73244 ... , • THE WORLD BANK II fRAGllIIHCONHln Providing Budget Aid In Situations Of Fragility: A World Bank - African Development Bank Common Approach Paper Disclaimer: This paper is a joint product of the Fragile and Conflict Affected Countries Group of the Operations Policy and Country Services Vice-Presidency at the World Bank, and the Governance, Economic and Financial Management Department at the African Development Bank. The findings of the paper reflect work in progress to encourage the exchange of ideas about ways to improve coordination in the provision of budget aid in fragile situations. The findings, interpretations, and conclusions expressed in this paper are entirely those of the staff of the respective units. They do not necessarily represent the views of the World Bank or the African Development Bank, their Executive Directors, or the countries they represent. Photo Credit (cover): Arne Hoel, World Bank Acknowledgements This paper was developed by a multi-institution core team led by Nicola Pontara (World Bank) and comprising Greg Ellis and Duncan Burrell (World Bank), Maylis Souque (Ministry of Foreign Affairs, France), Kate Tench and Emanuele Santi (African Development Bank), and Brendan Horton (Consultant). The paper has benefited at various stages from the guidance and support of Klaus Rudischhauser, José Correia-Nunes, Frans Baan, Sven Burgsdorff, Bronte Flecker, Cecile Tassin-Pelzer, Mikael Barfod, and David Johnson at the European Commission; Gabriel Negatu, Marlene Kanga, Carlos Santiso, and Damoni Kitabire at the African Development Bank; Louis Dicks-Mireaux and Dominique Desruelle at the International Monetary Fund; and Alastair McKechnie, Bill Byrd, Jan Walliser, Edward Mountfield and Francisco Carneiro at the World Bank. • i Background and Context The initial proposal to produce a Common Approach Paper on the Provision of Budget Support in Situations of Fragility (CAP) emerged in Tunis in June 2008, at the first tripartite meeting between the World Bank, European Commission and African Development Bank. The CAP was identified as an output in the thematic discussion on budget support in fragile situations. At the heart of the discussion was the belief that better coordination between the institutions was needed to increase the effectiveness of budget support when deployed with other aid instruments – in strengthening the transition from fragility to resilience. Given that presence of an International Monetary Fund-supported program is important for the engagement of the concerned institutions at the country level, the Fund was invited to join the group. Since June 2008, the scope of the CAP evolved, following an intensive consultative process between the concerned parties. It was decided that the focus should encompass the ensemble of instruments and sources of financing to the budget, rather than focusing solely on budget support operations. For this reason, the paper focuses on ‘budget aid’. Moreover, it was decided that in lieu of a legally-binding, signed statement of principles, the CAP should present a number of recommendations, including from a donor coordination standpoint, for the respective institutions to consider independently. The revised CAP also makes more extensive use of the available evidence on the use of budget aid in situations of fragility, including from retrospectives conducted by the institutions on budget support operations, and selected case studies. The World Bank and African Development Bank have been at the forefront of the development of this work. The European Commission has been consistently consulted for guidance and support during the writing of the CAP. The Commission is working on a Green Paper on the future of budget support with a specific focus on improving budget support operations in situations of fragility, which will take into account the findings of the CAP. The involvement of the IMF has been consultative in nature. • ii Acronyms and Abbreviations ADF African Development Fund AfDB African Development Bank ARTF Afghanistan Reconstruction Trust Fund AU African Union BSO Budget Support Operation CAP Common Approach Paper CAR Central African Republic CAS Country Assistance Strategy CPAR Country Procurement Assessment Review CPIA Country Policy and Institutional Assessment DBSL Development Budget Support Lending DCI Development Cooperation Instrument DPG Development Policy Grant DPL Development Policy Lending DPO Development Policy Operation EC European Commission EDF European Development Fund EPCA Emergency Post-Conflict Assistance FSF Fragile States Facility GBS General Budget Support GDP Gross Domestic Product HIPC Highly Indebted Poor Countries Initiative IBRD International Bank for Reconstruction and Development IDA International Development Association IMF International Monetary Fund LICUS Low-Income Countries Under Stress M&E Monitoring and Evaluation MDBS Multi-donor Budget Support MDGs Millennium Development Goals MDTF Multi-donor Trust Fund MOU Memorandum of Understanding NATO North Atlantic Treaty Organization NDS National Development Strategy OAS Organization of American States OECD-DAC Organization for Economic Development, Development Assistance Committee PAF Performance Assessment Framework PD Program Document PEFA Public Expenditure and Financial Accountability PEMFAR Public Expenditure Management and Financial Accountability Review PFM Public Financial Management PRS Poverty Reduction Strategy RCF Rapid Credit Facility RCW Recurrent Cost Window SSR Security Sector Reform UN United Nations WB World Bank • iii TABLE OF CONTENTS Executive Summary 1 1. Introduction 4 2. Conceptual Framework and Institutional Approaches to Budget Aid 5 2.1 The Concept of ‘Budget aid’ 5 2.2 Institutional approaches to the provision of budget support 8 3. Budget Aid in Situations of Fragility: Emerging Issues 12 3.1 Highlighting the advantages of budget aid 13 3.2 Managing risk in the provision of budget aid 15 3.3 Addressing the trade-offs between aid instruments, aid predictability and aid effectiveness 17 4. Selected Evidence on Budget Aid: Retrospectives and Case Studies 19 4.1 Evidence emerging from budget support retrospectives 19 4.2 Evidence emerging from selected case studies 20 5. Improving the Effectiveness in the Provision of Budget Aid 22 5.1 Key recommendations 23 5.2 Areas for further work and next steps 26 Annex 1. Retrospectives on Budget Support: Key Findings 26 References 33 Endnotes 36 List of Tables Table 1. Summary of Key Recommendations 3 Table 2. Selected Development Policy Operations in Fragile Situations 31 List of Boxes Box 1. Approaches to Fragility 6 Box 2. Definitions of Budget Support 7 Box 3. The Afghanistan Reconstruction Trust Fund (ARTF) 8 Box 4. Institutional Approaches to Resource Allocation and the Provision of Budget Support 9 Box 5. Typology of Risk 18 Box 6. World Bank’s Emerging Position on MoUs and PAFs 21 • iv Executive Summary This paper is concerned with improving coordination and the pooling of risk management in the provision of ‘budget aid’ in situations of fragility, among the World Bank and African Development Bank, in collaboration with the European Commission and IMF. Budget aid comprises all types of instruments and sources of assistance to the budget, including policy-based budget support lending and grants (instruments), as well as contributions channeled through multi-donor trust funds (sources of financing), which can finance recurrent expenditures subject to specific fiduciary arrangements. The overarching conclusion of the paper is that there is a need for greater coordination of approaches in the provision of budget aid between the participating institutions, with a view to helping fragile situations to embark on a path of stability and resilience. This can be done, inter alia, by providing budget aid that is more predictable and targeted at addressing the root causes of fragility and conflict. The paper conveys three main messages. First, rather than viewing budget aid as simply a transfer of financial resources to the country’s budget, and with a narrow focus on public financial management, it should be considered as a key element of an aid package that consists of evidence-based policy dialogue, analytical work, technical assistance, capacity building activities, as well as financial transfers. This package should be more ex- plicitly geared at addressing the underlying causes of fragility and supporting the transition toward resilience. This can be done by highlighting the role that budget aid can play in: stabilizing the macro-budgetary framework and allowing the state to carry out basic functions, to cement its legitimacy and contribute to maintaining political stability; supporting the longer-term endeavors of peace and state-building; and contributing to strengthening the capacity of recipient countries by channeling aid through national systems. Second, the risk elements surrounding the provision of budget aid need to be analyzed more deeply and shared more widely amongst the three institutions (World Bank, AfDB and EC). Working together to pool risk is a critical source of added value resulting from improved coordination of approaches. The typology of risk, the analysis of different categories of risk and the trade-offs that exist between them could be given more prominent attention in the documentation associated with budget aid. The risk of not engaging should be set against the benefits that can be reaped by successfully stabilizing a country, including the positive regional (and global) externalities that may be generated. Third, it is important to consider more systematically the choice and complementary nature of policy-based budget support lending and grants as well as other instruments to support recurrent expenditures such as MDTFs. Donors could fine tune the mix of instruments and sources of financing at their disposal depending on their exposure to risk to avoid negative consequences for countries resulting from the delay or withdrawal of disbursements through one mechanism. The paper also presents a number of recommendations – summarized in Table 1 – and areas for further engagement. The recommendations are intended to stimulate further debate and assist the three institutions in strengthening their effectiveness in the provision of budget aid, including from the standpoints of donor coordination and the pooling of risk. • 1 In the light of the diversity of approaches and progress made by each institution in addressing some of the issues raised in this paper, these recommendations may not apply uniformly to all of them. By the same token, the need for better coordination clearly requires the institutions to review these recommendations to see to what extent their adoption may facilitate more effective partnership between them. Given the great diversity of fragile situations, country-specific assessments should ultimately guide the decision with respect to the timing of engagement and choice of instrument and source of financing. • 2 Table 1. Summary of Key Recommendations Area of Intervention Specific Recommendations I. Improve donor coordination Ensure better coordination between institutions at both HQ and country levels at critical stages of the budget aid cycle and find non-legalistic, non-bureaucratic ways to strengthen working-level dialogue and partnerships. II. Reinforce the rationale of budget aid Articulate how budget aid can support the transi- tion toward resilience (macro-budgetary and po- litical stability, peace- and state-building, human security, regional and global dimensions), in com- bination with other aid instruments. III. Improve knowledge of the political Undertake/commission/share iterative politi- economy cal economy analysis (e.g. political stakeholder analysis) and use findings to make more informed decisions to engage in budget aid, manage risk and focus policy dialogue with partner countries. IV. Undertake a more comprehensive Undertake/commission/share risk analyses, adapt- analysis of risk ed to the specific characteristics of a fragile situa- tion, and use results to improve the coordination of approaches in the provision of budget aid. V. Consider non-traditional focus areas Consider non-traditional areas that are important to fostering stability and the legitimacy of the state (e.g. security sector reform (SSR), rule of law and justice, job creation), while maintaining selectivity, including at the sub-national level, as appropriate to country contexts. VI. Address issues of aid predictability and Enhance aid predictability by achieving a balanced effectiveness mix of instruments/sources of financing to the budget, and moving toward programmatic sup- port where possible with a series of single-tranche operations. VII. Improve Monitoring and Evaluation Develop peace and state-building indicators and (M&E) systems gauge the contribution of budget aid toward these objectives and include intermediate indicators in result frameworks to measure progress towards these long-term goals. • 3 1. Introduction The development community is increasingly focusing attention on the needs of fragile and weakly performing states. These countries, referred to here as ‘fragile situations’, face particularly severe development challenges. They are characterized by weak institutional capacity, poor governance, political instability, and ongoing violence or the legacy of past violence, with potential adverse effects on neighboring countries. They contribute signifi- cantly to the Millennium Development Goals (MDG) deficit (World Bank 2010a). Even though development assistance is inherently risky in these contexts, there is a strong rationale for engagement: the impact of well-designed and effectively implemented aid- financed programs can potentially be very high, because these countries start from a very low level of achievement. A successful exit from fragility also has a positive impact on neighboring countries. The question of collective impact in fragile situations becomes especially important, as it is rare that individual donors and agencies can exert a determining influence over out- comes single-handedly. Despite the considerable progress made by the donor community in adapting policies and procedures to dealing more effectively with fragile situations, the effectiveness of its engagement continues to be undermined by insufficient coordination and coherence and, in many cases, by the competing mandates of individual partners. While this problem is common across the development spectrum, its consequences can be particularly damaging in situations characterized by weak national capacity, which demand a more coherent, timely, and effective response. Indeed, poorly designed and managed development programs can exacerbate instability. This paper is concerned with improving coordination and risk management in the provi- sion of ‘budget aid’ in situations of fragility, among the World Bank, European Commis- sion (EC) and African Development Bank (AfDB). Given the importance of the presence of IMF-supported macroeconomic stabilization policies in the context of budget aid, the IMF has been actively engaged in this work. The overarching conclusion of the paper is that there is a need for greater coordination of approaches in the provision of budget aid between the participating institutions, with a view to helping fragile situations to em- bark on a path of stability and resilience. This can be done, inter alia, by providing budget aid that is more predictable and targeted at addressing the root causes of fragility and conflict. A major message of this paper is that budget aid should be considered as a key element of a package that consists of evidence-based policy dialogue, analytical work, technical as- sistance and capacity building activities, as well as financial transfers. This package should be geared more explicitly at addressing the underlying causes of fragility and strengthen- ing the legitimacy of the state. Another message is that the treatment of risk aspects in the provision of budget aid remains somewhat underdeveloped for all three institutions. Working more closely together to pool risk analysis and management information could potentially help spread and therefore reduce the risk faced by each individual institution. In riskier contexts, this is a critical source of added value resulting from improved harmo- nization of approaches in the provision of budget aid. • 4 This paper is structured as follows. Section 2 focuses on the conceptual framework, defi- nitions and institutional approaches to budget aid. Section 3 discusses the emerging is- sues on the provision of budget aid in situations of fragility. Section 4 (supplemented by Annex 1) reviews some evidence on the experience with budget aid, focusing mainly on budget support operations. Section 5 presents recommendations geared at improving coordination of approaches amongst the three institutions and sets out the next steps and areas for further work. 2. Conceptual Framework and Institutional Approaches to Budget Aid Because of its inherent cross-cutting nature, ‘budget aid’ can be at the core of an aid delivery package that is deployed in situations of fragility to strengthen the transition toward resilience, broadly defined here as being characterized by greater macro-economic and political stability, stronger public financial management, greater state legitimacy, re- duced risk of conflict, and improved human security. The conceptualizations of fragility by the institutions are summarized in Box 1. Section 2.1 reviews the definition of budget aid, including the budget support instrument and sources of budget financing. Section 2.2 takes stock of the institutional approaches to budget aid by the World Bank, EC and AfDB, focusing predominantly on general budget support. 2.1 The concept of ‘budget aid’ Budget aid is not limited to policy-based instruments (lending or grants) but can also in- clude the use of other sources of financing to the budget, such as multi-donor trust funds (MDTFs). The release of funds linked to policy-based budget support operations (BSOs, lending and grants) is usually triggered by the fulfillment of conditionality. Accordingly, the amount of aid provided through these instruments can vary with the recipient’s abil- ity to meet the requirements. MDTFs can finance discreet projects and also provide re- sources to the budget. For instance, they can be used to reimburse recurrent expenditures on a regular basis, subject to specific fiduciary arrangements (e.g. verification of payment) signed between the government and the manager of the trust fund (See Box 3). At times, BSOs are channeled through MDTFs using trust fund resources. As argued below in Sec- tion 3, the combined use of these instruments and sources of financing – that is, budget aid – can simultaneously address issues linked to aid effectiveness and aid predictability.1 Budget support. Budget support has become an increasingly important mode of devel- opment assistance, receiving growing attention from bilateral donors and international financial institutions in the context of a partnership-based approach to aid. Reflecting the recommendations of the Paris Declaration (2005) and the Accra Agenda for Action (2008), this shift has been motivated, inter alia, by: a move away from traditional project support reflecting the concerns of donors with parallel systems outside the government’s budget framework; greater emphasis on building country ownership, systems and capaci- ty; and, a shift from short- to medium-term support to reform. Increasing the proportion of aid funds that are included in the national budget process makes donors less prone to micromanage individual programs and allows governments to better plan and coordinate reform and development. • 5 Box 1. Approaches to Fragility The World Bank. The WB has recently switched from defining fragile states to ‘fragile situations’. A frag- ile situation is defined as having either: a) a composite World Bank, AfDB and Asian Development Bank (ADB) Country Policy and Institutional Assessment (CPIA) rating of 3.2 or less; or b) the presence of a UN and/or regional peace-keeping or peace-building mission (e.g. African Union, European Union, NATO), with the exclusion of border monitoring operations, during the past three years. The African Development Bank. The AfDB keeps a list of ‘fragile states’, which are eligible for additional resources under the Fragile States Facility. A country qualifies as fragile if it satisfies one of the following conditions: it has a composite AfDB/World Bank CPIA score of 3.2 or less; or, it has low income and does not have a CPIA score and it has had a UN or regional peace-building, peace-keeping or mediation opera- tion, in the past three years, with the exclusion of border monitoring operations. The European Union. The EU refers to ‘situations of fragility’, which apply to contexts where ‘the social contract is broken due to the State’s incapacity or inability to deal with its basic functions, and meet its obligations regarding: service delivery, management of resources, rule of law, equitable access to power, security and safety of the populace and protection and promotion of citizens’ rights and freedoms.’ The EU does not keep a formal list of fragile situations. However, EC regulations allow for the use of more flexible procedures and instruments once a country is declared ‘into crisis’ (Art. 72.73 of the Cotonou Agreement and Art. 168.2 of the Implementation rules of EC budget financial regulations). The International Monetary Fund. The IMF does not keep an official list of fragile states but refers to ‘fragile situations’ in its documents. The form and scope of the Fund’s engagement are determined by the underlying characteristics of fragility, such as institutional and policy implementation weaknesses. The Fund is actively engaged in fragile states in several ways: macroeconomic policy advice through surveil- lance; staff-monitored programs to establish a track record of policy performance; technical assistance and capacity building; and financial assistance through a variety of concessional lending instruments. Source: Institutions Websites and Guidelines (2009), EC (2007). Donor organizations use a range of definitions for budget support (Box 2). In current us- age, budget support refers to predictable, usually annual aid flows that are disbursed in re- sponse to a country’s progress in implementing a national poverty reduction strategy (PRS) or national development strategy (NDS) (Koeberle and Stavreski 2006). In fragile situations, budget support can be understood as financial assistance that addresses short-term needs in the aftermath of the conflict or supports a medium-term program linked to the recovery strategy; and is provided directly to a recipient country’s budget on a regular basis, using that country’s own financial management systems and budget procedures. As mentioned above, (policy-based) BSOs can be also channeled through MDTFs (e.g., the World Bank in the case of Afghanistan), notably in contexts where fiduciary risk is perceived to be very substantial. Budget support can be subject to high fiduciary risk in recipient countries with weak public financial management (PFM), as would be expected in fragile situations. In addition, policy- based budget support could increase the volatility of aid flows in countries where the policy environment is poor. However, evidence shows that budget support has been used effec- tively even in fragile situations with extremely weak fiduciary systems – as shown by the experiences of Central African Republic (CAR), Haiti and Afghanistan. In these cases, clear expenditure priorities and a strong government commitment to addressing institutional weaknesses reduced the risk to an acceptable level relative to expected benefits (World Bank 2005, Koeberle and Stavreski 2006). The OECD-DAC (2006), moreover, finds that budget • 6 Box 2. Definitions of Budget Support All definitions of budget support are constructed around funding country-led poverty reduction efforts by direct financial support to the country budgets. There are two types of budget support: general bud- get support (GBS) and sector budget support (SBS). GBS refers to an all-purpose contribution to the country’s budget. Assistance is typically provided in for- eign exchange, which is placed with the central bank to be converted into local currency and credited to the central government account (usually a treasury account) at the central bank. A government can use these funds to reduce or repay debts, as well as to secure or augment budget spending. SBS denotes different types of financial support that is combined with support for sector policy and insti- tutional reform. In un-earmarked SBS loan or grant proceeds go to the general budget. In earmarked SBS proceeds go to the budget of the sector ministry that conducts the dialogue and designs and implements the reform actions. In earmarked, traceable SBS, the proceeds are earmarked to specific sub-sectors or budget lines and expenditures and outputs are monitored. With earmarked SBS dialogue and condition- ality focus on expenditure items rather than sector policies. With un-earmarked SBS dialogue focuses more on sector policy. All the three organizations engage in general budget support. The EC and the AfDB can also provide un-earmarked sectoral budget support, although usually in non-fragile contexts. Moreover, on an excep- tional basis, the EC can use earmarked budget support (e.g. to reimburse expenditure ex-post), in cases of extreme fragility where the risks of engagement are particularly high. Source: World Bank 2006, World Bank 2010, AfDB 2010. support has played an important role in strengthening the economic and financial architec- ture of the state in post-conflict settings (e.g. in Rwanda). Sources of budget financing. Policy-based lending and grants are not the only instruments of assistance to the budget. MDTFs have grown into a significant source of support for rapid- ly scaling up donor assistance, particularly in high-risk, post-conflict (or post-disaster) situ- ations.2 They can be used as a source of financing for technical assistance, projects or BSOs that channel money indirectly to the budget, playing the role of intermediary between each donor and the recipient country. MDTFs can also fund programs outside of the government budget, in situations of weak capacity, and have played an important role in Afghanistan (See Box 3), Sudan, West Bank and Gaza, Timor Leste, Iraq, Sierra Leone and Indonesia. A re- cent review of post-crisis MDTFs concluded that MDTFs are more effective when providing sizeable resources to the budget rather than financing off-budget activities and small-scale projects (World Bank et al., 2007a). The World Bank has coordinated the mobilization of single- and multi-country MDTFs and currently administers many of them. It also provides resources to countries characterized by fragility through single and multi-country trust funds financed from IBRD’s net income. These funds are intended to complement the support provided through IDA, or finance coun- tries and purposes for which IDA resources cannot be used. At the AfDB, a country’s supple- mental allocation from the Fragile States Facility (FSF) can be used to support its recovery programs through MDTFs. By the same token, the EC can also channel resources through MDTFs administered by third parties (e.g. Vietnam and Afghanistan, both administered by the World Bank). The IMF does not contribute to MDTFs (on the approach of this paper to MDTFs, see again footnote 1). • 7 Box 3. The Afghanistan Reconstruction Trust Fund (ARTF) The ARTF is a partnership formed between the international community and the Afghan government to serve as a coordinated financing mechanism for the Government’s recurrent budget and priority re- construction efforts. Since its establishment in early 2002, 30 donors have contributed over US$3 bil- lion, making the ARTF the largest contributor to the Afghan budget. The ARTF’s Recurrent Cost Win- dow (RCW) finances operational and maintenance expenses. These funds enable the government to pay around half the civilian wage in all 34 provinces of the country. The process has helped build capacity in the Ministry of Finance by strengthening of the government’s PFM systems and ensuring some predict- ability to deliver essential services, such as education and health care, to the Afghan people. Disbursements are based on eligibility criteria that have been agreed with the government in line with the broader fiduciary framework for public expenditures. The RCW has disbursed around 100 percent of its allocation every year. In December 2008, ARTF donors and the Afghan government agreed on a new approach to support the government’s operating budget, the ARTF Incentive Program. The new approach establishes a phased decline in ARTF support for recurrent costs over the ten remaining years of the fund (through 2020). This approach supports the government’s revenue-raising efforts to attain eventual self- reliance – in conjunction with technical assistance and capacity building – and allows the ARTF to make a phased and predictable exit from the RCW. In addition it pro-actively supports the government’s eco- nomic reform agenda by empowering reformers to deliver on their programs. The program also ensures that the ARTF donors have a genuine stake in the performance of the Afghan government and in policy dialogue. Source: ARTF Website 2009. 2.2 Institutional approaches to the provision of budget support All three institutions undertake BSOs (including in fragile situations), except for the IMF.3 It is argued here that there are similarities in the approaches adopted by the three institutions, notably on what concerns the underlying rationale, eligibility criteria, and need for complementary action (e.g. technical assistance and capacity building) in concomitance with BSOs. However, differences also emerge concerning the design of BSOs, the criteria governing disbursement of funds, the use of conditionality, the assess- ment of progress and the expectations about the nature of donor coordination. These is- sues are tackled in the remainder of the paper, as they can pose barriers to a more coordi- nated approach to the provision of budget support in complex environments. Moreover, the treatment of risk remains somewhat underdeveloped for all donors, as discussed more at length in Section 3. Box 4 reviews the institutional approaches to the provision of budget support by the three institutions, according to their respective guidelines. Rationale and scope. All three institutions concur that budget support can be justified in countries experiencing credible post-conflict and political transitions (‘gradual improv- ers’). By contrast, in those characterized by conflict, prolonged crises or deteriorating eco- nomic governance, budget support may not be advisable. The rationale, notably in post- conflict settings, is articulated around the need for providing assistance to the processes of recovery in a manner that reinforces the centrality of the state. Emphasis is placed on seiz- ing a ‘window of opportunity’ for starting reforms and reviving the delivery of essential public services (and safety nets) by governments; avoiding deterioration of the economic and political environment that underpins stabilization; and supporting social stability, • 8 Box 4. Institutional Approaches to Resource Allocation and the Provision of Budget Support The allocation of resources to low-income countries takes place though the replenishments of the African Development Fund (ADF) at the AfDB, the International Development Association (IDA) at the World Bank, and the European Development Fund (EDF) and Development Cooperation Instrument (DCI) at the EU. The AfDB established a Fragile States Facility (FSF) in 2008, which provides additional assistance to ‘fragile states’ on a selective basis, through three windows: supplemental support; arrears clearance; and targeted support. IDA makes exceptional allocations to post-conflict and re-engaging countries, and can also support arrears clearance. The EC has no standing exceptional facility, but in the face of crises, reserves from the EDF and DCI can be mobilized to provide additional assistance. The IMF has recently established a Rapid Credit Facility (RCF). The RCF replaces the previous subsidized Emergency Post-Con- flict Assistance (EPCA) for low-income countries, and provides additional financial assistance to coun- tries in a wider range of fragile situations. The World Bank provides un-earmarked general budget financing, subject to the borrower’s own imple- mentation processes and systems, under the umbrella of Development Policy Lending (DPL). Develop- ment policy operations (DPOs) can be extended as loans, credits, or grants and are governed by op- erational policy OP/BP 8.60, which provides a unified framework while leaving room for customizing content and design to country circumstances. In 2005, the Bank produced a Good Practice Note for DPOs in Fragile States, which is being revised at the time of writing. The African Development Bank’s provision of budget support is governed by the guidelines on Develop- ment Budget Support Lending Operations (DBSL) and the Fragile States Facility (FSF). The latter indicate that FSF-financed operations can be implemented using any of the Bank’s instruments, including DBSLs. BSOs can be provided to post-conflict or transitional countries with governments committed to reforms but which have a weak or declining revenue base and where rapid response from the international com- munity is judged as crucial. The European Commission’s provision of budget support is financed by the EDF – for African, Caribbean and Pacific (ACP) countries – and the DCI. The guidelines on the Programming, Design and Management of General Budget Support (GBS) have been complemented by a Methodological Annex that deals specifi- cally with the provision of budget support in situations of fragility, based on lessons learned from past experiences. Some EU Member States also provide budget support on bilateral basis. The EC is currently working on a Green Paper on improving the delivery of budget support operations, including in fragile situations. Source: World Bank 2005, 2007, AfDB 2004, 2008, EC 2009. improvements in the security situation and the re-engagement of external partners. BSOs in fragile situations usually support key elements of nascent PRS/NDS and are delivered with complementary capacity building and technical assistance projects. In the absence of such programs, BSOs can assume the function of a broader ‘umbrella program’ until the government has articulated its strategy in a more comprehensive fashion. Eligibility criteria. The World Bank’s decision to extend DPL is based on an assessment of the country’s policy and institutional framework, including the economic situation, governance, environmental/natural resource management, and poverty and social as- pects. The Bank considers the country’s commitment to and ownership of the program against its track record. It also assesses the country’s institutional capacity and ability to implement effectively the program to be supported. The AfDB and EC require a NDS/ PRS; a viable macroeconomic framework including an on-track IMF program; strong • 9 leadership; satisfactory up-front assessment of the fiduciary environment; and, some track record on the PFM front. In practice, both the EC and AfDB have a flexible inter- pretation of these criteria in situations of fragility. The EC (2009) reformulates the above criteria as follows: some ‘ultra basic elements’ of a PFM system should be in place; a track record exists showing a positive trajectory in PFM reform; and a national policy, which addressees the challenges of fragility or instability, is defined or in the process of being defined. An assessment letter from the IMF is accepted in the absence of a Fund-support- ed program by both the EC and AfDB.4 Conditionality. There is broad consensus among the three institutions that core condi- tionality in fragile states should be limited in number and drawn from the PRS/NDS or PFM and procurement programs. This is also corroborated by the evidence that emerges in the retrospectives on budget support summarized in Annex 1. Conditionality should be results-oriented, focused on the most critical elements and realistic. This is due to the low levels of implementation capacity of recipient countries and the need for concentrat- ing on elements that are perceived as critical for strengthening the core functions of the state. However, there are differences in the ways in which conditionality is used across institutions. For instance, the World Bank supports result-oriented conditionality (that is, DPL is linked to outcome-oriented result frameworks, which are in turn linked to the NDS/PRS) but disburses against policy/institutional actions. The EC, by contrast, can in some cases use results as conditionality when it disburses against outcomes (see below). In these complex and unpredictable environments, moreover, conditionality is a tool that can be used to manage risks. Design of BSOs. The design of BSOs varies substantially across the three institutions, posing challenges for a more coordinated approach, notably on what concerns disburse- ment patterns. In fragile situations, the World Bank usually adopts an approach consist- ing of annual single-tranche operations, which may contain only tentative policy actions for follow-up operations. It gradually moves to a programmatic approach only when a medium-term program solidifies in country. The AfDB tends to deliver BSOs that have multiple-tranches in a single year, with a having greater control over resource use in re- sponse to government performance. The amount of tranches per operation is reduced as a country’s performance gradually improves. The EC uses a variety of approaches rang- ing from short-term operations based on fixed tranches at the early recovery stage (e.g. budget support for arrears clearance); to multi-years programs, in more robust contexts, which entail the gradual introduction of performance-based criteria through variable tranches. Disbursement criteria. Disbursement mechanisms also differ. The World Bank typically disburses against the fulfillment of prior actions, which are listed in each BSO (policy matrix). There are no effectiveness conditions, and disbursement normally occurs within 90 days of Board approval. Similarly, the disbursement of AfDB tranches depends on satisfactory performance against conditions precedent, which are usually drawn from national strategic documents and focused on PFM. Additional upstream conditions can be used, in extraordinary cases, prior to Board presentation. The EC disburses against the fulfillment of outcome-based general conditions, which demonstrate progress toward macro-economic stability, PFM reform, and key pillars of the PRS/NDS. In fragile situa- • 10 tions, specific conditions can be also used to address a particular concern or risk.5 Prog- ress against outcome-based specific indicators determines the level of disbursement of variable tranches – when they exist – usually linked to progress with respect to budget performance, e.g. disbursement rate of social spending, or PFM in general. Risk aspects. There is a general recognition within the three institutions that the po- tential benefits of BSOs in fragile situations often justify the risks associated with en- gagement, even though such benefits are not sufficiently spelt out in the documentation associated with BSOs. Risk aspects also tend to receive a relatively limited focus in the de- velopment and implementation of BSOs. Yet, at the same time, reservations are routinely put forward by some World Bank and AfDB shareholders, and a number of EU member states, on the opportunity of engaging with budget support in fragile environments. In these contexts, the three institutions tend to single out fiduciary risk as the most signifi- cant, but their approach to risk slightly differs.6 The World Bank argues that improved PFM may be an outcome, rather than a pre-condition, of support. The AfDB also acknowl- edges that PFM improvements can be obtained as result of policy dialogue and flow of funds through BSOs, but emphasizes both the need to assess fiduciary risk up-front and the use of fiduciary safeguards to prevent the misappropriation of funds. The EC high- lights the need to undertake a more comprehensive analysis of risk prior to engaging in budget support, covering the following aspects: political and policy risk, macro-economic risk, PFM risk, capacity-related risk, and donor coordination risk. Progress assessment. At the World Bank the monitoring of BSOs revolves around the completion of Implementation Completions Reports (ICRs), which evaluate the impact against outcomes, spelt out in result frameworks and linked to the key pillars of the NDS/ PRS. Similarly, the EC focuses on outcomes and maintains that the joint annual reviews of the NDS or PRS provide the best opportunity to assess the impact of BSOs. The AfDB relies on a focused operation-specific log frame which uses key indicators drawn from performance assessment frameworks (PAFs), which are also linked to the NDS/PRS.7 In better performing countries that benefit from programmatic lending, BSOs have been prepared in the context of multi-donor budget support frameworks (MDBS), aligned with PAFs. Some observers argue that PAFs have delivered mixed results, at times limiting the flexibility of each institution to adapt quickly its response to the client needs (World Bank 2009, 2009a, see Box 6). Others counter that these consideration do not necessarily apply to fragile and risky environments, but chiefly to better performing countries. Donor coordination mechanisms. All three institutions maintain that BSOs provide an opportunity to consolidate and coordinate external partners around a common, govern- ment-led program, like the NDS/PRS. This is particularly important during a transition, which is often characterized by significant increases in the volume of external assistance, a heavy policy agenda, and limited government capacity. The development of a common program may also reduce perceived risks, by increasing leverage on key policy issues and allowing greater sharing of information on fiduciary risk management. Yet, some differ- ences emerge. The EC advocates the elaboration of memoranda of understanding (MOUs), which typically contemplate carrying out joint diagnostic works (on risk, PFM, sectoral work), converging on conditionality, and putting in place systems for joint reviews of budget support with governments, through PAFs. The World Bank argues that MOUs are • 11 more efficient if such documents avoid complexity and rigid prescriptions: built-in flex- ibility facilitates donors’ responses in times of crisis. The AfDB uses joint MOUs where available, backed by detailed legal requirements in the loan or grant agreement for each operation. Suspension and exit. While multilaterals are less prone to interrupt budget support as a result of corruption or political risks than bilateral donors, the temporary suspension of budget support has been a reality in more than one country and for more than one donor. Yet, the identification of conditions under which such support would be scaled back, on account of less than satisfactory performance, is not articulated at great length in the guidelines governing BSOs. For the EC, breach of eligibility to budget support should in- form dialogue in order to identify measures needed to resume payments once programs (including with the IMF) have gone off-track. The AfDB reassesses eligibility criteria for each new BSO, which may be discontinued if criteria are not met. The World Bank main- tains that planned exits should address both the developmental and financing rationale for development policy assistance. BSOs may have to be suspended where program imple- mentation makes little progress or country conditions deteriorate significantly. Further- more, the World Bank would not disengage completely, unless the country accumulates arrears, and continues to undertake analytical work, standing ready to exploit possible windows of opportunity for re-engagement. 3. Budget Aid in Situations of Fragility: Emerging Issues This section identifies some emerging issues around the provision of budget aid in com- plex fragile environments that need to be addressed to improve the effectiveness of do- nor approaches. It is argued here that rather than viewing budget aid as simply a transfer of financial resources to the country’s budget, with a narrow focus on PFM (despite its importance), it should be considered as part of a package that consists of evidence-based policy dialogue, analytical work, technical assistance, capacity building activities, as well as financial transfers. This package should be more explicitly geared towards addressing the underlying causes of fragility. Doing so could help contribute to the prevention of future conflict – or the likelihood of conflict relapses – and strengthen the economic and financial architecture of the state. In addition, the risk elements surrounding the provi- sion of budget aid need to be analyzed in more depth and shared more widely amongst the three institutions. First, while the institutions’ Country Assistance Strategies (CASs) usually discuss how budget aid is integrated in overall assistance programs that aim at reducing poverty, the discussion on its potential role in supporting the transition from fragility to resilience (as defined in paragraph 6 above) is underplayed and in need of strengthening. This applies both to strategic documents like CASs and to the program documents (PDs) of BSOs. In the documentation associated to BSOs, the argument could be strengthened by highlight- ing the role that budget aid can play in: stabilizing the macro-budgetary framework and allowing the state to carry out basic functions (e.g. paying the salaries of civil servants), to cement its legitimacy and contribute to political stability; supporting the longer-term endeavors of peace and state-building; and contributing to strengthening PFM capacity by channeling aid through national systems with associated dialogue and technical assistance. • 12 Second, the treatment of risk aspects should be expanded substantially. At the moment, BSOs tend to have short, standardized sections on risk and risk mitigating factors. The typology of risk, the analysis of different categories of risk and the trade-offs that exist between them could be given more prominent attention. Transparent and shared analy- sis of risk (including fiduciary risk), and appropriate mitigation measures, could be better coordinated across donors to pool risk. In addition, the risk of not engaging should be discussed more explicitly and set against the benefits that can be reaped by intervening and successfully stabilizing a country, including the positive regional (and global) exter- nalities that may result. Third, greater attention needs to be paid to the choice of aid in- struments and associated trade-offs between aid predictability and aid effectiveness, by highlighting the complementary nature of different instruments and sources of financing to the budget. 3.1 Highlighting the advantages of budget aid Stabilize the macro-budgetary framework. In combination with other aid instruments, budget aid plays a key role in underpinning fragile social and political processes through supporting macro-budgetary stability. Due to their macro-economic vulnerability and in- ability to raise revenues from regular taxes, many fragile situations struggle to pay for basic state functions, clear external and domestic arrears and resume debt service. This, in turn, often undermines security (inability to pay military wages and thus maintain the loyalty of the security apparatus), state services (non-payment of wages for civil ser- vants), and related social services. Political and social unrest might ensue as a result, as governments are unable to plan and maintain the social contract by fulfilling the core functions expected of them in terms of the delivery of basic public services and main- tenance of peace and security. Moreover, constant cash pressures lead to poor budget planning and heavy reliance on firefighting in the ministry of finance, which undermines efforts at reforming PFM systems. Providing a reasonably stable flow of resources to the budget that allows the government to stay current with its salary payments and external creditors entails a different focus and objective than those in well-performing, general budget support countries, where it is often more about scaling up.8 The policy dialogue underpinning budget aid also needs to focus more squarely on strengthening fiscal practices, including building up domestic revenue generating capacities. Mobilizing domestic resources is important for the longer- term goal of achieving a sustainable fiscal position and reducing dependency on budget aid. Building tax institutions can also contribute to sustainable delivery of basic public services, which is crucial for stability and securing peace (IMF 2005). Measures here in- clude: providing technical assistance for drafting tax codes; setting up special tax admin- istration units, with a marked emphasis on pragmatism; linking the provision of aid to progress in domestic revenue performance; helping to curb extra-legal revenue extrac- tion (e.g. illegal taxes on trade); and reducing tax exemptions for post-conflict aid (Boyce 2009). Support peace- and state-building. Recent emphasis has been placed on the need to integrate efforts geared at promoting the endeavors of peace- and state-building (OECD- • 13 DAC 2009a, DFID 2009). Peace building encompasses measures in the context of conflict- affected or post-conflict situations for the explicit purpose of preventing violent conflict and promoting lasting and sustainable peace (OECD-DAC 2005, 2009). State building is seen as purposeful action to develop the capacity, institutions and legitimacy of the state in relation to an effective political process for negotiating the mutual demands between state and societal groups (OECD-DAC 2009). Violent conflict undermines the effective- ness and legitimacy of state institutions. At the same time, it is unlikely that peace will be sustained unless at least some minimum state functions are performed. Peace and state building are therefore the critical goals for stabilizing fragile situations, and their achieve- ment creates a desirable environment in which economic development can be pursued. A state-building agenda that focuses on rebuilding key institutions can support the growth process and enhance the absorptive capacity and management of aid flows of the recipient country, and have an impact on poverty reduction through service delivery. At the same time, the legitimacy of the state can be boosted by improving human secu- rity and the rule of law, which, by focusing on non-violent means to mediate disputes, contributes to reducing the chances of conflict. Legitimacy also depends on improving domestic accountability. This includes diversifying accountability mechanisms, both non- government (civil society) and the formal institutions of accountability (e.g. parliaments) and improving transparency in key areas of decision-making and revenue collection (e.g. management of natural resources). It is also important to focus on job creation, notably for unemployed youth. Strengthening peace and state-building along these lines is likely to call for greater resources beyond those needed for direct poverty reduction. Strengthen state capacity. The focus on improving state effectiveness requires short- and long-term technical assistance to build national capacity for responsive and accountable government administrations and service delivery within the public, non-governmental and/or private sector, through nationally-executed projects that complement the budget aid being provided. Bypassing the government by utilizing alternative delivery mecha- nisms does not strengthen state capacity, but rather shifts it, and may actually damage what little pre-existing capacity there is. Ultimately, capacity-building takes place best through gradual ‘learning by doing’. Except for the few large and lumpy investments, international assistance provided through budget aid would allow newly formed govern- ments to gradually exercise their most important functions: setting priorities, making hard choices, and managing expenditures. Supporting the government’s budget instead of financing costly discrete projects would be more conducive to capacity building over time. According to Francois and Sud (2006), there are additional advantages of budget over project aid. First, budget aid would enhance the prospect of donor funds reaching beyond the capital city into the rural areas thereby contributing to redress unbalanced develop- ment, itself a potential cause of fragility. It is argued that the government, by working through its own systems, would have greater ability to reach the rural poor. Donors have been ineffective in doing this for a number of reasons, including problems with ensuring security for their personnel. Second, the provision of budget aid would reduce the heavy ‘foreign footprint’ of donors that can lead to distortions in economic activity and labor markets (e.g. high salaries in project implementation units) and come to be resented • 14 before too long in most fragile situations, notably in countries emerging from conflict. Third, it would reduce demands from fragmented aid projects and processes on already limited government capacity, much of which gets diverted to fulfilling individual donor requirements. 3.2 Managing risk in the provision of budget aid The review of the evidence that informs this paper suggests that a more comprehensive analysis of risk elements in the documentation associated with both CASs and BSOs would be beneficial. Risks are generally higher in fragile situations. On the basis of solid political-economic analysis of the situation, donor coordination and the issue of collec- tive impact become very important to enable institutions to make informed decisions about whether to intervene in this manner in contexts where the control over outcomes is limited. There are also tensions and trade-offs between outcomes that agencies aim to promote (or prevent) by their interventions, such as stability, the absence of violence, and stronger, better governed state institutions; and those that are unintended adverse con- sequences of their intervention, such as the risk of fuelling corruption or of being unable to account for the use of funds disbursed. At one extreme, concern about ‘unintended consequences’ may lead an agency not to intervene at all; or it may cause it to intervene in ways that are highly conservative, restricted and inflexible. Finding the right balance is key, but this balance is likely to change over time thus requiring regular re-appraisal. Risk management. Risk management includes the three pillars of up-front risk identi- fication, mitigation and monitoring. Risk can be expressed as the probability of future harm – where ‘harm’ is taken to mean any undesired outcome. Risk management is about the attempt to manage the future by taking action in the present.9 Four scenarios can be depicted: (i) outcomes of pure chance where no one party can influence the outcome; (ii) situations where multiple factors influence the outcome and the risk taker may have little or no control; (iii) situations which are risky but where one can significantly influence the outcome; and (iv) situations where the outcome is entirely within one’s control. In ‘standard’ cases, most aid interventions tend to fall into category (iii). In fragile environ- ments, the ability to influence outcomes is often more limited. Here, it is possible that events would fall into category (ii), which is the most risky (OECD-DAC 2010). In these cases the question is: given the multiple risk factors, what measures can be adopted to minimize/mitigate identified risks and maximize the chance of the desired outcome(s) occurring? Political risk. BSOs use government systems and channel funds directly to the budget and therefore they can have political implications. These may at times be in line with desired outcomes, e.g. helping to reinforce state legitimacy in the presence of a leader- ship that is firmly committed to reform. However, if the political implications are not carefully assessed and monitored, BSOs may also have unintended negative impacts on national politics and stability. For instance, budget support payments prior to an election might send wrong signals on donor support and provide a competitive advantage to the incumbent political party. More attention to the political economy of fragile situations is therefore a sine qua non for assessing in country political risk. Political risk affects • 15 bilateral and multilateral donors in different ways. The latter tend to be less exposed to it and are able to be more predictable for partner governments, but should nevertheless undertake adequate ex ante political economy analysis to ensure the resources support legitimate regimes. Economic risk. While budget support provides a critical ingredient to improving eco- nomic and social outcomes, its provision needs to be consistent with the broader macro- economic environment. Specifically, a credible macroeconomic framework should be in place, preferably supported by some form of IMF program engagement, which can help mitigate the risk that the additional resources would be ultimately wasted through incon- sistent fiscal, monetary, and exchange rate polices. In addition, it is important to consider absorptive capacity, both at the macro level – where spikes in budget aid may cause infla- tionary and currency pressures – and at the micro level – where additional spending may be inefficient. Country specific analyses of the absorptive capacity of recipient countries are seldom carried out by the three institutions. Coordination with the IMF is important to make sure that budget support financing is factored into the macroeconomic program. Fiduciary risk. One common characteristic of fragile situations is their particularly weak institutional capacity, which gives rise to the opportunity for mismanagement of funds. Weaknesses in bureaucratic systems leave them vulnerable to corruption and incompe- tence, thereby raising the risk of leakage. Moreover, in most fragile situations, the social contract between the state and society is at best weak with civil society ill-prepared to play a watchdog role over government performance. These characteristics can increase the propensity for corruption, which is further compounded when the regulatory sys- tems to restrain this tendency are limited. In these situations, donors are particularly cautious in committing their resources through BSOs. Fiduciary risk analysis drawing on shared diagnostics like Public Expenditure and Financial Accountability (PEFA) reviews help establish the information to make informed decisions. The choice of aid instrument has important implications for the degree of exposure to risk. Using balance of payment support tied to government import expenditure may al- low better monitoring of expenditures and is therefore a type of budget aid, which is less susceptible to fiduciary risk. The other alternative is to channel resources through MDTFs, with disbursement governed by strict and specific fiduciary arrangements. Another short- term option is to put in place fiduciary safeguards, which can be used to address specific risks where the provision of budget aid to strengthen budgetary systems is important. These safeguards can be useful in protecting donor funds from misappropriation but come with additional costs both in terms of administrative costs but also in terms of the use of parallel systems which could undermine national system development (e.g. procurement or audit). As such, the decision to use safeguards and choice and design of appropriate safe- guards should be considered carefully to suit country circumstances.10 Yet, the attractiveness to partner governments of budget aid makes it particularly useful for establishing the incentives needed to secure their commitment to tackle corruption and PFM reform. The policy dialogue can focus on issues that are critical for improved gover- nance. Decisions over whether to embark on budget aid, and take on the risk of misap- propriation, need to be thus weighed against the benefits of substantive engagement with • 16 government for policy leverage on sensitive but critical anti-corruption and PFM reforms, which can be hard to achieve otherwise or through project aid. In considering whether the benefits outweigh the risks, donors need to keep in mind that although ring-fenced/proj- ect approaches go some way to giving greater assurance that funds will not be misused, they will not address the underlying systemic problem and, because of fungibility, project finance is ultimately no less susceptible to fiduciary risk. Risks associated with not engaging. The challenge in fragile situations is to find appropri- ate ways of operating that keep the risk within acceptable levels but do not themselves frus- trate the purpose of engaging. Most of the time risk analysis is geared to finding ways to reduce or eliminate identified risks, particularly risks perceived to be faced by the investing donor. Given that there is a perception that budget aid (notably budget support) is risky be- cause the resources transferred to government budgets are fungible, an assessment of the risks of leakages and potential misuse of funds is often a deciding factor in whether to en- gage or not. On the other hand, the knock-on effect on state effectiveness, macroeconomic stability and social and political volatility means that budget aid has a potentially critical role to play in breaking the propensity of a country to fall back into instability and conflict. In the context of macroeconomic crisis (such as the recent global economic downturn), the risks of not engaging with fast disbursing budget support are magnified, putting basic state functions like service delivery and maintaining security at risk. These aspects could receive more prominent attention in the documentation associated with CASs and BSOs. Towards a typology of risk. A more consistent and comprehensive risk analysis can help the institutions make truly informed decisions and may strengthen the case for budget aid interventions in difficult environments. In particular, it is argued that political risk receives little attention (hence the calls to focus more on the ‘political economy’) with some insti- tutions tending to concentrate principally on financing risk. This reinforces the view that risk analysis is mainly about institutional risk and that risk to peace, stability and reform in the partner country of providing or not providing budget aid is not given equal weight. To improve the quality of risk analysis, a typology of different types of risks faced in fragile situations could be developed, on a country-by-country basis, as the local context would change the relative weighting of different types of risks (See Box 5). 3.3 Addressing the trade-offs between aid instruments, aid predictability and aid effectiveness Budget aid provides the potential for increased predictability of funding. However, there are still degrees of uncertainty – deriving both from the design and disbursement pat- terns of BSOs as well as government performance – which may make it difficult for na- tional governments to plan and manage their budget with some degree of certainty. The issue becomes particularly acute when policy-based budget support payments are sus- pended or not approved, with potentially significant knock-on effects for political and social stability. The opportunity for policy-based support – which would finance expendi- tures that are more discretionary, such as expansion of the road network – are more ame- nable to funds disruption where a government fails to meet a key condition. However, at the same time, there are some core expenditures that must be protected, e.g. salaries and • 17 Box 5. Typology of Risk 1. Political and institutional risk. Risk of inadvertent fuelling of conflict through aid, e.g. becoming part of war economy or parallel economies, rendering aid recipients vulnerable; risk associated with unin- tended political bias in effects of aid, e.g. impacts on power-sharing and transitional arrangements; risk of favoring particular groups (e.g. dominant, entrenched business élites); risk associated with incoherent political strategies between donors. 2. Economic risk. Risk of macroeconomic instability (e.g. inflation, price volatility, currency fluctuations) that could erode the benefits of budget support; and risk that budget support itself could generate price or currency pressures; economic risks and opportunities of using local versus international procurement systems; risk associated with fiscal policy: lack of control over government spending, uncertain or limited tax revenue. 3. Financial risk. Risk associated with aid delivered through country systems given lack of local capaci- ties and capabilities (credibility of the budget, efficiency of budget executing processes and controls, reli- ability of accounting and reporting, quality, comprehensiveness and timeliness of external audit controls, procurement risk); associated risk of corruption, linked to lack of effective oversight or political checks. 4. Technical and operational risk. Operational security risks (e.g. safety of staff); risk of program failure or unintended effects associated with inadequate contextual understanding, flawed needs assessment, inadequate or unreliable data; risk associated with lack of capacity among donors/agencies to manage and monitor implementation; risk associated with lack of incentives or devolved authority to take and manage risks. Source: OECD-DAC 2010. some essential operating costs of government to avoid the deterioration of institutions. In the case of these essential functions, it is highly desirable for there to be a predictable flow of budget aid. However, trying too hard to ensure full predictability of policy-based BSOs could detract from the potential leverage and impact that budget aid can have in terms of reform and state-building actions. Medium-term predictability may be futile in fragile environments, but rapid decisions about the withdrawal of support in reaction to potential short-term events should be avoided. Similarly, the imposition of onerous audits regimes designed to identify infraction and not to inform governance reform often undermines the ideal of achieving a constructive partnership, besides compounding the uncertainty of disburse- ments. One under-explored option is to consider the complementary functions of policy- based BSOs and support to recurrent expenditures through alternative instruments and sources of financing – including through MDTFs – notably in cases where the state is unable to raise sufficient revenues to cover essential expenditures, and to avoid implo- sion. In an instance of policy reversal, donors could fine tune the mix of options at their disposal (including project aid) to avoid negative consequences for countries resulting from the delay or withdrawal of disbursements through one mechanism.11 By the same token, it is important to consider the trade-offs between an improved, more coordinated donor delivery of budget aid and the need to avoid global green light-red light signals on aid disbursements. • 18 4. Selected Evidence on Budget Aid: Retrospectives and Case Studies This section summarizes the results of the retrospectives carried out by the AfDB, EC and World Bank on policy-based lending, as well as the findings of four case studies – Burun- di, Central African Republic (CAR), Guinea Bissau and Sierra Leone – where two or more of the participating institutions undertook simultaneous BSOs. The aims of this section are to examine this selected evidence; compare it with the critical discussion conducted in Sections 2 and 3; and further ground the recommendations of this paper, discussed in Section 5, in country-specific experiences. 4.1 Evidence emerging from budget support retrospectives12 What emerges from the retrospective is that decisions to engage in the provision of budget support have been essentially political and characterized by the undertaking of stand-alone operations. In most cases, BSOs were not planned initially but became a re- ality in several countries because of the opening of windows of opportunity. The World Bank’s retrospective, for instance, finds that a programmatic series of BSOs in fragile situations is not usually contemplated (in accordance with its guidelines, see above) even though a series of stand-alone, single-tranche operations have subsequently ma- terialized in some countries. Hence there seems to be little certainty regarding predict- ability of support. While this approach may be understandable as the World Bank seeks to establish or continue engagement in a high-risk environment, it nevertheless may result in a lost opportunity for tying budget aid into broader and more coherent pro- gram of assistance, ex ante, and also help recipient countries’ governments to prioritize a medium-term reform plan. The EC retrospective finds that the eligibility criteria have been interpreted in a flexible way and with an eye on the direction of the transition (‘good willingness for change’) and potential risks and benefits. Conditionality has been streamlined by all institutions. The focus of BSOs remains overwhelmingly concentrated on economic governance, con- firming the overarching concerns with high fiduciary risk. By way of example, the AfDB retrospective finds that 74 percent of BSOs disbursement conditions are related to eco- nomic and financial governance. Of these conditions, 73 percent focus on PFM and 17 percent on economic management, including the business and investment environment. All retrospectives also point out that the rationale for budget aid could be strengthened. The links between the deployment of BSOs, the process of transitioning out of the crisis into resilience, and the impact that donors can have in contributing to the achievement of such outcomes, is not discussed at length in the documentation associated with these operations. The World Bank retrospective points out that the main risks that are commonly men- tioned in BSOs include the reigniting of domestic conflict/unrest, political risk (lack of political capacity/will to undertake reforms), capacity constraints, fiduciary risk (notably corruption), macroeconomic or fiscal instability (the inability to raise domestic revenue), • 19 and, in some cases, the threat of regional conflict spilling over the recipients’ borders. However, fiduciary issues are often singled out as the most prominent across the three in- stitutions, explaining the heavy focus of BSOs and supporting operations on PFM, good governance and transparency. Capacity constraints are usually addressed by complemen- tary capacity building programs. The World Bank retrospective maintains that the docu- mentation associated with BSOs is unduly optimistic regarding the ability to work around weak country capacity: follow-up operations tend to recognize that capacity development was less than anticipated in earlier rounds. All retrospectives confirm that stop-and-go patterns characterize the delivery of budget support, undermining its effectiveness. The AfDB retrospective reports progress in im- proving the speed of disbursement but notes that, even so, significant amounts of budget support funds have not been disbursed on time. Clearly, disbursement delays are often caused by poor performance in the government reform program, which in turn delays ap- provals.13 All retrospectives also highlight the importance of donor coordination in frag- ile situations. The EC retrospective maintains that donor coordination in the provision of budget support is essential, and emphasizes the useful role that MOUs and PAFs can play in the elaboration and assessment of progress achieved (or lack thereof). The World Bank, however, cautions against the rigidity that these tools can introduce, and argues that flexibility should prevail to allow donors to react to risks and respond rapidly in times of crisis or changes in country conditions, according to their operational guidelines (Box 6). 4.2 Evidence emerging from selected case studies In the preparation of this paper, a number of case studies were conducted on the pro- vision of budget aid in Burundi, CAR, Guinea Bissau and Sierra Leone (Horton 2009). These countries were, or are, emerging from severe internal conflict but have managed to remain in relative peace, albeit fragile (i.e. Guinea-Bissau, CAR) for the last few years. On aggregate, what emerges from this evidence corroborates the findings of the retrospec- tives and is relevant to the discussion conducted in Section 2.2, as well as the critique leveled in Section 3. Absorptive capacity and amounts of budget support. The case studies confirm that is rare to come across systematic assessments of how much assistance can be effectively absorbed by recipient countries. These important tasks are not undertaken on a system- atic basis by the three institutions, and only occasionally by the IMF (e.g. Guinea Bissau, Sierra Leone). The case studies also highlight that the amounts of budget support vary significantly from country to country. CAR – which is usually considered an ‘aid orphan’ – has received, on average, around 2 percent in budget support, as a proportion of GDP, during 2001-2008. Over the same period, budget support was around 13 percent of GDP in Burundi, nearly 9 percent in Sierra Leone and nearly 6 percent in Guinea-Bissau. It is unclear whether such differences can be explained by the relative size of country financial needs or by political considerations (or emerging windows of opportunity), or both. • 20 Box 6. World Bank’s Emerging Position on MOUs and PAFs MOUs. MOUs require agreement on a common policy framework, but donors’ objectives and disburse- ment requirements vary. The World Bank’s policy (like that of the AfDB) is to disburse directly into a country’s budget on the basis of prior policy reform actions; at the EC, the policy is to disburse against outcomes/targets. Moreover, some donors can earmark and channel their sectoral budget support for the financing of specific sector work plans and budgets, while others (like the World Bank) cannot. MOUs also face other challenges. For example, the mandate of bilateral donors may require them to include po- litical governance conditionality (such as conducting free and fair elections) with their financing, while multilateral agencies may be prohibited by their Articles of Agreement from including political consider- ations in their decision-making. In many cases the solution has been to include language explicitly recog- nizing that its own legislation or charter will bind each donor, but these solutions often come at the cost of time and good will. The process of agreeing MOUs is more efficient if the documents avoid complexity and rigid prescriptions. Built-in flexibility facilitates donors’ responses in times of crisis. In some cases, donors have agreed on the ultimate objective of the operations but have accepted the fact that differing institutional policies do not allow complete harmonization of donor procedures. In others, MOUs have used flexible language accommodating differences across donors, or have included special ‘exception’ clauses, excluding one or more donors from a particular clause or approach (e.g. Sierra Leone, Tanzania). PAFs. The institutions have made significant progress in harmonizing their DPLs through Multi-Donor Budget Support (MDBS) arrangements. This has contributed to progress in reducing transaction costs to governments, with donors working more closely together. This process has also increased coordina- tion in the definition and choice of reform actions and results to be supported and monitored. Common PAFs have been established in place of overlapping and occasionally contradictory program matrices. Yet, MDBS arrangements are posing new challenges for both donors and clients. In Benin, the government suggested that more flexibility should be introduced in the PAF to adjust disbursement conditions to changing circumstances. In Tanzania, government officials maintained that the MDBS framework fails to represent a full alignment between the donors and government interests and priorities. In Uganda, government officials criticized the outcome-focused approach of the PAF, pointing to the need for do- nors not only to monitor results but also to assist in identifying the policies and institutional reforms to deliver those results. In some cases, it has been difficult to adjust the content of PAFs to reflect new structural and medium-term issues that have emerged (e.g. food and global financial crises). The MDBS arrangements in Mozambique involve over 70 sector working groups, placing a large burden on donors and Government alike. Source: World Bank 2009a. Underlying causes of fragility and risk. Confirming the findings of the retrospec- tives, there is little evidence emerging from the case studies to suggest that BSOs have been thought of as a key element of a package that can be deployed to strength- en the transition from fragility toward resilience, however non-linear that path may turn out to be. In the documentation associated with BSOs, not enough attention has been given to showing that budget support in fragile situations can potentially gen- erate even greater benefits than in non-fragile contexts. There has been lip service but no explicit focus on the causes of fragility, and the reforms/changes needed to improve key institutions and service delivery, beyond PFM, which are needed to build an effective state and strengthen its legitimacy vis-à-vis its citizens and help restore trust (e.g. intervention in the law and justice sectors, at the rural/community levels). Risk analyses have tended to focus on risk to the donor institutions rather than on a comprehensive discussion of risk aspects based on a risk typology. • 21 Disbursement trends. Alignment with national systems has been relatively weak. Often, funds have been disbursed in the third or fourth quarter of the country’s fi- nancial year, rather than in the first or second quarters, hindering effective budget management. There have been delays with respect to planned disbursement dates and, at times, the amounts effectively disbursed have been significantly less than 100 percent (this applies in particular to the EC variable tranches). Delays in disburse- ments have required countries to either reduce current expenditures relative to bud- geted amounts or to maintain expenditure by obtaining additional financing. Mobi- lizing additional financing from the Central Bank or loans from commercial banks have been common (Burundi, CAR, Guinea Bissau and Sierra Leone) but domestic arrears were also accumulated in some cases (Guinea Bissau). In this way, slow dis- bursement has been costly in terms of higher interest rates, had a negative impact on debt sustainability and reduced the availability of credit to the private sector. Conditionality and program monitoring and evaluation. In line with the guidelines and best practice, all three institutions have reduced the number of conditions in BSOs, and have moved toward a common PFM matrix, in some cases a shared PAF, from which conditionality is drawn. There are still differences in terms of the na- ture of conditionality across institutions – which reflect the differences in the design of BSOs discussed above – with the EC maintaining its outcome-based approach in less fragile environments, and the World Bank and AfDB maintaining their focus on prior actions/conditions precedent. Overall, programs for M&E have been relatively neglected in all countries, which implies a risk that M&E will just continue to focus on the assessment of conditionality – and selected longer-term outcomes specified in result frameworks linked to key elements of NDS/PRS – rather than on measur- ing progress against key building blocks in the path toward stability and resilience, including progress towards peace and state-building through the development of in- termediate indicators. Donor coordination. While cooperation is improving and getting more structured (e.g. MOUs have been signed in all countries), friction can still occur. On PFM re- form, agencies have differed on priorities in part as a result of preferring one diag- nostic tool over another. When poorly managed, these disagreements can lead to delay in organizing support around key policy reforms. On disbursements linked to shared PAFs, tension has emerged in instances where delays have been due to block- ages imposed by one institution, but not agreed with by others. Thus, the discussion on MOU and PAFs needs to be nuanced according to the specific country context (Horton 2009, see Box 6). 5. Improving the Effectiveness in the Provision of Budget Aid in Fragile Situations Based on the discussion conducted in this paper and the review of the evidence that emerges from retrospectives and case studies, this section proposes some recommenda- tions. These recommendations are geared at improving coordination and risk manage- ment in the provision of budget aid to fragile situations amongst the three institutions. In light of the multiplicity of approaches in each institution and progress to date in ad- • 22 dressing some of the issues raised here, the recommendations proposed in this section may not apply uniformly to all partners. By the same token, given the great diversity of fragile situations, country-specific assessments should ultimately guide the decision with respect to the timing of engagement, choice of instrument and source of financing. The section concludes with some suggestions concerning next steps and areas for further research and enquiry. 5.1 Key recommendations I. Strengthen donor coordination. Consultation on this paper has shown that there is both a high degree of commonality of intent and a willingness to cooperate among the three institutions. Despite some divergence in operational policies and practices, this is not so great as to create real barriers to achieving a more coordinated approach to the provision of budget aid. Improved coordination seems particularly relevant for reaching informed deci- sions to engage, choosing the appropriate instruments to intervene, coordinating technical assistance and capacity development plans, and sharing information on risk and political economy analyses. Coordination between the institutions at both headquarters (HQ) and country levels can be strengthened, by having more interactions between country teams (including with management and partner country governments). There is room to reduce transaction costs by finding non-legalistic, non-bureaucratic ways to strengthen working- level dialogue and partnerships amongst institutions. It is important to coordinate with the IMF to ensure that the budget aid is anchored in a credible macroeconomic framework. Specific Recommendation: Ensure better coordination between institutions at both HQ and country levels at critical stages of the budget aid cycle and find non-legalistic, non-bureaucratic ways to strengthen working-level dialogue and partnerships. II. Strengthen the rationale of budget aid in fragile situations. Budget aid can play an important role in stabilizing the macro-budgetary framework and strengthening state ca- pacity by using national systems. In the documentation associated with CCASs and BSOs there should be more reference and attention paid to the increasingly recognized role of budget aid in supporting the longer-term enterprises of peace and state-building. This may require developing intermediate indicators to be included in results frameworks and used to measure progress towards these long-term endeavors.14 The institutions could therefore emphasize more explicitly how budget aid fits into overall assistance programs aimed not only at reducing poverty but also strengthening the transition from fragility towards re- silience. The institutions also need to recognize more explicitly the prospective national, regional and global benefits of budget aid, including its capacity to support political and economic stability. In addition, there may be merit in exploring the ways in which civil so- ciety and parliaments can be more involved in the design, implementation and monitoring of key elements of BSOs, to strengthen the relationship between the state and its citizens. Specific Recommendation: Articulate how budget aid can support the transition toward resilience (macro-budgetary and political stability, peace- and state-building, human security, regional and global dimensions) in combination with other aid instruments. • 23 III. Improve knowledge of the political economy. Decisions on the provision of budget aid in fragile situations also need to be informed by an understanding of the complex and highly political nature of (peace- and) state-building processes. It is insufficient for the engagement to be based simply around the delivery of short-term technical solutions that provide fiscal support, build capacity and repair weak and damaged institutions. Acquiring an understanding of the complex social and political interactions at work in these contexts, through an historical and cultural analytical lens, is of fundamental im- portance in considering the risks of budget aid. Such political economy analysis requires an understanding of the underlying and often multiple causes of past conflicts, the driv- ers of élite behavior and the politics of exclusion, the regional and global influences, and the expectations of the state by its citizenry. This analysis should be iterative involving continuous monitoring and assessment of the inevitable dynamism of the post-conflict and/or fragile situation. Specific Recommendation: Undertake/commission/share iterative political economy analysis (e.g. political stakeholder analyses) and use findings to make more informed decisions to engage in budget aid, manage risk and focus policy dialogue with partner countries. IV. Undertake a more comprehensive analysis of risk. The current emphasis on risk to the donor institutions (notably fiduciary risk) should be more balanced against the risk to state stability and the pursuit of nationally-led reform of not providing budget aid. The application of a risk typology, the discussion of management of risks and risk mitigation could be further expanded in the documentation associated with budget aid. Working more closely together to pool risk analysis and management information could poten- tially help spread and therefore reduce the risk faced by each individual donor. In riskier contexts, this is a critical source of added value resulting from improved coordination of approaches. Progress towards results should be also highlighted and set against the per- ceived risks. In all case studies benefits have been realized in terms of stronger country ownership, greater emphasis on medium-term results and national development objec- tives, improved PFM, greater promotion of fiduciary accountability, and lower transac- tions costs for government and donors alike. Specific Recommendation: Undertake/commission/share risk analyses, adapted to the specific characteristics of a fragile situation, and use results to improve the coordination of approaches in the provision of budget aid. V. Consider non-traditional focus areas. The design of programs financed by budget aid could consider focusing on non-traditional areas (and the sub-national level) where appropriate, while preserving selectivity and the need for prioritization. The focus on macro-budgetary stabilization, PFM, and the delivery of basic social services is warranted and necessary. But the institutions could pay more attention to other areas that have been relatively neglected to date but could go a long way toward strengthening the le- gitimacy of the state and maintaining political stability. These include: strengthening the capacity of the state to generate domestic revenues; focusing on sectors that are key to re-establish trust between the state and citizens, such as security sector reform15, rule of law and justice; focusing intervention at the sub-national level with a view to redressing • 24 specific regional/group grievances (horizontal inequalities); and fostering the creation of decent, sustainable employment opportunities. Doing so does not mean broadening the scope of programs, but addressing innovative areas while applying the same principle of selectivity. Specific Recommendation: Consider non-traditional areas that are important to fostering stability and the legitimacy of the state (e.g. security sector reform, rule of law and justice, job creation), while maintaining selectivity, including at the sub-national level, as appropriate to country contexts. VI. Address issues of aid predictability and effectiveness. National efforts to restore peace, rebuild the state and reduce poverty are likely to require external budget aid to be provided in a predictable manner. While multilaterals are considered less vulnerable to fiduciary and political risks than bilateral donors, challenges remain in the quest for pre- dictability. The institutions could consider more systematically the possibility of providing (programmatic) budget support in situations of fragility, ex ante. Doing so may help tie budget aid into broader and more coherent program of assistance and also help recipi- ent countries’ governments to prioritize a medium-term reform plan. While country per- formance may vary and therefore affect the ability to provide predictability of resources through performance-based lending, more attention can be paid to combining different instruments/and sources of financing that can be deployed to support the budget, thus al- lowing the state to fulfill its essential functions. Support to recurrent cost through, for in- stance, MDTFs, could complement more traditional means of aid delivery and could help avoid threats to state stability, in the event of policy reversals delaying budget support. Specific Recommendation: Enhance aid predictability by achieving a balanced mix of instruments/sources of financing to the budget, and moving toward programmatic support where possible with a series of single-tranche operations. VII. Improve M&E systems. Attention tends to be paid to joint donor M&E systems as well as support to national systems. M&E systems tend to focus on the BSO matrices and result frameworks linked to key elements of the PRS/NDS. Measuring progress toward peace- and state-building objective remains an elusive yet important task. This is most likely due to the difficulty of establishing links between BSOs and the programs they support with such long-term and sometimes amorphous objectives. Yet, the OECD-DAC principles for good international engagement in fragile situations (2006) and the Accra Agenda for Action (2008) recommended that, at the country level, donors and develop- ing countries should work out and agree on a set of realistic peace and state-building objectives that address the causes of conflict and fragility. It seems therefore sensible to develop M&E systems that are capable of capturing progress (or lack thereof) in these ar- eas. Intermediate indicators could be developed here and included in result frameworks, focusing, for instance, on inter and intra group grievances, horizontal inequalities and other relevant dimensions. Specific Recommendation: Develop peace and state-building indicators to gauge the contribution of budget aid toward these objectives and include intermediate indicators in result frameworks to measure progress towards these long-term goals. • 25 5.2 Areas for further work and next steps The analysis conducted in this paper has made an initial contribution to the discussion of the complex issues surrounding the provision of budget aid in fragile situations arguing for greater coordination of approaches between the participating institutions. Looking forward, there is also a need to strengthen cooperation in concrete ways, to ensure that interventions are better thought out and geared at providing predictable and coherent support in countries characterized by weak capacity and volatile environments. In this respect, this paper highlights the following areas for further work: c Review institutional policies, procedural guidance and implementation tools with a view to incorporat- ing some of the recommendations articulated in this paper, as relevant. This option can be taken up by each institution, independently, following consultation with senior management. c In the context of making changes to practice in the three agencies, seek stakeholder views on the rec- ommendations in this paper to help further refine and prioritize the issues. c Increase cooperation amongst the institutions by establishing operational mechanisms that better support the harmonization of modes of delivery of budget aid, with particular attention paid to the need for achieving greater coherence around the decisions to engage in this manner, the choice of aid instruments, the disbursement of funds and assessment of progress. c With a view to increasing the pooling of risk amongst the institutions, seek to develop a common risk typology that can be applied to the provision of policy-based budget support lending and grants, including those provided through MDTF’s or other pooled instruments. Links can be established here with the ongoing work on risk at the OECD-DAC INCAF. c Research and explore innovative ways of providing budget aid, with specific focus on non-traditional areas of engagement and the sub-national level, including ways in which civil society and parliament can be more actively engaged in the design, implementation and monitoring of this aid modality. c Carry out analytical work, in particular on: testing hypotheses on the impact of development assis- tance on preventing conflict/fostering resilience; assessing the contribution of budget aid to stabiliza- tion and state-building; and examining the correlation between predictability of funding, aid instru- ments and modes of delivery. Annex 1. Retrospectives on Budget Support: Key Findings The African Development Bank The AfDB finalized a DPL retrospective in 2008, covering budget and balance of payments support operations carried out during 1999-2007 (AfDB 2008). The analysis was updat- ed and extended to cover the period to end-2009 as part of the AFD-12 replenishment process (AfDB 2010). The period covered spans more than four full African Development Fund (ADF) cycles as well as lending to AfDB-eligible countries.16 Overall, the retrospec- tive shows an increase in the proportion of AfDB’s aid delivered through budget support to member countries, notably in the ADF countries, although overall lending volumes de- livered through this instrument fell during 2005-2007 before increasing again at the end of the period. The retrospective presents results with respect to conditionality, sectoral • 26 focus and complementarities, disbursement trends and selectivity against CPIA ratings. The ADF-12 replenishment paper updates the analysis and gives a specific focus on fragile situations. Conditionality. The reviews point to a reduction from an average of ten to seven condi- tions per BSO in ADF countries. This streamlining of conditionality is good practice fol- lowing international experience and has been made possible inter alia by the emphasis on common Performance Assessment Frameworks (PAFs). When drawn from a country’s PRS/NDS, or the equivalent national planning framework, streamlined conditionality reinforces ownership while increasing medium-term predictability (by reducing uncer- tainty in the interpretation of conditions) and having a positive effect on disbursement decisions. Sectoral focus and complementarities. Increased selectivity concentrates PBOs’ focus, frames policy dialogue and channels results. The Bank’s 2008 Governance Strategic Di- rections and Action Plan targets economic governance and public financial management objectives and, increasingly, aims to develop an enabling environment for the private sector. Reflecting this approach, 74 percent of PBO disbursement conditions are related to economic governance and public financial management. Of these conditions, 73 per- cent focus on public financial management and 17 percent on economic management, including the business and investment environment. All AfDB policy-based operations are usually complemented by capacity building projects. This applies to PBOs in fragile states as well either under their normal performance based allocation of ADF or from the additional resources provided under the Fragile States Facility. Disbursement trends. A key advantage of BSOs should be the predictability of fund- ing. Lack of predictability in fragile states in particular can lead to delays in key develop- ment activities and, in fragile situations, macro-economic instability, non-payment of essential expenditures (e.g. salaries) with concomitant risks of a return to civil strife. The retrospective review found that while alignment with national budget cycles has im- proved over time and disbursement delays – as measured between scheduled and actual final disbursements – have been reduced from an average of 3 months to 2 months for ADF countries, significant amounts of budget support funds have not been disbursed on time. Disbursement delays have affected mostly two-tranche operations. The review argues that a multi-year single tranche system would provide the necessary flexibility to respond to changing country performance, including varying the amount, recalibrating the program or accounting for delays in country implementation. Risk management. Under ADF-11, the Fund has approved performance-based opera- tions (PBOs) in both fragile and non-fragile countries. In the former, PBOs can help re- build core systems and institutions for economic and financial management, mitigate persistent vulnerabilities that threaten recovery, and support countries that demon- strate a commitment to reform and the potential for progress. The ADF recognizes that conducting PBOs in situations of fragility can occasion great benefits but also entail con- siderable risk. Risk management, particularly for fiduciary risks, is therefore critical to achieving objectives. For that reason, PBOs in fragile states are preceded or accompanied by complementary ADF-financed capacity-building investment projects and technical as- • 27 sistance targeting areas of fiduciary risk. In addition to the use of up-front fiduciary risk assessments common to all PBOs, PBOs under the FSF are subject to additional criteria, including the use of specific fiduciary safeguards (audits, support for country public fi- nancial management systems, and multiple tranches). When warranted, import support has also been used in fragile states, as it is less susceptible than budget support to fidu- ciary risk. Key recommendations. The latest retrospective analysis (AfDB 2010) highlights a series of priorities for consolidating efficiency and effectiveness and promoting innovation in the use of PBOs going forward. These are captured in current work including an ongoing independent evaluation, preparation of new guidance and a new consolidated PBO policy in 2011. First, a new Fiduciary Risk Management Framework is being finalized to im- prove the transparency and comprehensiveness of risk identification, appraisal, monitor- ing and mitigation. The Bank is working to harmonize the assessment and management of fiduciary risks including safeguards with other donors in fragile states. Second, the Bank will strengthen field office capacity to engage in policy dialogue so as to strengthen Bank leadership in the chairing of budget support groups in Africa. Third, the Bank is working to improve resource allocation and operational planning through country and joint donor dialogue to ensure Fund and other donor resources for complementary tech- nical assistance and institutional support projects. Fourth, working with the internation- al development and evaluation community as well as with African governments, the Bank is working to address the common challenge of better monitoring and reporting PBO results and impacts beyond initial governance effects basing shared analysis on national reporting systems. And finally, the Bank will develop clear parameters and guidelines for using PBOs as a crisis-response instrument. The European Commission The EC has conducted a retrospective on the provision of budget support to fragile situ- ations, based on evidence collected on BSOs conducted during the 9th EDF (EC 2008). The review was carried out for the preparation of the 10th EDF and its findings have also contributed to the elaboration of the Methodological Annex to the Guidelines of General Budget Support, discussed above. The review focuses on five topics: decision and timing of engagement; interpretation of eligibility criteria; underlying causes of fragility; dis- bursement trends; and coordination with donor organizations. Decision and timing of engagement. Decisions to engage have been essentially political and on a stand-alone basis. In most cases, budget support was not planned initially but became a reality in several countries because of the opening of windows of opportunities (e.g. CAR, Haiti, and Guinea Bissau). Although this finding demonstrates that the EDF is a flexible tool, the review recommends that future EDFs should account for the possibil- ity that some fragile situations may become eligible for budget support, and identify ex ante resources that can be used to this end. According to the retrospective, transitions are characterized by three phases: humanitarian and military emergency, socio-economic recovery, and macro-economic stabilization. The review maintains that the bulk of bud- get support has been delivered to support macro-stabilization and contributed to exter- nal arrears clearance operations (EIB/EDF, AfDB on a case by case basis). It argues that, • 28 notwithstanding the risks, this tool can play a critical role earlier on, to support socio- economic recovery. Interpretation of eligibility criteria. The general eligibility criteria for budget support pertain essentially to the macro-economic environment, the state of PFM and the quality of the national strategy. The EC review points out that these criteria have been interpret- ed in a flexible way and with an eye on the direction of the transition (‘good willingness for change’) and potential risks and benefits. The relaxation of certain criteria from the ‘standard’ case is also noted and was discussed above in Section 2. However, it is argued that in many cases the program documents (PDs, conventions de financement) of the BSOs under scrutiny fall short of providing appropriate reference to the articles that gov- ern the provision of budget support under the Cotonou Agreement.17 Underlying causes of fragility. The review argues that the rationale for intervening with budget support in fragile situations, as well as the description of the specific objectives of such interventions, is relatively weak in the BSOs under review. In particular, the links between the deployment of budget support and the process of transitioning out of the crisis into resilience, and the collective impact that donors can have in achieving such an outcome, is not discussed at length in the PDs. In terms of modalities of engagement, the review identifies a gradual move from a single-tranche approach, to two-, fixed tranch- es per year, to the inclusion of variable tranches within a multi-year programmatic ap- proach. Such sequencing takes place as the perceived risks decline. Disbursement trends. Delays continue to undermine the effectiveness of budget sup- port. These are due to difficulties linked to fulfilling the eligibility criteria as well as the weak capacity of recipient countries. Concerning the eligibility criteria, delays are often due to weaknesses in the macro-economic framework that sometimes cause IMF-sup- ported programs to go off-track. In such cases, disbursements may re-start only once corrective measures have been implemented that are acceptable to the IMF. Second, the EC may delay disbursements if complementary information requests pertaining to PFM are not satisfied and/or specific action are not implemented (e.g. these may be specific conditions for the release of variable tranches). Weak capacity may contribute to delays in the transmission of information and the required documentation for legal examination before disbursement. Coordination with donors. The review stresses the importance of donor coordination in fragile situation and highlights the links with the IMF and the World Bank in the provi- sion of budget support. It maintains that the donor conferences that usually take place before re-engaging in a post-crisis situation should pay more attention to the ‘big pic- ture’, notably to the link between project and budget aid and the macro-economic issues that are likely to be faced during transition in the medium-term, such as the reduction of debt. In addition, the review points out that donor coordination in the provision of bud- get support is essential, and emphasizes the useful role that MOUs and PAFs can play in the elaboration and assessment of progress achieved. Key findings. The review presents four findings that in the future should inform the provision of budget support by the EC. First, the PDs should make more explicit refer- ence to the legal basis for engagement, and provide a more robust treatment of the ratio- • 29 nale for the use of budget support. Second, coordination among development partners on macroeconomic issues and PFM is indispensable and in need of further strengthen- ing. Third, emphasis is placed on ensuring the availability of human resources in both donor organizations and recipient countries, in view of the risks associated with fragile situations. Finally, a typology should be developed of different approaches to budget support (e.g. single-tranche, double fixed tranche, and so on) together with an indication of their applicability according to the specific country circumstances (EC 2008). In the follow up, a new tool to assess risk has been developed: the “grille de risques� as part of the newly agreed Methodology is used to decided engaging into budget support, design and review program. The World Bank In 2009, the World Bank finalized a development policy lending (DPL) Retrospective, which reviews 166 development policy operations (DPOs) approved between FY06 and FY09.18 As part of this exercise, a background paper focused specifically on fragile situ- ations, covering 12 IDA DPOs in nine countries during the same period (see Table 2). All operations have been disbursed in the form of grants, given the high probability of debt distress in these countries. Four of the 12 grants have been used for arrears clear- ance. Most operations tend to be stand-alone, single-tranche operations. The following discussion pays particular attention to program scope, underlying causes of fragility, risk assessment, predictability and exit, participation and partnership, and re-engagement. Program scope. Except in Côte d’Ivoire, the DPOs and the Country Assistance Strategies (CASs) or Interim Strategy Notes (ISNs) upon which they were based focused on weak PFM systems and public sector governance and transparency, with technical assistance provided by a parallel project.19 Fiscal stabilization also features prominently in the BSOs. With respect to public sector governance and transparency, PFM systems are the central focus and most conditions/prior actions involve the publication and/or submission of a variety of documents, including quarterly budget execution reports (CAR), ministerial financial accounts (Afghanistan), the budget (Liberia, Togo) the budget communiqué of the Council of Ministers (Cote D’Ivoire), the transfer formulas used to finance decentral- ized government (Sierra Leone) and procurement laws (Liberia, CAR and Togo). Some DPOs chose, as secondary areas of focus, government delivery of social services, particularly education and health (less often, infrastructure), sometimes linked to World Bank investment projects in those areas. DPOs in both Sierra Leone and Burundi have supported government efforts to realign the budget with the PRS and have monitored budget expenditure shares for health and education (Sierra Leone) or pro-poor expendi- tures (Burundi) during implementation. In Lao PDR, the DPO supported the regulariza- tion of timely payment of staff in these sectors to ensure continued provision of service. Most DPOs also included a private sector component, focusing either on regulatory re- form or divestiture of state-owned enterprises. • 30 Table 2. Selected Development Policy Operations in Fragile Situations Fiscal Mgmt, PFM, Fiduciary Stds, Audit Civil Service / Public Administration State Owned Enterprise Reform Transparency & Governance Private Sector Regulation Education and/or Health Financial Sector Reform Project Financing By Fiscal Stabilization Business Climate Number of DPOs Decentralization Client Country Active FY Tranche Afghanistan FY07 1 Single x x x x x x IDA, ADB PFM grants Burundi FY07 1 Multiple x x x x x IDA PFM grant Central African Rep FY07, FY08 2 Single x x x x LICUS on PFM, IDA on Social Services Cote D’Ivore FY08 1 Single x x x x x x Plan for LICUS, IDA Haiti FY07 1 Multiple x x x LICUS Lao PDR FY06, FY07, 3 x x x x x x x IDA capacity building FY08 Single Liberia FY08 1 Single x x x IBRD TF, LICUS, IDA grant on PFM Sierra Leone FY07 1 Single x x x x x IDA Grant Togo FY08 1 Single x x x x x LICUS on SOE governance Source: World Bank 2009a, Nelson 2009. Underlying causes of fragility. All PDs present a description of the country situation, including its position along the spectrum of fragility, the political and security situation, and the risks of instability. The DPOs invariably address PFM and transparency of state budget processes, and the policy measures needed in these areas. By the same token, those related to the social sectors are mainly directed toward compliance with the condi- tions of the HIPC process. However, the link between the provision of budget aid and the bigger picture of strengthening the transition toward resilience (with specific policies geared at doing so) is seldom made in the PDs. Moreover, except in the case of Sierra Le- one, which supports a government fiscal decentralization process, DPOs do not support activities that reach out to rural communities. Risk assessment. The main risks that are commonly mentioned include the reigniting of domestic conflict/unrest, political risk (lack of political capacity/will to undertake re- forms), capacity constraints, fiduciary risk (notably corruption), macroeconomic or fis- cal instability (the inability to raise domestic revenue), and, in some cases, the threat of regional conflict spilling over the recipients’ borders. Fiduciary issues are often singled out as the most prominent, which explains the heavy focus of supporting operations on PFM, good governance and transparency. Capacity constraints are usually addressed by complementary capacity building programs. Yet, PDs are unduly optimistic regarding the ability to work around weak country capacity: follow-up DPOs tend to recognize that ca- pacity development has been less than anticipated in earlier rounds. Predictability. All DPOs are single-tranche while in two cases – Haiti and Burundi – they featured a floating second tranche. Some DPOs are part of a planned annual series under the CAS/ISN, such as in the case of Lao PDR. In practice, follow-up operations have been carried out in most countries listed in Table 2. Similarly to the point raised by the EC • 31 retrospective, a programmatic series of BSOs in fragile situations do not seem to be sys- tematically planned or discussed in the CAS/ISNs, even though a series of stand-alone, single-tranche operations may subsequently materialize. Hence there seems to be little certainty regarding predictability of support, or of ensuring that the overarching stra- tegic development framework is factored in BSOs. While this approach may be under- standable as the World Bank seeks to establish or continue re-engagement in a high-risk environment, it nevertheless may result in a lost opportunity for tying budget aid into broader and more coherent program of assistance. Participation and partnership. According to the Bank’s operational policy and corpo- rate guidelines on development policy lending (OP 8.60), as part of the country dialogue, the Bank advises borrowing countries to consult with key stakeholders and engage their participation in the process of formulating the country’s development strategies. For a DPO, the country draws on this process of strategy formulation to determine the form and extent of consultation and participation in preparing, implementing and evaluating the operation. Partners are limited in these operations, usually restricted to the Bank, the Government, the IMF, frequently a regional development bank (notably the AfDB) and sometimes the EC. Re-engagement. Macroeconomic management, fiscal transparency, and institution building were the objectives of DPL operations to four countries that had stopped receiv- ing budget support from the donor community (CAR, Côte d’Ivoire, Liberia, and Togo). All of these countries sought support mainly to clear arrears to the World Bank and other multilateral and bilateral creditors, so that donors could reengage. These operations were successful in this respect but had very limited policy content. For example, the FY07 operation in the CAR limited its requirements in the area of transparency to a prior ac- tion permitting donor-funded technical advisors to attend Treasury Committee meet- ings. This very limited and light policy content illustrates a tension inherent in the use of the DPL instrument for clearing arrears from those with broader stability/development objectives (Nelson 2009, World Bank 2009a). • 32 References Accra Agenda for Action, 2008, Accra Agenda for Action, Accra: High-Level Forum. African Development Bank, 2004, Guidelines on Development Budget Support Lending (DBSL), Tunis: the African Development Bank. African Development Bank, 2008, Operational Guidelines for the Fragile States Facility, Tunis: the African Development Bank. African Development Bank, 2008a, Policy Based Lending in Retrospective, Governance and Economic and Financial Management, Tunis: the African Development Bank. African Development Bank, 2010, Review of Policy Based Operations Under ADF-11, Discussion Paper prepared for the ADF-12 Replenishment: February 2010, Cape Town. African Development Bank, 2010a, Fiduciary Risk Management Framework for PBOs, forthcoming. Boyce, J. and M. O’Donnell (eds.), 2009, Peace and the Public Purse: Economic Policy for Post-war State-building, Centre for International Cooperation Studies in Multilateralism. Collier, P., and A. Hoeffler, 2002, ‘Aid, Policy and Growth in Post-conflict Societies’, World Bank Policy Research Working Paper No. 2902, Washington DC: The World Bank. European Commission, 2007, Communication from the Commission to the Council,the European Parliament, the European Economic and Social Committee and theCommittee of the Regions: Towards an EU response to situations of fragility, Engaging in difficult environments for sustainable development, stability and peace,Brussels:TheEC. European Commission, 2008, Rapport de synthèse de l’étude sur les appuis budgétaires de la Commission européenne dans les pays ACP relevant de situations de fragilité, Bruxelles: DG Développent European Commission, 2009, Aide budgétaire dans les situations de fragilité, Annexe méthodologique aux Lignes directrices pour la programmation et la mise en œuvre de l’aide budgétaire de la CE, Bruxelles: DG Développent. (Budget Support in situations of fragility, Methodological annex to the Guidelines on programming, design and man agement of budget support.) Department for International Development, 2009, Building the State and Securing the Peace, Emerging Policy Paper, London: DFID. Francois and Sud, 2006, Promoting Stability and Development in Fragile and Failed States, Development Policy Review, Vol. 24, No. 2, pp. 141-160. Horton, Brendan, 2009, Harmonization of Budget Support by IDA, the African Development Bank and the European Commission in Situations of Fragility, Vol. I and II, Washington DC. Koeberle, S., Stravreski, Z., and J. Walliser (eds.), 2006, Budget Support as More Effective • 33 Aid, Recent Experiences and Emerging Lessons, Washington DC: The World Bank. Koeberle, S., and Z. Stavreski, 2006, ‘Budget Support: Concept and Issues’ in Koeberle, Stavreski and Walliser (eds.), Budget Support as More Effective Aid? Recent Experiences and Emerging Lessons, Washington DC: The World Bank. International Monetary Fund, 2005, Rebuilding Fiscal Institutions in Post-conflict Countries, Washington DC: The International Monetary Fund. International Monetary Fund, 2008, The Fund’s Engagement in Fragile States and Post- Conflict Countries – A Review of Experience – Issues and Options, Washington DC: The International Monetary Fund. International Monetary Fund, 2009, A New Architecture of Facilities for Low-income Countries, Washington DC: The International Monetary Fund. Multi-lateral Development Banks (MDBs), 2007, Toward a more Harmonized Approach to MDBs Engagement in Fragile States, Background Paper for the Meeting of the Heads of the MDBs in Washington DC (October 2007). Nelson, Eric, 2009, Development Policy Operations in Fragile and Conflict-affected States, Background Paper for the World Bank DPL Retrospective 2009, mimeo, Washington DC. OECD-DAC, 2006, Joint Evaluation of General Budget Support Synthesis Report (1994- 2004), Published by DFID, Glasgow, UK OECD/DAC, 2007, Principles for Good International Engagement in Fragile States and Situations, Paris: OECD. OECD/DAC, 2008, Concepts and Dilemmas of State Building in Fragile Situations: From Fragility to Resilience, Paris: OECD. OECD-DAC, 2009, INCAF Guidance Paper on Peace-building and State-building, Paris: OECD. OECD-DAC, 2009a, The Legitimacy of the State in Fragile Situations: A Discussion Paper, Paris: OECD-DAC. OECD-DAC, 2010, Understanding Risk in Fragile and Conflict-affected Settings, Concept Note prepared for the INCAF Financing and Aid Architecture Task Team, Paris: OECD. Paris Declaration, 2005, Paris Declaration on Aid Effectiveness: Ownership, Harmonization, Alignment, Result and Mutual Accountability, Paris: High-Level Forum. World Bank, 2005, Good Practice Note for Development Policy Lending: Development Policy Operations and Program Conditionality in Fragile States, Operational Policies and Country Services, Washington DC: The World Bank. World Bank, 2007, Operational Approaches and Financing in Fragile States, Operational Policies and Country Services and Resource Mobilization Department, Washington DC: The World Bank. • 34 World Bank, 2007a, Review of Post-Crisis, Multi-Donor Trust Funds, in cooperation with Canadian International Development Agency (CIDA), Netherlands Ministry of Foreign Affairs and UK Department for International Development. World Bank, 2009, IDA Mid-Term Review: Aid Effectiveness, Operational Policies and Country Services, Washington DC: The World Bank. World Bank, 2009a, Development Policy Lending Retrospective: Flexibility, Customization and Results, Washington DC: The World Bank. World Bank, 2009b, IDA’s Support to Fragile and Conflict-affected Countries: Progress Report 2007-2009, Operational Policies and Country Services, Washington DC: The World Bank. World Bank, 2009c, Post Conflict Performance Indicators: Recommendations of External Panel Review and Next Steps, Washington DC: The World Bank. World Bank, 2010, Review of Instruments: Informal Discussion with the Executive Board, OPCS Slides, Washington DC: The World Bank. World Bank, 2010a, Global Monitoring Report, Washington DC: The World Bank. World Bank. Forthcoming, World Development Report 2011: Conflict, Security and De- velopment. Washington, DC:World Bank • 35 Endnotes 1 There may be other instruments and sources of financing – other than MDTFs – that could be used to finance recurrent costs in fragile situations that are unable to raise sufficient revenues to cover key expenditures, or where the policy dialogue follows a stop-go pattern. This paper does not single out MDTFs per se as an alternative or necessary complement to policy- based lending and grants. It simply highlights their potential role in addressing the recurrent cost challenge in situations of fragility. 2 These situations are often characterized by the presence of a transitional body under UN mandate, a government of national unity which cannot legitimately make long-term commit- ments, or a weak government with very limited revenue mobilization capabilities; large un-met needs regarding both immediate provision of basic services, and more medium-term reconstruc- tion of social and economic infrastructure; and extremely low capacity and weak economic gover- nance and administrative systems. 3 The IMF provides financial assistance for balance of payments purposes. However, a member’s use of Fund resources for budget support is not inconsistent with its legal framework. From an economic perspective, balance of payments needs may coincide with fiscal gaps. In such cases, the use of Fund resources may result in some degree of budget support, either directly (via the Treasury) or indirectly (via the Central Bank extending credit to the government). 4 Overall, all 3 agencies, do not require minimum threshold levels of PFM in order to pro- vide budget support. Rather, the trajectory of change on PFM, potential for results from budget support and government commitment to reform are considered more important indicators in deciding eligibility for budget support. 5 Disbursement can be released upon the implementation of measures that reinforce those ‘ultra-basic’ elements of the PFM system such as the publication of a decree that sets up a treasury committee or the transmission of budget execution reports. 6 Other risks identified include: uncertainty regarding the ‘whole of government’ com- mitment and capacity to implement reforms; scant information on economic, social and political conditions; and concerns on aid dependency. 7 The main objective of PAFs is usually to provide a concise and monitorable set of indica- tors, drawn from government’s own existing commitments under the NDS/PRS and other related agreements with cooperating partners. These are typically selected and revised on an annual ba- sis. The PAF matrix is also typically used by multi-donor budget support frameworks as the main focus for the joint assessment of progress in the implementation of the Government’s reform program. 8 The available evidence suggests that in post-conflict environments aid tends to taper off after initial spikes, while in chronically fragile environments it often follows a stop-and-go pat- tern (Collier and Hoeffler 2002). 9 This poses challenges. The first is an analytical one. Risk management depends on the ability to foresee the likely future given the available knowledge of the past and present. The second is a practical one. Even if the future can be foreseen, it may not be possible to influence the causal factors involved to reduce the probability of adverse outcomes. The third concerns influencing be- • 36 havior. Even if some events can be predicted, other people need to be persuaded to do something about it (e.g. climate change). 10 Examples of safeguards for budget risks: publication of financial and management in- formation; public expenditure tracking studies; community level disclosure of budgets. Examples of budget control measures: safeguards for audit risks: ex-ante audits of particular budget lines; audit performed by external auditors/agencies appointed by the donor. Examples of safeguards for procurement risks: use of external procurement specialist; publication of public procurement information (upcoming bidding). Examples of safeguards for corruption risks: CSOs involvement in tracking the use of public funds; CSOs monitoring of payment of wages and salaries; NGOs undertaking social audits; Introduction of ‘negative list’ (AfDB 2010a). 11 MDTFs are not a panacea and can also be affected by coordination problems, notably when bilateral partners wish to introduce political conditions or triggers that cannot be taken on board by multi-lateral institutions. See also footnote 1 on the CAP’s approach to MDTFs. 12 Annex 1 reviews in detail the retrospective conducted by the three institutions on BSOs, including those implemented in fragile states (AfDB 2008, 2010, EC 2008, World Bank 2009). 13 At the World Bank delays in disbursements of BSOs are often due to delays in Board pre- sentation because of non-compliance with prior actions. 14 The World Bank is elaborating a set of peace- and state-building indicators, in collabora- tion with key international partners, such as the OECD-DAC. 15 It should be noted that the articles of agreements of the institutions may not allow their intervention in security-related sectors. 16 Funds from the ADF finance the bulk of Africa’s low-income countries, while ADB-eli- gible countries include only thirteen nations, which are middle-income countries. Many of the ADF-eligible countries whose operations are examined in the retrospective are classified as fragile situations. Hence, this experience is very relevant to the topics discussed in this paper. 17 The Cotonou Agreement, signed on 23 June 2000, aims to promote the economic, social and cultural development of the African, Caribbean and Pacific states, peace and security, and a stable and democratic political environment. 18 The report presents three key messages. On flexibility, DPOs have supported country- owned reforms aimed at achieving specific development results in countries with different needs, in the form of grants, credits, or loans or emergency financing. On customization, DPOs have provided financing and policy advice in areas where authorities required expertise and technical knowledge. On results, DPOs have been associated with positive outcomes in the delivery of so- cial services in health, education and PFM (World Bank 2009a). 19 Some of these operations were accompanied by technical assistance/capacity-building projects funded by grants—in the form of economic management support projects—from IDA (Afghanistan, Côte d’Ivoire, Burundi, Haiti, Laos, Liberia, and Sierra Leone), IBRD surplus (Af- ghanistan, Haiti, and Liberia), LICUS (CAR, Haiti, Liberia, Togo, and planned for Côte d’Ivoire), and PFM projects financed by MDTFs (Afghanistan, Côte d’Ivoire, Togo) (Nelson 2009). • 37