroeconomic variables for the past 10 years may provide some guidance about the extent of realism in the baseline scenario projections. The IMF has indicated that explicit justification is required if sustainable debt ratios are driven by DSA assumptions that deviate sharply from historical norms. We assessed the realism of the 2010 DSA projections for three countries--Burundi, Ethiopia, and Tanzania--and found that a too optimistic tone potentially prevailed. While conclusions from this sample are not generalizable to all LICs, these examples are illustrative of how DSAs are conducted. For the sample we reviewed, we based our conclusion on our analysis of the divergence between the DSA projections and their historic values,[Footnote 64] as well as on the reasonableness of the DSAs' underlying assumptions that countries (1) would realize growth-enhancing investments; (2) would implement agreed- to reforms, such as tax reforms that would boost government revenues; and (3) would not be subject to adverse country-specific factors, such as recurring droughts, floods, and political instability. For the three countries we reviewed, we found that macroeconomic projections did not adequately consider the country's vulnerabilities, such as failure to implement reforms, inability to make planned investments, or recurrence of adverse weather or political instability. Burundi's DSA Projections: The 2010 DSA for Burundi projects that real GDP growth rate will increase from its 10-year historical average of 2.7 percent to an average of 4.7 percent over the medium-term. This projected strong GDP growth depends on several factors, including an increase in anticipated export earnings from privatization of the coffee sector, which accounts for about two-thirds of total exports, and integration into the East Africa Community, which could give Burundi access to a broad market of about 120 million people and attract more investment. However, Burundi might not meet the GDP projection if it does not realize the higher export earnings from reform of the coffee sector. Privatization of the coffee sector is occurring more slowly than expected, with only 13 of 117 coffee processing facilities sold because there were few interested buyers. While the government plans to accelerate the sale of the remaining stations beginning in 2011, it is not clear whether investors will buy them. In June 2011, the IMF lowered Burundi's 2011 growth projection from 4.5 percent to 4.2 percent, noting that higher food and fuel prices were likely to continue to increase throughout the year. The IMF reported that risks to Burundi's macroeconomic outlook are significant and include higher food and fuel prices and a worsening of the political, social, and security situation, which would endanger donor support and could further worsen debt indicators.[Footnote 65] Nonetheless, the 2010 DSA assumes Burundi's security and political situation will continue to improve. Moreover, Burundi's 2010 DSA projected a large decrease in the fiscal deficit-to-GDP ratio, from a 5 percent average during 2007 through 2009, to 1.3 percent in 2015, based on a widening of the tax base as a result of continued tax reforms as well as reductions in spending. The IMF said that mobilizing domestic revenue is critical for Burundi's fiscal sustainability. However, the fiscal deficit projections assume that Burundi will control government wages and reduce defense and security spending. Ethiopia's DSA Projections: In 2010, the IMF and World Bank changed Ethiopia's risk of debt distress from "moderate" to "low" based on the inclusion of workers' remittances, which means that Ethiopia now receives 100 percent concessional loan financing instead of 50 percent grants and 50 percent concessional loans.[Footnote 66] In addition, Ethiopia's 2010 DSA projected that Ethiopia would achieve strong export growth and implement key reforms. Ethiopia's 2010 DSA projected exports as a percent of GDP to rise to 19.1 percent by 2015, compared to the 10- year average of 13.5 percent, and to further increase to a 31 percent average during 2016 through 2030. However, in April 2011 IMF staff expressed concern that Ethiopia's failure to implement monetary reforms, including removing the government-imposed bank credit ceilings, as well as highly negative real interest rates, were hindering the commercial banks' financing role, which is fundamental to higher growth.[Footnote 67] In May 2011, IMF staff lowered estimates for Ethiopia's real GDP growth rate from the 7.7 percent forecasted in Ethopia's 2011 DSA to 6 percent for 2011 through 2012 due to high inflation, restrictions on private bank lending, and a more difficult business environment. The projected large increase in growth in the 2010 DSA depended on anticipated growth in service exports resulting from an expected increase in electricity exports based on current energy investments, greater investment in the national airline, and continued good harvests supporting agriculture. The DSA notes Ethiopia's debt profile is very sensitive to export growth assumptions. The inclusion of workers' remittances as a source of debt repayment was a main reason for the improvement in Ethiopia's debt distress rating in 2010. While the DSA projects workers' remittances to remain large and stable at 8.5 percent of GDP, it did not provide historical data or additional information upon which to base this conclusion. [Footnote 68] Moreover, IMF staff reported that Ethiopia's risk of external shocks, such as droughts and high international commodity prices, is high. Ethiopia depends on rain-fed agriculture, which accounts for nearly half of GDP and 85 percent of employment. For the last 30 years, Ethiopia has been hit by droughts every 5 to 7 years, as well as frequent increases in international prices. However, staff told us that country-specific factors, such as weather-related shocks, were not specifically incorporated in the baseline scenario as such. [Footnote 69] Tanzania's DSA Projections: Tanzania's 2010 DSA projects higher export and GDP growth. The 2010 DSA projects a 6.5 percent real GDP growth rate in 2011, rising to 7.5 percent in 2015. The growth in real GDP is based on expected returns from the increase in infrastructure investment, including a rise in agricultural productivity and improved food distribution through investment in rural roads and markets, which is to be financed by additional domestic and external borrowing on less concessional terms. Also, Tanzania's DSA projects an increase in the export-to-GDP ratio from 24.1 percent in the medium term to 28.5 percent in the long-term based on the country's potential to substantially increase commodity and manufacturing exports. Following discussions with Tanzanian government officials, in Tanzania's 2011 DSA the IMF projected a slowdown in real GDP growth for fiscal year 2011/12 (July through June), from the earlier projected rate of 7.1 percent to 6.6 percent. This revision was based on adverse weather, rising fuel prices, and lagging investment. The poor rainfall disrupted electricity generation and lagging investment coupled with higher demand for electricity led to power rationing, which adversely impacted growth. The rising cost of fuel increased the replacement cost of power generation. In addition, the ongoing drought could adversely affect the 2011 food harvest. If Projections Prove Too Optimistic, Creditors Could Provide More Grants: To the extent that DSA projections prove too optimistic, debt problems may reemerge. This could become evident in the projections in future DSAs, and, if the deviations from the prior projections are significant enough, the country's debt distress rating could change, meaning the country could receive more grants than loans. IMF and World Bank staff advise that the quality of the DSA hinges critically on the projections and assumptions underlying the baseline scenario, since alternative assumptions can lead to substantially different debt dynamics. The causes of optimism in the DSAs could be at the global macroeconomic level as well as at the country level. According to the IMF, there are increased risks to global economic recovery, including slower growth and extreme volatility in commodity prices. At the country level, our assessment of three countries' DSAs illustrates how projections and assumptions can change over a relatively short period of time, potentially affecting a country's risk of debt distress. For example, Burundi's present value of debt-to-exports ratio already exceeds the country-specific threshold by a wide margin throughout the projection period. Burundi faces challenges in generating higher government revenue through tax reform and increasing export earnings due to slower than anticipated coffee sector reforms. Lower fiscal revenues or declining GDP growth would lead to a considerable deterioration of its debt ratios, according to its 2010 DSA. This could lead to Burundi being classified as "in debt distress." However, since Burundi is already to receive only grants from creditors, options to assist the country financially might be limited. Similarly, if Ethiopia does not achieve the export growth projections in its DSAs, its debt ratios and performance ratings could worsen, and, if significant enough, could lead to a worsening of its debt distress rating. If this occurs, the terms of Ethiopia's financing from certain lenders could change. Ethiopia now receives its financial assistance as all concessional loans, but a change in risk rating could lead to financing with a larger grant component. Regarding Tanzania, IMF staff reported in 2011 that revenue collection had fallen short of ambitious targets, and that the rapidly increasing fiscal deficit was being financed by increasingly expensive resources due to a shift from mostly grants to loans. Though Tanzania's current risk of debt distress is low, maintaining current spending policies could widen the fiscal deficit, leading to rising debt servicing costs, with an adverse impact on the debt indicators. In May 2011, IMF reported that Tanzania faces formidable challenges given widespread poverty, high population growth, and tremendous dependence on foreign aid, and the near-term economic outlook is subject to considerable uncertainty with a rising risk of donor aid shortfalls and higher international fuel prices. Moreover, the effectiveness of the debt distress rating and associated financing requirements in keeping countries' debt at sustainable levels depends on their broader use by borrowers and creditors. Some other multilateral institutions--including the African Development Bank, Asian Development Bank, Inter-American Development Bank and International Fund for Agricultural Development--and developed countries use the debt distress rating system to make decisions about their terms of financing. If projections ultimately prove too optimistic and countries' ability to repay their debt declines significantly, some multilateral institutions could subsequently choose to provide more grants than loans to help lower the risk of debt problems reemerging. Concluding Observations: The food, fuel, and financial crises resulted in slower economic growth, higher deficits, and increased inflation for LICs. While the overall impact of the crises on LICs may have been milder than in the advanced economies, the high rate of poverty in these countries increases their overall vulnerability. For example, our previous work shows that many LICs were experiencing protracted food emergencies and had severe and widespread malnourishment even prior to the onset of the crises. The IFIs responded to the crises by increasing the amount of resources made available to the LICs. The IMF increased lending to LICs more than sixfold to almost $5 billion. The World Bank committed $18.1 billion through regular lending and five new crises response initiatives that committed $12.2 billion in financial assistance to LICs, including $1.4 billion in new funding. The World Bank provided funding as a mix of loans and grants, depending on the performance and debt vulnerability of each country. The World Bank was able to meet its goal of increasing the speed of disbursement for several initiatives, but the overall picture is mixed, especially when compared to the pre-crises period. Furthermore, in the case of both institutions, the impact of these new resources on LIC government spending during the crises has been difficult to establish. According to the World Bank and IMF, the crises did not significantly impair the countries' ability to repay their future debt, because they expected the world economy to reestablish its pre-crises growth levels and the LICs to implement reforms necessary to achieve projected future growth levels. However, the increased risk to the global recovery and the extreme volatility of commodity prices may undermine the realization of these expectations. To date, the mix of loans and grants provided by the multilateral development banks have been largely unaffected by the crises. We found that the projections for our case study countries may prove too optimistic, which could contribute to debt problems re- emerging as the amount of loans countries receive could be greater than what would be considered sustainable. However, given that the IFIs update the DSAs on a regular basis, any excessive optimism should become evident over time, and the World Bank and other lenders could then increase the amount of grants they provide which would help mitigate potential debt problems. Agency Comments and Our Evaluation: The U.S. Treasury, World Bank, and IMF provided written comments on a draft of this report, which are reprinted in appendixes IV, V, and VI, respectively. The U.S. Treasury commented that the IFIs appropriately responded to the crisis and effectively managed the trade-offs associated with quickly disbursing funds in an environment of limited capacity to absorb aid and that the United States strongly advocated for increased IFI engagement in LICs during the crises. The Treasury letter also stated that we provided a good overview of how the IMF responded forcefully to the crisis. In our discussion on the impact of IMF programs on government spending, they suggested we should have examined the impact of IMF programs on social spending. However, we would emphasize that accounting for the differences between program and non-program countries is critical to estimating the impact of IMF programs on spending during the crisis, which the IMF did not do in their 2010 report. The U.S. Treasury also noted that speed of disbursements is just one measure of effective crisis response and that it is important to consider trade-offs between speed of disbursements and the need to ensure adequate governance structures and fiduciary controls are in place. We included this information in our report. In addition, the Treasury stated that by working closely with the World Bank and other multilateral development banks (MDBs) to put in place the right fiduciary arrangements and strengthen country capacity to absorb and manage MDB assistance, the Treasury can improve the quality of the World Bank's and other MDBs' interventions and improve the monitoring and reporting of development results. The World Bank stated that it welcomes and agrees with our overall conclusion that IFIs, including IDA and IFC, met many goals in response to the crises in LICs. The World Bank also stated that it achieved a significant increase in both its commitments and its disbursements to LICs. We acknowledge that the World Bank responded to crises in LICs by increasing its commitments and disbursements through regular IDA lending and by establishing initiatives. Our analysis focused on the initiatives because these were specifically designed to respond to the crises. Our calculations for the overall commitments and disbursements, as well as the disbursement rates, differed from the World Bank's because our methodology sought to isolate those activities which were explicitly undertaken in response to the crises. The World Bank said that IDA accelerated assistance delivery without compromising attention to governance and aid effectiveness. We acknowledge that disbursement rates, which vary over time, depend on a number of factors, including recipient country capacity, need, and governance, and the type of lending. The World Bank said there is growing evidence that IDA-supported public spending for essential services increased. As we previously reported, IFIs do not independently track developing countries' poverty-reducing expenditures and instead rely upon developing countries' governments to provide such data, even though the accuracy of these data and country capacity to provide this information is questionable. Finally, regarding debt sustainability, the World Bank noted that our analysis is based on a sample of just three countries and thus cannot assess the realism of the 2010 DSA projections. We based our conclusions on our assessment of the realism of the 2010 DSA projections for three countries as well as on the current risks to global economic recovery, reported by the IMF in August 2011. The IMF indicated broad agreement with the findings of our report, including the overview of the impact of the crisis on LICs. While the IMF suggested that our assessment is narrow, we paid sufficient attention to a range of response efforts, mentioning the IMF's call for countercyclical policy responses, improved macroeconomic conditions in LICs, doubling of access levels, and modifications to lending instruments. The IMF acknowledged that comparing program with non-program countries does not prove a causal link from program engagement to higher spending and notes a recent related study. We include a reference to the study entitled "What Happens to Social Spending in IMF-Supported Programs" but also note that it does not necessarily reflect the results during the crises response period. The IMF also stated that growth assumptions underlying LIC DSAs have been borne out so far and that DSAs have built-in methods for addressing risks. We described the IMF's use of alternative scenarios and stress tests to arrive at a country's debt distress rating. However, we noted that these tests are very general and do not adequately reflect country-specific risks including political instability, adverse weather, global economic crises, and failure to implement reforms or make planned investment. Our analysis of the three countries' DSAs is intended to be illustrative and not generalizable. Our conclusion, that projections which might be too optimistic could be mitigated by future DSAs and additional grants, is not dependent on these three countries. As agreed with your offices, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days from the report date. At that time, we will send copies to Members of Congress; U.S. Treasury, the IMF, and the World Bank. In addition, the report will be available at no charge on the GAO Web site at [hyperlink, http://www.gao.gov]. If you or your staffs have any questions about this report, please contact me at (202) 512-9601 or melitot@gao.gov. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. GAO staff who made major contributions to this report are listed in appendix VII. Signed by: Thomas Melito: Director, International Affairs and Trade: [End of section] Appendix I: Objectives, Scope, and Methodology: Our objectives were to examine (1) the economic impact of the crises on low-income countries (LIC), (2) international financial institutions' (IFI) responses and reported results, and (3) IFIs' assessment of the impact of the crises on LICs' ability to repay their debt. Economic Impact of Crises: To examine the impact of the crises on LICs' economic performance, we collected and analyzed key macroeconomic data from 1990 through 2010 for 38 of the 40 LICs for which data were available, except as where noted. We analyzed variables including real gross domestic product (GDP), current account, fiscal deficit, government expenditures and revenue, consumer price index measure of inflation, and foreign direct investment. We obtained these data series from the widely used IMF and World Bank databases--World Economic Outlook, International Financial Statistics, and World Development Indicators. We computed LICs' average economic performance with respect to each of the key economic variables, including current account and fiscal deficits and inflation, using a real GDP weighted average based on purchasing power parity (PPP) GDP, so that the resulting weighted average reflects each country's size in terms of their share in total GDP of the entire group of LICs. We used the PPP GDP weights to construct weighted averages for the other variables, including fiscal and current account deficits, government revenue and expenditures, and the consumer price index measure of inflation. We analyzed the LIC group's macroeconomic performance over the 2007 through 2009 crises period, and compared this to the group's performance during the pre-crisis period from 2004 through 2006 to determine whether economic performance improved or deteriorated. We also disaggregated the results to determine which countries experienced improvements and deteriorations in each of the key macroeconomic variables. We corroborated our results with data from the economic and financial forecasting firm, IHS Global Insight. We also examined and assessed the DSA's incorporation of World Economic Outlook assumptions concerning the global pace of recovery, including those of the country's key trading partners. We also compiled information on international food and oil price data from the UN Food and Agriculture Organization's Food Price Index and the U.S. Department of Energy. Additionally, we reviewed IMF country reports and World Bank Country Assistance Strategies, which also contain limited historical information on the key macroeconomic variables. Some IMF data is based on developing country government data with greatly varying statistical capacity across countries, and we discussed these limitations with IMF officials. We determined that the data were sufficiently reliable for summarizing countries' past macroeconomic performance. Crises Response: To examine IFIs' responses to the crises and reported results, we analyzed documents and data from the World Bank and the IMF. For the World Bank, these documents include proposals and framework documents for the crisis response initiatives; Country Assistance Strategies and Interim Strategy Notes; Project Information Documents and Implementation Completion Reports; and Independent Evaluation Group reports and approach papers. For the IMF, we reviewed IMF country reports; countries' letters of intent; research papers; and the Independent Evaluation Office reports. We also reviewed joint World Bank-IMF publications. We interviewed officials from these institutions, as well as from the Department of State and the U.S. Agency for International Development. To evaluate whether the World Bank's crisis response was consistent with the institution's stated goals of increasing the speed of disbursements, we analyzed World Bank data on financial commitments and disbursements made to LICs between 2005 and 2010. To do this, we determined the first year disbursement rate for all projects approved during the crises response period (2008 through 2010); all projects approved during the pre-crises period (2005 through 2007); and all projects approved during the crises response period under each initiative, as well as those approved during the crises response period outside of any initiative. We did not use the World Bank's standard disbursement rate methodology because our analysis sought to isolate those activities which were explicitly undertaken in response to the crises. For our analysis, this included only projects approved after the World Bank first stated its intention to respond to any of the three crises. This occurred in early 2008, with the establishment of the Global Food Crisis Response Program. We also identified the total number of projects and the amount of funding committed in association with any of the World Bank's crisis response initiatives. To measure the speed of disbursements, we first calculated the total disbursements for each project that took place during the first four quarters following project approval. We then determined the average disbursement rate by using a weighted average, which is computed as the ratio between the sum of first year disbursements and the sum of the commitments. One type of project, "additional financing" projects, tracks disbursements under the "parent" project, though commitment amounts are recorded under the "additional financing" project. In our dataset, there were 160 of these projects, out of a total of 622. To ensure that disbursements from "additional financing" projects were captured in our analysis, we developed a methodology that calculated the remaining commitment balance of the "parent" project as of the quarter in which the additional financing project is approved. We then added the balance to the new commitment from the additional financing project to form the denominator of the disbursement ratio. We then calculated the first year disbursement rate by determining the first four quarters of disbursements under the "parent" project following the approval date of the "additional financing" project. We used disbursement data through June 2011, the latest available, to ensure as many projects as possible (97 percent of all projects) had 4 quarters of disbursements. Eighteen projects approved in the fourth quarter of 2010 had only 3 quarters of disbursement data. We used disbursement data through June 2008 for the pre-crisis period to make our analysis comparable. We then compared various disbursement rates to one another to reach our conclusions. We assessed the reliability of the data we used in our analysis by comparing the consistency of the data among the various sources, and discussing the data with World Bank and IMF officials. For the data used to determine World Bank disbursement rates, we interviewed World Bank officials to understand their database and correct errors in the data. We determined that the data used in our analysis were sufficiently reliable for our purposes. To determine U.S. dollar values associated with the IMF's response to the crises, we used information on IMF program and funding levels. To calculate the portion of the $250 billion in support that the LICs received, we totaled the amount each country received in Special Drawing Rights, the IMF's unit of account, and multiplied that by the August 28, 2009 conversion rate to arrive at a U.S. dollar value. To calculate that IMF loans to LICs increased more than sixfold from approximately $748 million between 2005 and 2007 to about $4.9 billion between 2008 and 2010, we used data from the "IMF Lending Arrangements" online tool.[Footnote 70] For each year, we totaled new lending to LICs in Special Drawing Rights, then converted that total to dollars using the year-end exchange rate. We also assessed the sensitivity of the results of an IMF analysis of government spending in LICs in 2009 using data from the IMF and the Worldwide Governance Indicators. We compared program LICs and non-program LICs using measures of institutional quality from the World Governance Indicators and indicators of macroeconomic policy from the IMF's World Economic Outlook database. We conducted sensitivity analyses by omitting countries with the lowest scores on certain measures of institutional quality or the least favorable pre-crisis macroeconomic policies from the sample of non-program LICs. We assessed the reliability of data used in our sensitivity analyses and found them to be sufficiently reliable for summarizing and ranking countries' institutional quality and macroeconomic policy. Debt Sustainability Analyses: To examine the extent to which IFI's assessments of LICs' ability to repay their debt was impacted by the crises, we reviewed the changes in each of the 40 LIC's Country Policy and Institutional Assessments (CPIA) and debt distress ratings over the period 2007 through 2010 to determine if the crises led to a change in a country's rating. We used CPIA data from the World Bank's online database and the debt distress ratings from each of the LIC's debt sustainability analyses (DSA) over the period. To illustrate how the DSAs are conducted and how macroeconomic projections affect the reliability of the debt distress rating, we focused on three case study countries: Burundi, Ethiopia, and Tanzania. We selected these countries based on criteria that include the number of IFI projects, amount of IFI financial support, and country conditions. For example, between fiscal years 2008 and 2010, Ethiopia and Tanzania were in the top three recipients of IDA assistance by dollar value, receiving approximately $4.8 billion, while Burundi is a post-conflict fragile state. For each of our three case-study countries, we analyzed the country's 2010 World Bank-IMF Bank DSAs, as well as the associated IMF program reviews and Article IV consultations, and World Bank Country Assistance Strategies. We also reviewed prior and subsequent DSAs to make comparisons and check for data consistency. In addition, we met with IMF staff responsible for the DSA preparation for preparing each of the case-study countries. We based our assessment of the DSA's ability to accurately reflect the country's debt vulnerabilities on our analysis of the DSA's macroeconomic projections and the underlying assumptions, which form the basis of the country's risk of debt distress. These include the DSA's projections and assumptions regarding key macroeconomic variables, such as real GDP and export growth, and the divergence of the growth rates of these variables from their historic values; assumptions regarding the LIC's implementation of reforms and productivity-increasing investment; and assumptions concerning the country's vulnerability to external shocks, including adverse weather, political or regional instability, and rising food and fuel prices. To determine economic performance over the study period for our three case study countries, and to calculate historical growth rates for key macroeconomic variables, we relied primarily on IMF and World Bank databases--World Economic Outlook, International Financial Statistics, World Development Indicators, and Balance of Payments Statistics. We also utilized data from IMF Article IV consultations and country program reviews. We compared GAO-calculated 10-year historical averages for the most important macroeconomic variables with the DSA's historical averages for these variables, which form the basis for the DSA's projections and debt ratios. We evaluated the IFIs' determination of a country's risk of debt distress partly based on the extent to which the values of historical key macroeconomic variables diverge from their projected values; and, if so, whether the DSA provides reasonable justification for this divergence. In making this assessment, we also considered additional information available in the Article IVs, World Bank Country Assistance Strategies, and Global Insight Country Intelligence Reports. We also based our assessment of the IFIs' determination of the country's risk of debt distress on the DSA's consideration of country-specific factors, such as the country's susceptibility to weather-related shocks; political instability; and implementation of institutional reforms that would enhance a country's growth prospects, particularly in the economic and debt management areas. We discussed our approach and preliminary findings with officials from the IMF. We assessed the reliability of data used in our country analysis based on the consistency of data across various sources and determined them to be sufficiently reliable to make nongeneralizable assessments of the DSAs for the three case study countries. We conducted this performance audit from September 2010 to September 2011 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. [End of section] Appendix II: Economic, Demographic, and Financial Assistance Data for LICs: Country: Afghanistan; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 30; Life expectancy (2008, years): 44; GNI per capita ($USD, 2008): $310; Percent of population: Rural: 76%; Access to improved water source: 48%; Access to improved sanitation facilities: 37%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = 5,557; IDA: 4%; United States: 46%; All multilateral assistance: 15%; All bilateral assistance: 85%. Country: Bangladesh; Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 162; Life expectancy (2008, years): 66; GNI per capita ($USD, 2008): $520; Percent of population: Rural: 73%; Access to improved water source: 80%; Access to improved sanitation facilities: 53%; Average Annual Official Development Assistance (ODA) (Dollars in millions): $2,650; All donors total = $2,650; IDA: 22%; United States: 4%; All multilateral assistance: 50%; All bilateral assistance: 50%. Country: Benin; Medium Performer; Moderate risk of debt distress; Population (nearest million, 2009): 9; Life expectancy (2008, years): 61; GNI per capita ($USD, 2008): $700; Percent of population: Rural: 59%; Access to improved water source: 75%; Access to improved sanitation facilities: 12%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $680; IDA: 14%; United States: 7%; All multilateral assistance: 52%; All bilateral assistance: 48%. Country: Burkina Faso; Medium Performer; High risk of debt distress; Population (nearest million, 2009): 16; Life expectancy (2008, years): 53; GNI per capita ($USD, 2008): $480; Percent of population: Rural: 80%; Access to improved water source: 76%; Access to improved sanitation facilities: 11%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,085; IDA: 19%; United States: 3%; All multilateral assistance: 54%; All bilateral assistance: 46%. Country: Burma (Myanmar); Performer: No rating; Risk of debt distress: No rating; Population (nearest million, 2009): 50; Life expectancy (2008, years): 62; GNI per capita ($USD, 2008): No data available; Percent of population: Rural: 67%; Access to improved water source: 71%; Access to improved sanitation facilities: 81%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $454; IDA: 0%; United States: 12%; All multilateral assistance: 24%; All bilateral assistance: 76%. Country: Burundi; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 6; Life expectancy (2008, years): 50; GNI per capita ($USD, 2008): $140; Percent of population: Rural: 90%; Access to improved water source: 72%; Access to improved sanitation facilities: 46%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,060; IDA: 35%; United States: 4%; All multilateral assistance: 68%; All bilateral assistance: 32%. Country: Cambodia; Weak Performer; Moderate risk of debt distress; Population (nearest million, 2009): 15; Life expectancy (2008, years): 61; GNI per capita ($USD, 2008): $630; Percent of population: Rural: 78%; Access to improved water source: 61%; Access to improved sanitation facilities: 29%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $755; IDA: 6%; United States: 9%; All multilateral assistance: 37%; All bilateral assistance: 63%. Country: Central African Republic; Weak Performer; Moderate risk of debt distress; Population (nearest million, 2009): 4; Life expectancy (2008, years): 47; GNI per capita ($USD, 2008): $410; Percent of population: Rural: 61%; Access to improved water source: 67%; Access to improved sanitation facilities: 34%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $534; IDA: 39%; United States: 6%; All multilateral assistance: 77%; All bilateral assistance: 23%. Country: Chad; Weak Performer; Moderate risk of debt distress; Population (nearest million, 2009): 11; Life expectancy (2008, years): 49; GNI per capita ($USD, 2008): $540; Percent of population: Rural: 73%; Access to improved water source: 50%; Access to improved sanitation facilities: 9%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $565; IDA: 3%; United States: 22%; All multilateral assistance: 42%; All bilateral assistance: 58%. Country: Comoros; Weak Performer; In debt distress; Population (nearest million, 2009): 1; Life expectancy (2008, years): 65; GNI per capita ($USD, 2008): $750; Percent of population: Rural: 72%; Access to improved water source: 95%; Access to improved sanitation facilities: 36%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $53; IDA: 5%; United States: 0%; All multilateral assistance: 47%; All bilateral assistance: 53%. Country: Democratic Republic of Congo; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 66; Life expectancy (2008, years): 48; GNI per capita ($USD, 2008): $150; Percent of population: Rural: 66%; Access to improved water source: 46%; Access to improved sanitation facilities: 23%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $2,261; IDA: 17%; United States: 10%; All multilateral assistance: 54%; All bilateral assistance: 46%. Country: Eritrea; Weak Performer; In debt distress; Population (nearest million, 2009): 5; Life expectancy (2008, years): 59; GNI per capita ($USD, 2008): $280; Percent of population: Rural: 79%; Access to improved water source: 61%; Access to improved sanitation facilities: 14%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $149; IDA: 14%; United States: 2%; All multilateral assistance: 60%; All bilateral assistance: 40%. Country: Ethiopia; Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 83; Life expectancy (2008, years): 55; GNI per capita ($USD, 2008): $280; Percent of population: Rural: 83%; Access to improved water source: 38%; Access to improved sanitation facilities: 12%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $3,598; IDA: 22%; United States: 21%; All multilateral assistance: 48%; All bilateral assistance: 52%. Country: The Gambia; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 2; Life expectancy (2008, years): 56; GNI per capita ($USD, 2008): $400; Percent of population: Rural: 44%; Access to improved water source: 92%; Access to improved sanitation facilities: 67%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $231; IDA: 52%; United States: 4%; All multilateral assistance: 85%; All bilateral assistance: 15%. Country: Ghana; Strong Performer; Moderate risk of debt distress; Population (nearest million, 2009): 24; Life expectancy (2008, years): 57; GNI per capita ($USD, 2008): $1,150; Percent of population: Rural: 50%; Access to improved water source: 82%; Access to improved sanitation facilities: 13%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,495; IDA: 17%; United States: 8%; All multilateral assistance: 45%; All bilateral assistance: 55%. Country: Guinea; Weak Performer; In debt distress; Population (nearest million, 2009): 10; Life expectancy (2008, years): 58; GNI per capita ($USD, 2008): $340; Percent of population: Rural: 66%; Access to improved water source: 71%; Access to improved sanitation facilities: 19%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $394; IDA: 6%; United States: 13%; All multilateral assistance: 37%; All bilateral assistance: 63%. Country: Guinea-Bissau; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 2; Life expectancy (2008, years): 48; GNI per capita ($USD, 2008): $460; Percent of population: Rural: 70%; Access to improved water source: 61%; Access to improved sanitation facilities: 21%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $155; IDA: 15%; United States: 1%; All multilateral assistance: 66%; All bilateral assistance: 34%. Country: Haiti; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 10; Life expectancy (2008, years): 61; GNI per capita ($USD, 2008): No data available; Percent of population: Rural: 53%; Access to improved water source: 63%; Access to improved sanitation facilities: 17%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,448; IDA: 16%; United States: 20%; All multilateral assistance: 56%; All bilateral assistance: 44%. Country: Kenya; Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 40; Life expectancy (2008, years): 54; GNI per capita ($USD, 2008): $730; Percent of population: Rural: 78%; Access to improved water source: 59%; Access to improved sanitation facilities: 31%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,805; IDA: 10%; United States: 29%; All multilateral assistance: 32%; All bilateral assistance: 68%. Country: Kyrgyzstan; Medium Performer; Moderate risk of debt distress; Population (nearest million, 2009): 5; Life expectancy (2008, years): 67; GNI per capita ($USD, 2008): $790; Percent of population: Rural: 64%; Access to improved water source: 90%; Access to improved sanitation facilities: 93%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $389; IDA: 10%; United States: 15%; All multilateral assistance: 47%; All bilateral assistance: 53%. Country: Laos; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 6; Life expectancy (2008, years): 65; GNI per capita ($USD, 2008): $750; Percent of population: Rural: 69%; Access to improved water source: 57%; Access to improved sanitation facilities: 53%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $508; IDA: 10%; United States: 1%; All multilateral assistance: 38%; All bilateral assistance: 62%. Country: Liberia; Weak Performer; Low risk of debt distress; Population (nearest million, 2009): 4; Life expectancy (2008, years): 58; GNI per capita ($USD, 2008): $170; Percent of population: Rural: 40%; Access to improved water source: 68%; Access to improved sanitation facilities: 17%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,081; IDA: 3%; United States: 25%; All multilateral assistance: 30%; All bilateral assistance: 70%. Country: Madagascar; medium Performer; Low risk of debt distress; Population (nearest million, 2009): 20; Life expectancy (2008, years): 60; GNI per capita ($USD, 2008): $410; Percent of population: Rural: 70%; Access to improved water source: 41%; Access to improved sanitation facilities: 11%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $663; IDA: 19%; United States: 12%; All multilateral assistance: 59%; All bilateral assistance: 41%. Country: Malawi; Medium Performer; Moderate risk of debt distress; Population (nearest million, 2009): 15; Life expectancy (2008, years): 53; GNI per capita ($USD, 2008): $260; Percent of population: Rural: 81%; Access to improved water source: 80%; Access to improved sanitation facilities: 56%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $861; IDA: 8%; United States: 12%; All multilateral assistance: 48%; All bilateral assistance: 52%. Country: Mali Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 13; Life expectancy (2008, years): 48; GNI per capita ($USD, 2008): $610; Percent of population: Rural: 68%; Access to improved water source: 56%; Access to improved sanitation facilities: 36%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,021; IDA: 14%; United States: 8%; All multilateral assistance: 43%; All bilateral assistance: 57%. Country: Mauritania; Medium Performer; Moderate risk of debt distress; Population (nearest million, 2009): 3; Life expectancy (2008, years): 57; GNI per capita ($USD, 2008): $980; Percent of population: Rural: 59%; Access to improved water source: 49%; Access to improved sanitation facilities: 26%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $328; IDA: 13%; United States: 5%; All multilateral assistance: 48%; All bilateral assistance: 52%. Country: Mozambique; Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 23; Life expectancy (2008, years): 48; GNI per capita ($USD, 2008): $380; Percent of population: Rural: 63%; Access to improved water source: 47%; Access to improved sanitation facilities: 17%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $2,039; IDA: 12%; United States: 12%; All multilateral assistance: 35%; All bilateral assistance: 65%. Country: Nepal; Medium Performer; Moderate risk of debt distress; Population (nearest million, 2009): 29; Life expectancy (2008, years): 67; GNI per capita ($USD, 2008): $400; Percent of population: Rural: 83%; Access to improved water source: 88%; Access to improved sanitation facilities: 31%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $949; IDA: 14%; United States: 8%; All multilateral assistance: 43%; All bilateral assistance: 57%. Country: Niger; Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 15; Life expectancy (2008, years): 51; GNI per capita ($USD, 2008): $330; Percent of population: Rural: 83%; Access to improved water source: 48%; Access to improved sanitation facilities: 9%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $556; IDA: 9%; United States: 8%; All multilateral assistance: 51%; All bilateral assistance: 49%. Country: North Korea; Performer: No rating; Risk of debt distress: No rating; Population (nearest million, 2009): 24; Life expectancy (2008, years): 67; GNI per capita ($USD, 2008): No data available; Percent of population: Rural: 37%; Access to improved water source: 00%; Access to improved sanitation facilities: No data available; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $140; IDA: 0%; United States: 55%; All multilateral assistance: 19%; All bilateral assistance: 81%. Country: Rwanda; Medium Performer; Moderate risk of debt distress; Population (nearest million, 2009): 10; Life expectancy (2008, years): 50; GNI per capita ($USD, 2008): $440; Percent of population: Rural: 82%; Access to improved water source: 65%; Access to improved sanitation facilities: 54%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $940; IDA: 14%; United States: 14%; All multilateral assistance: 48%; All bilateral assistance: 52%. Country: Sienna Leone; Weak Performer; Moderate risk of debt distress; Population (nearest million, 2009): 6; Life expectancy (2008, years): 48; GNI per capita ($USD, 2008): $320; Percent of population: Rural: 62%; Access to improved water source: 49%; Access to improved sanitation facilities: 13%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $407; IDA: 11%; United States: 4%; All multilateral assistance: 54%; All bilateral assistance: 46%. Country: Solomon Islands; Weak Performer; Moderate risk of debt distress; Population (nearest million, 2009): 1; Life expectancy (2008, years): 66; GNI per capita ($USD, 2008): $ 1,050; Percent of population: Rural: 82%; Access to improved water source: No data available; Access to improved sanitation facilities: No data available; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $220; IDA: 0%; United States: 0%; All multilateral assistance: 4%; All bilateral assistance: 96%. Country: Somalia; Performer: No rating; Risk of debt distress: No rating; Population (nearest million, 2009): 9; Life expectancy (2008, years): 50; GNI per capita ($USD, 2008): No data available; Percent of population: Rural: 63%; Access to improved water source: 30%; Access to improved sanitation facilities: 23%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $711; IDA: 0%; United States: 31%; All multilateral assistance: 24%; All bilateral assistance: 76%. Country: Tajikistan; Weak Performer; High risk of debt distress; Population (nearest million, 2009): 7; Life expectancy (2008, years): 67; GNI per capita ($USD, 2008): $600; Percent of population: Rural: 74%; Access to improved water source: 70%; Access to improved sanitation facilities: 94%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $386; IDA: 11%; United States: 13%; All multilateral assistance: 60%; All bilateral assistance: 40%. Country: Tanzania; Strong Performer; Low risk of debt distress; Population (nearest million, 2009): 44; Life expectancy (2008, years): 56; GNI per capita ($USD, 2008): $460; Percent of population: Rural: 74%; Access to improved water source: 54%; Access to improved sanitation facilities: 24%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $2,655; IDA: 20%; United States: 10%; All multilateral assistance: 47%; All bilateral assistance: 53%. Country: Togo; Weak Performer; Moderate risk of debt distress; Population (nearest million, 2009): 7; Life expectancy (2008, years): 63; GNI per capita ($USD, 2008): $410; Percent of population: Rural: 58%; Access to improved water source: 60%; Access to improved sanitation facilities: 12%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $503; IDA: 19%; United States: 1%; All multilateral assistance: 45%; All bilateral assistance: 55%. Country: Uganda; Strong Performer; Low risk of debt distress; Population (nearest million, 2009): 33; Life expectancy (2008, years): 53; GNI per capita ($USD, 2008): $420; Percent of population: Rural: 87%; Access to improved water source: 67%; Access to improved sanitation facilities: 48%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,745; IDA: 17%; United States: 21%; All multilateral assistance: 41%; All bilateral assistance: 59%. Country: Zambia; Medium Performer; Low risk of debt distress; Population (nearest million, 2009): 13; Life expectancy (2008, years): 45; GNI per capita ($USD, 2008): $950; Percent of population: Rural: 65%; Access to improved water source: 60%; Access to improved sanitation facilities: 49%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $1,211; IDA: 4%; United States: 19%; All multilateral assistance: 41%; All bilateral assistance: 59%. Country: Zimbabwe; Weak Performer; In debt distress; Population (nearest million, 2009): 13; Life expectancy (2008, years): 44; GNI per capita ($USD, 2008): $320; Percent of population: Rural: 63%; Access to improved water source: 82%; Access to improved sanitation facilities: 44%; Average Annual Official Development Assistance (ODA) (Dollars in millions): All donors total = $676; IDA: 0%; United States: 35%; All multilateral assistance: 15%; All bilateral assistance: 85%. Sources: GAO analysis of World Bank, IMF and OECD-DAC data; Map Resources (map). Note: Average Annual Official Development Assistance (ODA), for which the main objective is economic development and welfare, was calculated using data from 2008 and 2009 and does not include funding for 2010. ODA excludes certain items such as military aid and antiterrorism activities. [End of table] [End of section] Appendix III: World Bank Commitments and Disbursements to LICs through Crisis Response Initiatives: [End of section] Between 2008 and 2010, the World Bank committed $12.2 billion in financial assistance to 38 LICs through five crisis response initiatives, including $10.8 billion from existing International Development Association (IDA) funds, and $1.4 billion from new financial assistance. Figure 14 shows World Bank commitments to 38 LICs between 2008 and 2010. Figure 14: Distribution of World Bank Commitments to 38 LICs, by Country, 2008-2010: [Refer to PDF for image: horizontal bar graph] Country: Ethiopia; Existing IDA funds: $1.964 billion; New funds: $86.85 million. Country: Bangladesh; Existing IDA funds: $1.610 billion; New funds: $50.53 million. Country: Tanzania; Existing IDA funds: $1.289 billion; New funds: $89.03 million. Country: Kenya; Existing IDA funds: $895.46 million; New funds: $88.93 million. Country: Ghana; Existing IDA funds: $790.81 million; New funds: $47.59 million. Country: Democratic Republic of Congo; Existing IDA funds: $710 million; New funds: $40 million. Country: Uganda; Existing IDA funds: $544.91 million; New funds: $70.09 million. Country: Nepal; Existing IDA funds: $327.74 million; New funds: $94.65 million. Country: Burkina Faso; Existing IDA funds: $349.96 million; New funds: $36.54 million. Country: Afghanistan; Existing IDA funds: $257.98 million; New funds: $69 million. Country: Mali; Existing IDA funds: $274.77 million; New funds: $50.38 million. Country: Mozambique; Existing IDA funds: $230.92 million; New funds: $75.23 million. Country: Rwanda; Existing IDA funds: $248.03 million; New funds: $40.53 million. Country: Benin; Existing IDA funds: $134.93 million; New funds: $39.84 million. Country: Haiti; Existing IDA funds: $135.02 million; New funds: $38.32 million. Country: Kyrgyzstan; Existing IDA funds: $136.31 million; New funds: $26.08 million. Country: Zambia; Existing IDA funds: $120.34 million; New funds: $19.66 million. Country: Burundi; Existing IDA funds: $113 million; New funds: $25.4 million. Country: Laos; Existing IDA funds: $94.21 million; New funds: $38.11 million. Country: Liberia; Existing IDA funds: $83.21 million; New funds: $42.28 million. Country: Niger; Existing IDA funds: $110 million; New funds: $7.11 million. Country: Tajikistan; Existing IDA funds: $49.38 million; New funds: $52.38 million. Country: Sierra Leone; Existing IDA funds: $50 million; New funds: $40.98 million. Country: Malawi; Existing IDA funds: $46.79 million; New funds: $34.1 million. Country: Togo; Existing IDA funds: $56.97 million; New funds: $20.41 million. Country: Cambodia; Existing IDA funds: $65 million; New funds: $8.18 million. Country: Central African; Existing IDA funds: $48.61 million; New funds: $11.76 million. Country: Madagascar; Existing IDA funds: $40 million; New funds: $10 million. Country: Gambia, The; Existing IDA funds: $12.13 million; New funds: $21.6 million. Country: Guinea-Bissau; Existing IDA funds: $5 million; New funds: $27.98 million. Country: Mauritania; Existing IDA funds: $9.96 million; New funds: $15.54 million. Country: Solomon Islands; Existing IDA funds: $7.2 million; New funds: $3.17 million. Country: Guinea; Existing IDA funds: $0; New funds: $10 million. Country: Comoros; Existing IDA funds: $3.79 million; New funds: $5.12 million. Country: Zimbabwe; Existing IDA funds: $0; New funds: $7.3 million. Country: Somalia; Existing IDA funds: $0; New funds: $7 million. Country: Eritrea; Existing IDA funds: $0; New funds: $0. Country: Chad; Existing IDA funds: $0; New funds: $0. Source: GAO analysis of World Bank data. [End of figure] Within the four crisis response initiatives that sought to increase the speed of disbursements of commitments from existing IDA funds, first year disbursement rates varied by country, as shown in figures 15 through 18. Figure 15: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds through the Global Food Crisis Response Program, by Country: [Refer to PDF for image: horizontal bar graph] Country: Cambodia; Disbursement rate: 105.7%. Country: Ethiopia; Disbursement rate: 102.49%. Country: Bangladesh; Disbursement rate: 95.05%. Country: Mozambique; Disbursement rate: 94.49%. Country: Nepal; Disbursement rate: 38.03%. Country: Tanzania; Disbursement rate: 34.22%. Country: Kyrgyzstan; Disbursement rate: 27.07%. Country: Kenya; Disbursement rate: 3.67%. Country: Madagascar; Disbursement rate: 0%. Country: Haiti; Disbursement rate: 0%. Source: GAO analysis of World Bank data. Note: We determined the first year disbursement rates by calculating the total disbursements for each project for the first four quarters following project approval, then taking the ratio of total disbursements to the total committed funds. These rates include projects that were approved under the Global Food Crisis Response Program between 2008 and 2010. Some disbursement rates may exceed 100 percent due to exchange rate fluctuation between 2008 and 2010. [End of figure] Figure 16: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds through the Fast Track Facility, by Country: [Refer to PDF for image: horizontal bar graph] Country: Central African Republic; Disbursement rate: 102%. Country: Laos; Disbursement rate: 102%. Country: Burkina Faso; Disbursement rate: 100%. Country: Democratic Republic of Congo; Disbursement rate: 86%. Country: Ghana; Disbursement rate: 50%. Source: GAO analysis of World Bank data. Note: We determined the first year disbursement rates by calculating the total disbursements for each project for the first four quarters following project approval, then taking the ratio of total disbursements to the total committed funds. These rates include projects that were approved under the Fast Track Facility between 2009 and 2010. Some disbursement rates may exceed 100 percent due to exchange rate fluctuation between 2008 and 2010. [End of figure] Figure 17: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds through the Rapid Social Response Program, by Country: [Refer to PDF for image: horizontal bar graph] Country: Rwanda; Disbursement rate: 102.01%. Country: Burkina Faso; Disbursement rate: 92.72%. Country: Benin; Disbursement rate: 90.17%. Country: Malawi; Disbursement rate: 75.42%. Country: Nepal; Disbursement rate: 69.06%. Country: Haiti; Disbursement rate: 59.89%. Country: Ghana; Disbursement rate: 41.09%. Country: Kyrgyzstan; Disbursement rate: 40.18%. Country: Tanzania; Disbursement rate: 38.91%. Country: Togo; Disbursement rate: 36.55%. Country: Democratic Republic Congo; Disbursement rate: 28.47%. Country: Ethiopia; Disbursement rate: 23.79%. Country: Sierra Leone; Disbursement rate: 19.42%. Country: Uganda; Disbursement rate: 15.3%. Country: Liberia; Disbursement rate: 12.32%. Country: Kenya; Disbursement rate: 11.69%. Country: Laos; Disbursement rate: 10.5%. Country: Comoros; Disbursement rate: 9.29%. Country: Afghanistan; Disbursement rate: 7.89%. Country: Central African; Disbursement rate: 7.2%. Country: Solomon Islands; Disbursement rate: 5.27%. Country: Bangladesh; Disbursement rate: 0.74%. Country: Guinea-Bissau; Disbursement rate: 0. Source: GAO analysis of World Bank data. Note: We determined the first year disbursement rates by calculating the total disbursements for each project for the first four quarters following project approval, then taking the ratio of total disbursements to the total committed funds. These rates include projects that were approved under the Rapid Social Response between 2008 and 2010. Some disbursement rates may exceed 100 percent due to exchange rate fluctuation between 2008 and 2010. [End of figure] Figure 18: First Year Disbursement Rates of World Bank Commitments from Existing IDA Funds, Pilot Crisis Response Window, by Country: [Refer to PDF for image: horizontal bar graph] Country: Benin; Disbursement rate: 96.24. Country: Rwanda; Disbursement rate: 93.13. Country: Tajikistan; Disbursement rate: 92.42. Country: Burkina Faso; Disbursement rate: 87.67. Country: Mali; Disbursement rate: 80.14. Country: Malawi; Disbursement rate: 74.51. Country: Togo; Disbursement rate: 64.71. Country: Laos; Disbursement rate: 50.83. Country: Central African; Disbursement rate: 50.64. Country: Kyrgyzstan; Disbursement rate: 44.05. Country: Uganda; Disbursement rate: 43.88. Country: Sierra Leone; Disbursement rate: 36.13. Country: Guinea-Bissau; Disbursement rate: 32.92. Country: Democratic Republic of Congo; Disbursement rate: 26.71. Country: Mauritania; Disbursement rate: 26.06%. Country: Liberia; Disbursement rate: 25.39%. Country: Nepal; Disbursement rate: 20.1%. Country: Afghanistan; Disbursement rate: 12.72%. Country: Gambia, The; Disbursement rate: 12.6%. Country: Ghana; Disbursement rate: 9.82%. Country: Comoros; Disbursement rate: 9.29%. Country: Tanzania; Disbursement rate: 9.17%. Country: Bangladesh; Disbursement rate: 5.99%. Country: Haiti; Disbursement rate: 5.33%. Country: Mozambique; Disbursement rate: 2.86%. Country: Kenya; Disbursement rate: 2.83%. Country: Zambia; Disbursement rate: 1.83%. Country: Ethiopia; Disbursement rate: 0. Country: Burundi; Disbursement rate: 0. Source: GAO analysis of World Bank data. Note: We determined the first year disbursement rates by calculating the total disbursements for each project for the first four quarters following project approval, then taking the ratio of total disbursements to the total committed funds. These rates include projects that were approved under the Pilot Crisis Response Window during 2010. Some disbursement rates may exceed 100 percent due to exchange rate fluctuation between 2008 and 2010. [End of figure] [End of section] Appendix IV: Comments from the U.S. Department of the Treasury: Note: GAO comments supplementing those in the report text appear at the end of the appendix. Department Of The Treasury: Washington, D.C. September 12, 2011: Thomas Melito: Director, International Affairs and Trade: Government Accountability Office: 441 G Street N.W. Washington, D.C. 20548: Dear Mr. Melito, Thank you for the opportunity to comment on the draft report regarding the International Financial Institution (IFI) response to the crises in low-income countries. The United States strongly advocated for increased IFI engagement in low-income countries (LICs) during the recent food, fuel, and financial crises. We recognized the potentially severe impact that these crises could have for the poorest in these countries, and we worked closely with other G-20 countries to shape the IFI response in a way that hoped for the best, but prepared for the worst. As such, we have a strong interest in the findings of your report. The GAO report provides a good overview of how the IMF responded forcefully to the crisis in order to protect the poorest countries. As the GAO report indicates, the IMF significantly reformed its concessional facilities in order to respond to the crisis, more than doubling resources available to lend to LICs, raising access limits, and increasing the flexibility of its instruments. The United States was instrumental in pressing for these reforms in the IMF Board and in G-20 meetings. Treasury also worked closely with Congress to ensure that $1 billion in profits from gold sales would be allocated towards increased concessional resources for LICs. The GAO report, in assessing the Funds response to the crisis in the poorest countries, focuses on a comparison of program and non-program LICS, indicating that the contribution of IMF programs to government spending in LICs is difficult to establish. The U.S. Treasury closely monitors IMF lending to LICs to ensure that programs appropriately preserve pro-poor spending. In our analysis, IMF programs agreed to in the wake of the crisis had appropriate fiscal targets, and as a general policy the IMF sought to accommodate an increase in pro-poor spending in these programs. In contrast to prior crises, almost 90 percent of LICs in both program and non-program countries were able to increase real expenditures during the crisis. LIC spending on public investment, health, and education grew at a faster rate during the crisis, and many LICs took steps to strengthen social protection systems even as revenues declined. The goal of an IMF program is to provide countries with balance-of-payments support, which in turn can help to create fiscal and monetary space to implement structural adjustments and stabilize the economy within a sound macroeconomic framework commensurate with economic reality. Thus, while we appreciate GAO's analysis of program vs. non-program countries, a more meaningful comparison would have been to examine the impact of IMF programs on an individual country's expenditures on health, education and social safety nets. [See comment 1] Treasury also closely monitors World Bank lending to LICs. As the GAO report correctly notes, the World Bank undertook five initiatives to complement the response of the International Development Association (IDA) to the crisis, specifically the: 1) Global Food Crisis Response Program: 2) IDA Fast-track Facility; 3) Rapid Social Response Program; 4) Infrastructure Recovery and Assets Platform; and 5) Pilot Crisis Response Window. Treasury was actively involved in the creation of these facilities and has been working hand-in-hand with the World Bank to ensure that these facilities adhere to the highest standards of fiduciary controls, results reporting and capacity building at the recipient country level. It is also worth noting that these facilities were designed to achieve multiple objectives, notably to introduce greater flexibility and improve the speed with which operations are developed, approved and funds committed; to increase the awareness with the intent of raising additional resources; and to work with recipient governments to improve disbursements for development. The GAO report's focus on speed of disbursements is just one measure of effective crisis response, and it is important to keep these other objectives in mind in considering the effectiveness of the World Bank's efforts. [See comment 2] We also believe it is important to consider the trade-off between speed of disbursements and the need to ensure adequate governance structures and fiduciary controls are in place. This is a particular challenge in a crisis environment, but we believe that the MDBs have managed it well in recent years. The MDBs have strong fiduciary controls, as well as social and environmental protections, all of which serve to ensure that U.S. taxpayer funds are being used appropriately. And, as GAO notes, countries' absorptive capacity affects the rate at which the MDBs can disburse effectively. We do not believe the MDBs should sacrifice these standards even in a crisis environment. The evidence suggests to us that the World Bank and other MDBs managed an effective crisis response without undermining these critical controls. Further, we believe it is important to note the large and real increase in resource flows (including disbursements) generated by the World Bank and other MDBs in response to the call by the G20 for an aggressive crisis response. Collectively, the MDBs have increased (or are on track to increase) commitments by over $100 billion over the period 2009-11, an effort that had a strongly positive impact in support of the developing world. From the U.S. perspective, the MDB's crisis support helped to bolster economies that represent critical export markets. In this way, the MDBs were supporting our recovery. The lessons and experience from this crisis period have fed into the creation of a Crisis Response Window (CRW) in IDA-16, designed to be deployed in countries suffering from acute natural disasters or exogenous economic shocks. In addition to the IDA-16 Crisis Response Window, we are actively working with the Bank to improve the way it approaches and delivers on crisis assistance. More broadly, we believe that the best role for the Bank (and its core expertise) is to provide long-term support in helping developing countries build greater resilience to future crises. By working closely with the Bank and other MDBs to put in place the right fiduciary arrangements and strengthen country capacity to absorb and manage MDB assistance, Treasury can improve the quality of the Bank's and other MDBs' interventions and improve the monitoring and reporting of development results. We appreciate the work that has gone into this report. We believe that the IFIs appropriately responded to the crisis and effectively managed the tradeoffs associated with quickly disbursing funds in an environment of limited capacity to absorb aid. Sincerely, Signed by: Scott Morris: Deputy Assistant Secretary: International Development Finance and Debt: Signed by: Mark Sobel: Deputy Assistant Secretary: International Monetary and Financial Policy: The following are GAO's comments on the U.S Treasury letter, dated September 12, 2011. GAO Comments: 1. We emphasized that accounting for the differences between program and non-program countries is critical to estimating the impact of IMF programs on spending during the crisis, which the IMF did not do in their 2010 report. 2. We acknowledged that the speed of disbursements is one measure, among others, of effective crisis response. We acknowledged the Treasury's statement that the need to ensure that recipient country capacity and governance controls were sufficiently robust to absorb the additional resources provided during the crisis period played a role in the speed of disbursements. [End of section] Appendix V: Comments from the World Bank: Note: GAO comments supplementing those in the report text appear at the end of the appendix. The World Bank: Joachim von Amsberg: Vice President and Head of Network: Operations Policy and Country Services: Washington, D.C. 20433: U.S.A. September 14; 2011: Mr. Thomas Melito: Director, International Affairs and Trade: U.S. Government Accountability Office: 441 G Street, NW: Washington, DC 20548: Dear Mr. Melito: Thank you for the opportunity to comment on the United States Government Accountability Office (GAO) draft report on the Response of International Financial Institutions (IFIs) to Crises in Low-Income Countries (LICs) (GA0-11-832). We appreciate the time and effort that has gone into producing this report and are grateful for GAO staffs kind interactions with us during preparation. We welcome and agree with GAO's overall conclusion that IFIs, including IDA and IFC, met many goals in response to the crises in LICs. The World Bank Group (WBG) is proud of its response to the crises in the LICs and we are continuously seeking ways to further enhance our efficiency and development effectiveness, in consultation with our stakeholders. The World Bank Group's response to the crises in LICs: In response to the crises, the WBG achieved a significant increase in both its commitments and its disbursements of financial resources to LICs. The increase in IDA commitments to the 36 LICs included in GAO's report was dramatic: a 48.6% increase in IDA commitments to LICs in the four years following the onset of the crisis (FY2008-2011) compared to the four year period preceding the crisis (FY2004-2007). IDA disbursements to the 36 LICs also increased, from $16.6 billion to $19.6 billion, comparing the two four year periods; an 18.1 percent increase in disbursements to LICs. [Figures: vertical bar graphs: Commitments and Disbursements] IDA as a whole committed a record $56 billion in FY08-11 to promote economic growth, fight poverty and protect the poor in the context of the crises. This was a 43 percent increase over the pre-crisis period (FY04-07) and a record high for the global development institution. Over the same period of time, disbursements to IDA countries as a whole increased by 20 percent. This support has helped countries deal with the devastating effects of the global financial meltdown. In addition to significantly scaling up its response, the Bank also streamlined procedures. The average preparation time for operations was cut by one third. We cut red tape to enable swifter implementation of shovel-ready projects. As the report acknowledges, the WBG's lnternational Finance Corporation responded to the crises through investments, trade initiatives and enhanced cooperation with donors. Like IDA, it did so within resource constraints and taking into appropriate account the limited capacity of some countries to implement programs quickly and effectively. Key international partners have noted the WBG's swift and effective response to the crises, helping to lay the foundations for recovery (see box below). In addition to financial support, by providing cutting-edge expertise and policy advice the Bank helped Governments put in place policy and institutional reforms to prevent future crises and to promote good governance including in the private and financial sectors. The Views of Some Key International Partners: The Bank "fast-tracked resources for the poorest countries" - United Kingdom Government, Department for International Development. The Bank "accelerated processing and approval procedures with front- loading of IDA resources" - Organization for Economic Cooperation and Development "The increased lending by dramatic amount by the World Bank in this context of recession reflects a role it has now played in helping to stabilize world economic growth" - Independent Evaluation Group. More importantly, millions of poor bear testament to FDA's role in building the foundations for resilience and recovery in the context of the crisis. Between FY09-11, IDA financed 23 new public works projects in response to the crisis. Two-thirds of this lending was in Africa. For example, Ethiopia's Productive Safety Net workfare program, a flagship for Africa in terms of a safety net system, reaches more than eight million rural dwellers with much-needed support while generating community assets. Since FY09, IDA support for short- and medium-term food supply crisis response measures are estimated to have reached 5.9 million faun households to date, and are expected to directly benefit at least 8.8 million farm households when all activities are fully implemented. Specific comments on the GAO's draft report: Despite the overall positive conclusions, we feel that the report could more fully reflect the speed and scale of our response. We have the following specific comments in this regard: * The major thrust of the Bank's response to the crises was through mainstream IDA. However, the GAO report focuses heavily on a marginal slice of the Bank's response ” namely additional resources that were leveraged through various incremental crisis response initiatives. We would emphasize that the crisis facilities were created to complement overall IDA development financing during the crises: they were only a small part of the Bank's total response. [See comment 2] * Measured using standard definitions, the disbursement ratio for IDA projects in the 36 LICs increased from 25.4 percent pre-crisis to 26.4 percent post-crisis, Taking a portfolio-wide view, IDA's overall disbursements to LICs did not slow during the crises. The GAO report compares the first year of disbursements of operations approved in the crisis period to the first year of disbursements of operations approved in the pre-crisis period. We disagree with the use of this ad hoc indicator, because it overlooks disbursements made from the much larger existing portfolio during the crises. Total disbursements from the World Bank ongoing and new projects allowed the continuation of investments and programs that would otherwise have slowed down or stopped because of the crises. Furthermore GAO's indicator ignores the very impact of the crises on the countries' capacity to launch and implement new projects in more challenging circumstances. [See comment 3] * IDA accelerated delivery without compromising attention to governance and aid effectiveness (which remain central priorities for both the WBG and the United States Treasury). As the report emphasizes, disbursements did not increase uniformly across all countries. IDA makes no apologies in this regard and is proud of its continued commitment to sustainable and responsible development. The needs, capacity and governance of countries vary over time. [See comment 4] * There is growing evidence that IDA supported public spending for essential services. The World Bank never claimed it would single handedly maintain LIC public spending intact”-merely that it would contribute to this goal in priority areas. Financial support to LICs was through scores of credits and grants that contributed to public spending: the Bank routinely analyzes and monitors poverty reducing expenditures in IDA countries and there is ever greater evidence on this in country-specific public expenditure reviews and dedicated economic reports. It is also unclear why the report only focuses on two initiatives: the Fast Track Facility and the Infrastructure Platform when assessing IDA support to public spending. Moreover, we believe the report should acknowledge more clearly that definitive evaluation will only be possible when the projects in question are complete ” most are still in train. [See comment 5] * Finally, on debt sustainability aspects, we note that the analysis provided in the report is based on a sample of just three countries and so cannot achieve its stated purpose of assessing the realism of the 2010 Debt Sustainability Analysis projections. [See comment 6] Again, let me thank you for the opportunity to review and comment on this draft report. We look forward to working closely with the United States authorities, along with other development partners, as we strive to continuously improve our effectiveness and performance. Sincerely, Signed by: Joachim von Amsberg: Vice President and Head of Network Operations Policy and Country Services: The following are GAO's comments on the World Bank's letter, dated September 14, 2011. GAO Comments: 1. We included information about commitments and disbursements in our report, although our figures differ from the Bank's because we used different methodologies. Our methodology isolated those activities which were explicitly undertaken in response to the crises. 2. We acknowledged that the World Bank responded to crises in LICs through regular IDA lending and by establishing initiatives. Our analysis focused on the initiatives because these were specifically designed to respond to the crises. 3. Our analysis sought to isolate those activities which were explicitly undertaken in response to the crises. We did not assess the World Bank's standard approach to calculating disbursements. To measure the speed of disbursements, we first calculated the total disbursements for each project that took place during the first four quarters, including the quarter of project approval. We then determined the average disbursement rates for different groups of projects by using a weighted average, which is computed as the ratio between the sum of first year disbursements and the sum of the commitments for all projects that belonged to a group. 4. We acknowledged that disbursement rates, which vary over time, depend on a number of factors, including recipient country capacity, need, and governance, and the type of lending. 5. As we previously reported, IFIs do not independently track developing countries' poverty-reducing expenditures and instead rely upon developing countries' governments to provide such data, even though the accuracy of these data and country capacity to provide this information is questionable. We focused on the Fast Track Facility and the infrastructure program because these two initiatives were explicitly designed to support domestic spending in recipient countries during the crises. 6. We based our conclusion that the underlying macroeconomic projections might prove too optimistic on the realism of the 2010 DSA projections for three countries--Burundi, Ethiopia, and Tanzania--as well as on the current risks to the global economic recovery and rising commodity prices, reported by the IMF in August 2011. Our analysis of the three countries' DSAs is intended to be illustrative and not generalizable. Our conclusion, that projections which might be too optimistic could be mitigated by future DSAs and additional grants, is not dependent on these three countries. [End of section] Appendix VI: Comments from the International Monetary Fund: Note: GAO comments supplementing those in the report text appear at the end of the appendix. International Monetary Fund: Deputy Managing Director: Washington, D.C. 20431 September 8, 2011: Mr. Thomas Melito: Director, International Affairs and Trade: United States Government Accountability Office: Washington, DC 20548: Dear Mr. Melito: We welcome this opportunity to comment on your report entitled "International Financial Institutions Met Some Goals in Response to the Crises in Low-Income Countries But Impact on Spending Difficult to Establish." We broadly agree with the findings of the report, including the overview of the impact of the crisis on low-income countries (LICs), which aligns closely with our own analysis. Our comments are focused on three broad issues related to the IMF's engagement with our LIC members. Overall coverage of the IMF's response to the crisis: The report focuses its assessment rather narrowly on IMF disbursements, fiscal spending, and debt risk ratings. While these are important areas, other key elements of the IMF's comprehensive initiatives to assist LICs during the crisis receive little or no attention, and could have been more fully explored to provide a more balanced and complete account. In particular: [See comment 1] * The Fund's macro-policy advice to LICs in the face of the crisis, and especially the early call for a countercyclical policy response. This response was a first for LICs, which in past global downturns often had to cut spending as revenues and financing dried up. The IMF was able to advocate such a response because of the improved macroeconomic conditions in LICs, which in turn reflected these countries' policy efforts over an extended period prior to the crisis” in most cases, with IMF support. * The doubling of access levels (the amounts countries could draw under IMF concessional facilities) early on in the crisis (April 2009) was a key measure-”more important for countries, in practical terms, than the subsequent funding pledges. * The modifications to these facilities, to make them more flexible, easier to use, and tailored to country needs. These started with the reforms to the Exogenous Shock Facility (ESF) in September 2008, followed by a comprehensive reform of the concessional facilities in July 2009 that generalized and extended the earlier reforms. Discussion of fiscal spending: The report acknowledges the finding of a 2010 IMF staff paper that LICs with programs supported by the Fund were able to increase government spending more than non-program countries during the crisis. Indeed, almost 90 percent of LICs with a Fund-supported program were able to increase real primary expenditure at the height of the crisis in 2009.[Footnote 1] The GAO report goes on to argue that simply comparing program with non-program countries does not prove a causal link from program engagement to higher spending. This is, of course, a valid point, which is why we were careful not to claim causality in the 2010 paper. We agree with the GAO report's conclusion that a more rigorous analysis would be needed to conclude that IMF-supported programs resulted in increased government spending. In this context, however, we would note that a recently published IMF Staff Discussion Note conducted just such an analysis, albeit over a longer time period (1985-2009) and larger sample of countries (140), and found that IMF- supported programs have had a positive and significant effect on education and health spending in LICs, after controlling for other relevant factors.[Footnote 2] [See comment 2] Impact of the global crisis on debt sustainability: The GAO report rightly states that the basis for the IMF and World Bank's generally sanguine view on the impact of the crisis on debt sustainability depends on the "realism of the projections, which include quick recovery, implementation of policy reforms, and low inflation." (page 1). The GAO is correct to note that, like any projections, these may turn out to be too optimistic after the fact. However, we would note that the "quick recovery" assumption has generally been borne out so far, and see no reason to believe that the growth path for LICs over the 20-year horizon used for Debt Sustainability Analysis (DSA) will be materially different than currently assumed. Moreover, the DSA does have some built-in constructs that address potential concerns raised by the GAO regarding downside risks. In particular, stress tests and alternative scenarios performed in the context of DSAs are designed as a check on the realism of baseline projections, and feed directly into the debt distress ratings. As regards policy implementation, while some specific policy slippages are cited in two of the three ease studies in the GAO report, the report rightly emphasizes elsewhere that such findings cannot he generalized. No evidence is given of a systematic weakening of policy implementation in LICs since the crisis. Let me thank you once again for the opportunity to comment on your report and to offer our perspective on such an important topic. Sincerely yours, Signed by: Min Zhu: [End of section] Appendix VI Footnotes: [1] Emerging From the Global Crisis”Macroeconomic Challenges Facing Low-Income Countries, October 2010. [2] What Happens to Social Spending in IMF-Supported Programs, IMF SDN 11/15, August 2011. The following are GAO's comments on the IMF's letter, dated September 8, 2011. GAO Comments: 1. We reported on a wide range of IMF responses, including the IMF's call for countercyclical policy responses, improved macroeconomic conditions in LICs, doubling of access levels, and modifications to lending instruments. 2. We believe that comparisons of program and non-program country performance can be misleading without appropriate context or analysis. The IMF acknowledged that comparing program with non-program countries does not prove a causal link from program engagement to higher spending and notes a recent related study. The study found that IMF- supported programs were associated with increased spending on education and health as percentage of GDP or a percentage of spending in LICs, based on data from 1985 through 2009. We include a reference to the study, entitled "What Happens to Social Spending in IMF- Supported Programs"--which was released after our audit work had concluded--and we described its conclusions and relevance in a footnote. In particular, the study's results represent the average effect of an IMF-supported program over the time period, and therefore do not necessarily reflect the results during the crises response period. 3. We discussed the DSAs' assumption for "quick recovery" in relation to both LICs and the global economy. The IMF forecasted LICs' recovery in line with the forecast of a quick recovery for the global economy, which was expected to boost demand for LICs' exports. In August 2011, the IMF reported renewed risks to the global recovery, which means that projections for future export growth could be too optimistic. However, in commenting on this report, IMF noted that the "quick recovery" assumption had been borne out so far. Our report described the IMF's use of alternative scenarios and stress tests to arrive at a country's debt distress rating. However, we noted that these tests are very general and do not adequately reflect country-specific risks including political instability, adverse weather, global economic crises, and failure to implement reforms or make planned investment. [End of section] Appendix VII: GAO Contact and Staff Acknowledgments: GAO Contact: Thomas Melito, (202) 512-6901 or melitot@gao.gov: Staff Acknowledgments: Cheryl Goodman, Assistant Director; Marc Castellano; RG Steinman; Michael Hoffman; Elizabeth Kowalewski; and Arian Terrill made key contributions to this report. The team benefited from the expert advice and assistance of Adam Vogt, Etana Finkler, Bruce Kutnick, Fang He, Shirley Min, Leah DeWolf, Kathryn Buldoc, Martin de Alteriis, Tom McCool, Heneng Yu, Mary Moutsos, Mae Liles, and Holly Dye. [End of section] Footnotes: [1] The World Bank defines low-income countries as the 40 countries with a per capita gross national income of $995 or less for fiscal year 2011. [2] We focused on two World Bank organizations, the International Development Association (IDA) and the International Finance Corporation (IFC), and the International Monetary Fund (IMF). In this report we use the term IFIs to refer to this group. We did not include other international financial institutions, such as the African Development Bank and the Asian Development Bank. [3] The United States and other countries contribute to a fund that finances IDA activities. [4] See Supplemental Appropriations Act, 2009, Pub. L. No. 111-32, § 1402, 123 Stat. 1859, 1918, June 24, 2009. [5] Countries experiencing balance-of-payments problems have difficulty obtaining the financial resources needed to meet their payments to foreigners. [6] Our analysis will sometimes refer to a smaller group of countries based on the availability of data. Data were generally not available for North Korea and Somalia. Also, some World Bank documents refer to LICs as a broader group of about 80 countries eligible for IDA assistance, which includes the 40 in the scope of this report. When necessary, we use the term "IDA-eligible countries" to differentiate the larger group. [7] The G-20, an organization of finance ministers and central bank governors representing both industrialized and developing economies, was created in response to financial crises during the late 1990s as an annual forum to facilitate international economic policy cooperation. In response to the global economic crisis, the G-20 convened at a summit in November 2008, at which time major political leaders, including heads of states and governments, agreed on a joint plan of action to prevent the global economy from collapsing. Since then, the G-20 has convened on a regular basis. [8] The sale of IMF gold stocks concluded in December 2010. According to the IMF, the primary motivation for the gold sales was to support the IMF's new income model through the creation of an endowment funded by profits from these sales; support for LICs also included contributions from member governments; and the portion of gold resources intended to support LICs are expected to subsidize future financing to LICs. [9] IDA financing includes a 0.75 percent service charge. [10] Investment lending provides financing for a wide range of activities aimed at creating the physical and social infrastructure necessary to reduce poverty and create sustainable development. Development policy lending provides rapid financial assistance in the form of direct, untied budget support to governments for policy and institutional reforms aimed at achieving a set of specific development results. [11] The World Bank Group includes three other institutions: the International Bank for Reconstruction and Development, the Multilateral Investment Guarantee Agency, and the International Centre for Settlement of Investment Disputes. We did not include these institutions in our review because they are only minimally, if at all, engaged with LICs. [12] IMF lending to LICs is funded by member governments and some IMF contributions. [13] GAO, Developing Countries: The United States Has Not Fully Funded Its Share of Debt Relief, and the Impact of Debt Relief on Countries' Poverty-Reducing Spending Is Unknown, [hyperlink, http://www.gao.gov/products/GAO-09-162] (Washington, D.C.: Jan. 26, 2009). Beginning in 1998, we reported that even after receiving debt relief, many countries remained vulnerable to future debt problems even with sound economic policies. We reported that World Bank and IMF assumptions that countries would achieve strong export growth may have been optimistic for some countries, particularly those that rely upon a few commodities for export earnings. Such countries are particularly vulnerable to economic events such as a decline in the price or output of a primary export. [14] Data for most economic indicators were not available for North Korea and Somalia. [15] Real GDP is a measure of the value of all the goods and services produced in the economy in a given year, adjusted for changes in the price level. [16] From January 2007 to June 2008, the price index rose by 57 percent. [17] GAO, International Food Assistance: Better Nutrition and Quality Can Further Improve U.S. Food Aid, [hyperlink, http://www.gao.gov/products/GAO-11-491] (Washington, D.C.: May 2011); and GAO, Global Food Security: U.S. Agencies Progressing on Governmentwide Strategy, but Approach Faces Several Vulnerabilities, [hyperlink, http://www.gao.gov/products/GAO-10-352] (Washington, D.C.: March 2010). [18] According to the World Bank, in general, increases in domestic food prices have been smaller than increases in international food prices due to, for example, changes in exchange rates and domestic foods that are only loosely connected to international markets. However, some countries are dependent on imported food and therefore sensitive to fluctuations in international food prices. [19] The current account consists of the balance of trade in exports and imports of goods and services, and also includes current transfers, such as worker remittances--personal funds that the foreign- born send to their home countries. [20] A countercyclical policy moves counter to economic cycles, by cutting taxes or increasing spending in recessionary times and raising interest rates to tighten credit and curb spending during inflationary periods. [21] Real primary spending includes all expenditures (current and capital) except interest, adjusted for inflation using the GDP deflator. The composition varies by country. It could include spending to safeguard social safety nets in the context of declining revenues or capital spending, such as infrastructure. [22] Data for North Korea was not available. Foreign direct investment is net inflow of investment to acquire a lasting management interest (10 percent or more of voting stock) in an enterprise operating in an economy other than that of the investor. It is the sum of equity capital, reinvestment of earnings, other long-term capital, and short- term capital as shown in the balance of payments. [23] To measure the speed of disbursements, we first calculated the total disbursements for each project that took place during the first four quarters, including the quarter of project approval. We then determined the average disbursement rates for different groups of projects by using a weighted average, which is computed as the ratio between the sum of first year disbursements and the sum of the commitments for all projects that belonged to a group. [24] Similarly, first year disbursements from projects approved between 2008 and 2010 totaled $6.3 billion, an increase of about $700 million, or 12.7 percent, from first year disbursements for projects approved between 2005 and 2007. [25] IDA resources are replenished by donor countries every 3 years and committed in fixed allocations to countries' long-term development programs. [26] According to IFIs, the MDGs were designed to provide a framework for the entire international community to work together toward a common end: making sure that human development reaches everyone, everywhere. Among others, MDGs include eradicating extreme poverty and hunger. [27] The World Bank reported that effective social protection programs directly reduce poverty and inequality and build resilience by helping to provide individuals and families with the flexibility to adjust their consumption over time and cope with shocks. [28] For more information, see the Framework Document for a Global Food Crisis Response Program (Report Number 43841), available at [hyperlink, http://go.worldbank.org/4GNDMS8VT0]. [29] For more information, see the Proposal for an IDA Financial Crisis Response Fast-Track Facility (Report Number 46735), available at [hyperlink, http://go.worldbank.org/MP3GUZTYU0]. [30] Safety nets are social protection programs targeted to the poor or vulnerable. They include cash transfers, school lunch programs, public works projects, and fee waivers for essential services such as health, nutrition, education and heating. [31] For more information, see the Framework Document for a Rapid Social Response Program (Report Number 48121), available at [hyperlink, http://go.worldbank.org/YXCMLIJYD0]. [32] For more information, see the Infrastructure Recovery and Assets Platform overview, available at [hyperlink, http://siteresources.worldbank.org/INTSDNET/Resources/5944695- 1247775731647/InfrastructureRecoveryandAssets_overview_01.12.2010.pdf]. [33] For more information see the Proposal for a Pilot IDA Crisis Response Window (Report Number 51848), available at [hyperlink, http://siteresources.worldbank.org/IDA/Resources/Seminar%20PDFs/73449- 1257448780237/CRW_Official_Use.pdf]. [34] In March 2009, the World Bank also requested that donors provide a portion of their domestic economic stimulus funding to help address the impacts of the crises on developing countries. However, the World Bank's Independent Evaluation Group reported that donors' minimal response led the World Bank to instead focus on replenishing IDA funds and developing a crisis response window. See Independent Evaluation Group, The World Bank Group's Response to the Global Economic Crisis (Washington, D.C.: World Bank, 2010). [35] New financial assistance was provided from internal World Bank resources through the Pilot Crisis Response Window and the food program. The internal resources available through these initiatives include $1.2 billion linked to arrears clearance and about $300 million in income earnings. Countries were required to clear their arrears, or pay any late principal and interest charges, to IDA before they could receive debt relief under the Heavily Indebted Poor Countries Initiative. Donors provided some of the funding to clear these arrears. Funds that were provided by donors to compensate for prior arrears clearance costs or to be used for future arrears clearance were instead used by the Pilot Crisis Response Window. Income earnings refer to higher-than-anticipated returns on IDA's investments and income earned from interest payments to the World Bank by middle-income borrower countries. [36] Efforts to improve data through the social protection program's trust fund operations in LICs were to include funding technical diagnostics assessing the impacts of increasing prices and financial shocks on various communities and social groups; more effective identification and enrollment of beneficiaries and the disbursement of benefits; and monitoring and evaluation, among other activities. [37] Over the next 3 years, the permanent Crisis Response Window will be able to draw upon $2.1 billion. [38] Our analysis included only projects approved after the World Bank first stated its intention to respond to any of the three crises. This occurred in early 2008, with the establishment of the Global Food Crisis Response Program. [39] For the social protection program, we were also able to identify a comparison group of similar projects approved from 2005 through 2007, using project "theme codes" provided by the World Bank. [40] In July 2011, the World Bank's Independent Evaluation Group recommended that the World Bank engage consistently with governments during stable times to help countries develop safety nets, place a continued emphasis on building social safety net systems and institutional capacity, and strengthen engagement in LICs. World Bank Management agreed with these recommendations. See Independent Evaluation Group, Social Safety Nets, An Evaluation of World Bank Support 2000-2010 (Washington, D.C.: World Bank, 2011). [41] Even though more countries received disbursements faster during the crises period, the average first year disbursement rate for all LICs is slower because this average takes into account the amount of commitments in addition to disbursement rates. We calculated the average first year disbursement rate for all countries using a weighted average. Countries with faster disbursement rates received 43 percent of all commitments and were given a lower weight in the weighted average calculation. [42] For more information, see G-20 Chair Consultations of LICs on Flexibility and Adaptability of IFIs in Freetown (8/14/09) and London (8/17/09). [43] http://www.gao.gov/products/GAO-09-162]. [44] By contrast, the World Bank prepared about 20 technical diagnostics for middle-income countries. [45] Other initiatives established in response to the financial crisis have relatively small investments in LICs, including approximately $13.2 million for microfinance, advisory services, and bank capitalization activities. [46] According to IFC, the level of trade supported is determined for each transaction by the value associated with the trade and IFC's guarantee percentage. [47] According to the World Bank Independent Evaluation Group, the amount available for IFC to respond to impacts of the financial crisis in all member countries was $36 billion. In contrast, other institutions at the World Bank had $142 billion available to lend to these same countries, and at the April 2009 G-20 Summit, world leaders pledged to support a tripling of the IMF's lending resources from about $250 billion to $750 billion. Also, IFC as a whole estimated that to respond to the financial crisis, it could invest 5 percent more annually in fiscal years 2009 through 2011 than it did in 2008. In contrast, IDA could increase lending overall by 25 percent. [48] According to U.S. Treasury, total IMF lending in 2009 and 2010 to a larger group of LICs (IMF generally uses a broader definition than the 40 countries in the scope of this report) exceeded $5.5 billion, in line with 2009 legislation which requires the Secretary of the Treasury to seek to ensure that the IMF provide support to LICs of not less than $4 billion as a condition of the gold sales. [49] The $250 billion in support was a general Special Drawing Right (SDR) allocation implemented in August 2009. An SDR is an interest- bearing international reserve asset created by the IMF. An SDR allocation is a way of adding to members' international reserves, allowing members to reduce their reliance on more expensive domestic or external debt for building reserves. [50] The countries included are the 62 countries for which data were available, using a broader definition of LICs. [51] In a 2010 report, Emerging from Global Crisis: Macroeconomic Challenges Facing Low-Income Countries, the IMF found that in 2009 LICs with IMF-supported programs were more likely to increase primary expenditures (spending not including interest payments) and on average increased primary expenditures more than non-program LICs. As noted above, the countries included in the IMF analysis are the 62 countries for which data were available, using a broader definition of LICs. Liquidity constraints are limits on borrowing that could make it difficult to finance spending when the economy suffers a downturn. The analysis includes 36 of the 40 LICs we reviewed. [52] We repeated the analysis for the 40 LICs and found that while spending increased in program LICs more than non-program LICs, this result also remained sensitive to small changes in the sample of non- program LICs. [53] We used indicators of institutional quality and governance from the Worldwide Governance Indicators. See Daniel Kauffman, Aart Kraay, and Massimo Mastruzzi, "The Worldwide Governance Indicators: Methodology and Analytical Issues," Policy Research Working Paper 5430 (The World Bank, September 2010). [54] We replicated the IMF's analysis of 62 LICs, which compared average (median) growth in real primary expenditures in 2009, each time omitting certain LICs from the sample of non-program countries that were least comparable to program LICs due to self-selection into IMF programs. In each case we compared mean and median growth in real primary expenditures for the two samples. For the two sensitivity analyses described in this report our results held for both median and mean spending growth. [55] See, e.g., Aye Y. Evrensel, "Effectiveness of IMF-supported Stabilization Programs in Developing Countries," Journal of International Money and Finance (2002) or Zlata Hajro and Joseph P. Joyce, "A True Test: Do IMF Programs Hurt the Poor?" Applied Economics (2009). Multivariate techniques, including those that address selection bias such as Heckman selection models, instrumental variables, or synthetic control methods, would be necessary to properly account for differences between program and non-program LICs. In a separate and more sophisticated analysis, IMF researchers found that IMF-supported programs were associated with increased spending on education and health as a percentage of GDP or a percentage of spending in LICs, based on data from 1985 through 2009. The study's results represent the average effect of an IMF-supported program over the time period, and therefore do not necessarily reflect the results during the crises response period. Benedict Clements, Sanjeev Gupta, and Masahiro Nozaki, "What Happens to Social Spending in IMF-supported Programs?" IMF Staff Discussion Note SDN/11/15 (2011). [56] According to IMF and World Bank documents, until the impact of the crisis dissipates, the IMF and World Bank will prepare full DSAs for LICs annually. After that, full DSAs are expected to be prepared once every 3 years, with short annual updates in interim years. [57] IDA assigns countries a CPIA rating based on 16 indicators in 4 categories. The 4 categories with examples of their indicators follow: economic management (macro, fiscal, debt policy); structural policies (trade, financial sector, business regulatory environment); policies for social inclusion and equity (gender equality); and public sector management and institutions (property rights and rule-based government). [58] The present value of debt is a measure that takes into account the concessional, or below-market, terms that underlie most of these countries' loans. Present value of debt is defined as the sum of all debt-service obligations (interest and principal) on existing debt, discounted at the market interest rate. [59] According to the research group Development Finance International, the "alternative scenarios" generated are not comprehensive, because they do not vary all relevant macroeconomic variables, exclude certain effects, and do not necessarily reflect the risks that a government may believe are likely to occur in its economic or borrowing prospects. While useful as an overall general tool to compare debt sustainability across multiple countries, individual countries need to interpret their results with caution in analyzing the risk of a debt crisis. [60] Each category of debt distress has specific debt thresholds. According to the World Bank and IMF, in the low risk category, all debt indicators are well below relevant country-specific debt-burden thresholds, and stress testing and country-specific alternative scenarios do not result in indicators significantly breaching thresholds. In the moderate risk category, the baseline scenario does not indicate a breach of thresholds; however alternative scenarios or stress tests result in a significant rise in debt-burden indicators over the projection period (nearing thresholds) or a breach of debt or debt-service thresholds. In the high risk category, the baseline scenario indicates a protracted breach of debt or debt-service thresholds but the country does currently not face payment difficulties. This is exacerbated by the alternative scenarios, or stress tests. For a country categorized as "in debt distress," current debt and debt-service ratios are in significant or sustained breach of thresholds. Actual or impending debt restructuring negotiations or the existence of arrears would generally suggest that a country is in debt distress. [61] Countries received full and irrevocable debt relief from creditors by meeting specific criteria such as maintaining good performance under an IMF-supported reform program, when they completed the Heavily Indebted Poor Countries Initiative. To qualify for this debt relief, countries had to meet additional criteria, including having unsustainable debt burdens. See GAO, Developing Countries: The United States Has Not Fully Funded Its Share of Debt Relief, and the Impact of Debt Relief on Countries' Poverty-Reducing Spending is Unknown, http://www.gao.gov/products/GAO-09-162] (Washington, D.C.: Jan. 26, 2009). [62] To the extent that countries are receiving more grants, IDA and the African Development Bank's African Development Fund reduce the volume of grant assistance provided under this system. Specifically, IDA and the African Development Fund reduce grant assistance by 20 percent for countries classified at a high or moderate risk of debt distress, thereby reducing available resources. The 20 percent volume reduction is divided into an "incentives"-related portion and a "charges"-related portion. The incentives-related portion is reallocated to IDA-only countries based on performance, and the charges-related portion is provided to creditworthy blend countries. According to IDA, this grant reduction was instated to maintain IDA's performance incentive. We estimate that, during the crisis response period, IDA committed about $1.5 billion less in assistance to the LICs due to its policy of reducing assistance by 20 percent when grant financing is provided. IDA was to reallocate these funds to all eligible IDA countries. This amount is slightly more than the total new financial assistance IDA allocated to these countries as part of its crisis response. [63] IMF, Staff Guidance Note on the Applications of the Joint Bank- Fund Debt Sustainability Framework for Low-Income Countries (Jan. 22, 2010). [64] In our analysis of the divergence of projections from historical averages, we focused on GDP growth and the current account and fiscal deficits, because GDP, exports, and government revenues are key measures of the country's capacity to repay debt. We also examined inflation, as inflation is vulnerable to food supply and energy price shocks and can affect the growth outcome, and foreign direct investment, which helps meet LICs' financing needs. We compared the historical averages of key macroeconomic variables with medium-and long-term projections for these variables, taking into account the underlying rationale for these projections as presented in the DSA and accompanying program papers, and interviewed IMF and IDA staff. We also examined the extent to which program reforms and key country- specific risk factors were included in the baseline scenario. See appendix I for additional information on our methodology. [65] Burundi, one of the poorest countries in the world, is emerging from more than a decade of civil conflict. [66] Remittances are recorded through official and unofficial channels, including estimates based on banking system flows. Remittances can be used explicitly in the DSA analysis when they are a large and stable source of income and the breaches under the analysis excluding remittances are not protracted. [67] According to the U.S. Treasury, the Ethiopian government has since removed the government-imposed bank credit ceilings, but replaced them with a still-repressive central bank directive. [68] Our analysis of remittance data from the World Bank's World Development Indicators database yielded significantly different results than the data in Ethiopia's DSA. For example, according to the World Development Indicators database, the remittances-to-GDP ratio averaged 1.4 percent from 2007 through 2009, whereas, according to the DSA, that ratio is expected to be 8.5 percent from 2010 onward. [69] IMF staff told us that such country-specific factors are addressed by additional risk assessments or stress tests that include shocks to export growth and GDP growth that last for only 2 years. [70] Available at [hyperlink, http://www.imf.org/external/np/fin/tad/extarr1.aspx]. These figures refer to the LICs in the scope of this report, but the IMF generally uses a broader definition of LICs. [End of section] GAO's Mission: The Government Accountability Office, the audit, evaluation and investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through GAO's Web site [hyperlink, http://www.gao.gov]. Each weekday, GAO posts newly released reports, testimony, and correspondence on its Web site. To have GAO e-mail you a list of newly posted products every afternoon, go to [hyperlink, http://www.gao.gov] and select "E-mail Updates." Order by Phone: The price of each GAO publication reflects GAO‘s actual cost of production and distribution and depends on the number of pages in the publication and whether the publication is printed in color or black and white. Pricing and ordering information is posted on GAO‘s Web site, [hyperlink, http://www.gao.gov/ordering.htm]. Place orders by calling (202) 512-6000, toll free (866) 801-7077, or TDD (202) 512-2537. Orders may be paid for using American Express, Discover Card, MasterCard, Visa, check, or money order. Call for additional information. To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Web site: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]: E-mail: fraudnet@gao.gov: Automated answering system: (800) 424-5454 or (202) 512-7470: Congressional Relations: Ralph Dawn, Managing Director, dawnr@gao.gov: (202) 512-4400: U.S. Government Accountability Office: 441 G Street NW, Room 7125: Washington, D.C. 20548: Public Affairs: Chuck Young, Managing Director, youngc1@gao.gov: (202) 512-4800: U.S. Government Accountability Office: 441 G Street NW, Room 7149: Washington, D.C. 20548:

The Justia Government Accountability Office site republishes public reports retrieved from the U.S. GAO These reports should not be considered official, and do not necessarily reflect the views of Justia.