CPR Working Papers Social Development Department Environmentally and Socially Sustainable Development Network Paper No. 6 June 2003 Financing and Aid Management Arrangements In Post-Conflict Situations ____________________________________ Salvatore Schiavo-Campo 2 Acronyms and Abbreviations AACA Afghan Aid Coordination Authority ADB Asian Development Bank AfDB African Development Bank AMA Aid Management Agency ARTF Afghanistan Reconstruction Trust Fund BiH Bosnia and Herzegovina BoP Balance of Payments CG Consultative Group CPR Conflict Prevention and Reconstruction Unit (World Bank) DGF Development Grant Facility (World Bank) EC European Commission ETFA Emergency Trust Fund for Afghanistan (UNDP) EU European Union FRY Federal Republic of Yugoslavia IDF Israeli Defense Force IMF International Monetary Fund MDTF Multi-Donor Trust Fund GDP Gross Domestic Product IDA International Development Association OECD Organization for Economic Cooperation and Development OED Operations Evaluation Department (World Bank) OHR Office of the High Representative (Bosnia and Herzegovina) PA Palestinian Authority PIC Peace Implementation Council (Bosnia and Herzegovina) PIU Project Implementation Unit PLO Palestinian Liberation Organization PCF Post Conflict Fund (World Bank ­ financed from DGF) PECDAR Palestinian Economic Council for Development and Reconstruction PISG Provisional Institutions of Self-Government (Kosovo) RS Republika Srpska (entity within BiH) SFOR Stabilization Force (Bosnia and Herzegovina) TATF Technical Assistance Trust Fund (West Bank and Gaza) TFBH Trust Fund for Bosnia and Herzegovina TFET Trust Fund for East Timor TFGWB Trust Fund for Gaza and West Bank TFK Trust Fund for Kosovo UNDP United Nations Development Programme UNMIK United Nations Mission in Kosovo UNTAET United Nations Transitional Administration for East Timor 3 Table of Contents I. INTRODUCTION...........................................................................................................................1 II. CASES..........................................................................................................................................3 WEST BANK AND GAZA.........................................................................................................3 The Authorizing Environment......................................................................................................3 The Context...........................................................................................................................3 The Mandate for International Aid and Key Milestones ............................................................4 Donor Coordination, Reconstruction Funding and Aid Management...............................................5 Donor Coordination................................................................................................................5 The Technical Assistance Trust Fund (TAFT)..........................................................................5 The Trust Fund for Gaza and West Bank (TFGWB).................................................................6 The Holst Fund ......................................................................................................................6 Aid Management and Implementation Arrangements--PECDAR..............................................8 Accomplishment and Problems ....................................................................................................9 Accomplishments and Success Factors....................................................................................9 Problems and Issues............................................................................................................. 10 West Bank and Gaza: A Chronology.......................................................................................... 11 BOSNIA.................................................................................................................................. 12 The Authorizing Environment.................................................................................................... 12 The Context......................................................................................................................... 12 The Mandate for International Aid and Key Milestones .......................................................... 14 Donor Coordination, Reconstruction Funding and Aid Management............................................. 14 Donor Coordination.............................................................................................................. 14 The Trust Fund for Bosnia and Herzegovina (TFBH).............................................................. 15 Aid Management and Implementation Arrangements.............................................................. 16 Accomplishments and Problems ............................................................................................ 17 Bosnia and Herzegovina: A Chronology ..................................................................................... 17 EAST TIMOR.......................................................................................................................... 18 The Authorizing Environment.................................................................................................... 18 The Context......................................................................................................................... 18 The Mandate for International Aid and Key Milestones .......................................................... 19 Donor Coordination, Reconstruction Funding and Aid Management............................................. 19 The Trust Fund for East Timor (TFET).................................................................................. 20 Aid Management and Implementation Arrangements.............................................................. 21 Accomplishments and Problems................................................................................................. 22 Accomplishments and Success Factors.................................................................................. 22 Problems and Issues............................................................................................................. 23 East Timor: A Chronology......................................................................................................... 24 AFGHANISTAN...................................................................................................................... 25 The Authorizing Environment.................................................................................................... 25 The Context......................................................................................................................... 25 The Mandate for International Aid and Key Milestones .......................................................... 26 Donor Coordination, Reconstruction Funding and Aid Management............................................. 26 Donor Coordination.............................................................................................................. 26 The Afghanistan Recovery Trust Fund (ARTF)...................................................................... 27 Aid Management and Implementation Arrangements.............................................................. 29 4 Accomplishments and Problems................................................................................................. 29 Accomplishments and Success Factors.................................................................................. 29 Problems and Issues............................................................................................................. 30 Afghanistan: A Chronology...................................................................................................... 32 III. ESTABLISHING A MULTI-DONOR TRUST FUND AND AID MANAGEMENT AGENCY IN A POST CONFLICT SITUATION ........................................................................................ 33 Structural and Design Issues...................................................................................................... 33 Organizational and Procedural Issues......................................................................................... 35 Internal Bank Processes............................................................................................................. 37 The Aid Management Agency.................................................................................................... 38 A Trilateral Division of Responsibilities..................................................................................... 39 Stages Toward Establishing an MDTF in Post-Conflict Situations ................................................ 41 IV. GENERAL ISSUES AND SELECTED RECOMMENDATIONS................................................ 43 The General Context ................................................................................................................. 43 Insist on the Script, Not the Cast: Relating to Donors.................................................................. 44 Relating to the UN.................................................................................................................... 44 The Importance and Limits of Ownership: Relating to the Client................................................. 45 Go Soft, Go Local: Relating to NGOs and Local Communities.................................................... 46 Selected References......................................................................................................................... 48 Annex: Partial List of Persons Contacted.......................................................................................... 50 Annexes (available on request from the CPR Unit): 1. Grant Agreement for the Technical Assistance Trust Fund (TATF) for West Bank and Gaza 2. Grant Agreement for the Afghanistan Reconstruction Trust Fund (ARTF) 3. Sample Donor Agreement for the ARTF 4. Terms of Reference for the Monitoring Agent for the ARTF 5. Organizational Structure of an Aid Management Agency 5 Preface and Acknowledgements This paper was produced at the request of, and funded by, the Conflict Prevention and Reconstruction Unit and the East Asia Region of the World Bank, based on terms of reference developed by Ian Bannon and Mary Judd. It is intended primarily as operational reference for Bank staff and other international and national public officials working on financing and aid management arrangements in current and future post-conflict situations. However, because of the inherently broad context of such situations, the conclusions and recommendations are not limited to operational and technical matters but extend to strategic and policy issues. The research was grounded on a variety of published sources and websites; Bank reports (particularly from OED and the CPR Unit) and internal documents; discussions with key World Bank staff currently or previously involved in financing post-conflict reconstruction (and a handful of persons at the IMF, UN and ADB); and my direct personal experience in Palestine, the Balkans, and Afghanistan. I am grateful to all the respondents for their time, help and insight, as well as to the helpful staff of the Legal Files Library. However, a special acknowledgement is due of the valuable inputs and extensive comments provided by Hadi Abushakra, Rema Balasundaram, Anders Bjorgung, Sean Bradley, Sarah Cliffe, R. Kyle Peters, and Christine Wallich. Because this paper lays no claim to originality other than in its identification and synthesis of key issues, detailed footnotes and references are kept to a minimum. Also, several of the issues are sensitive; many of the personal views were provided on a not-for-attribution basis; and the non-Bank respondents do not wish to be identified. The customary caution is thus especially relevant: while the assessments, conclusions and recommendations in this paper are grounded largely on the sources consulted and information received, they are my own--and so is the responsibility for any errors. S. Schiavo-Campo Executive Summary The general setting Severe conflict, especially its most virulent ethnic form, destroys much more than buildings and power plants. It short-circuits the rules that keep human interaction constructive and predictable, targets primarily the organizations and individuals who administer those rules, and wipes out most positive forms of social capital. Post-conflict reconstruction is first and foremost an institutional challenge. Hence, the first lesson of experience for aid in post-conflict situations is the imperative of assuring robust linkages between the aid and the rebuilding of local institutions, and the core challenge is the balancing of immediate reconstruction priorities with long-term institutional development. This paper is concerned with the specific issues of financing modalities and aid management arrangements in post-conflict situations, and advances a number of recommendations on the basis of a review of several recent cases, among which four are assessed in detail: West Bank and Gaza, Bosnia and Herzegovina, East Timor (Timor-Leste) and Afghanistan. While generally applicable recommendations do emerge from the review, the most important recommendation is to tailor the design and sequencing of financing and aid coordination to the circumstances of the specific case. Each case has its own unique core features. Thus, for example: · In West Bank and Gaza, the key features to which the funding and programming of post-conflict reconstruction had to adapt were the constraints posed by Israel, the excessive centralization of decision-making in the Palestinian Authority, and the patchwork nature of the territories under Palestinian control. · In Bosnia and Herzegovina, the extreme interethnic hostility and political and administrative fragmentation conditioned everything, the main challenge of aid arrangements was to recognize the reality of dysfunctional fragmentation without thereby strengthening it, and the strategic principle for future recovery can be formulated as "infant social capital" protection. · In East Timor, external assistance was affected from the start by the virtual absence of any administrative capacity in the territory, the extremely weak human resource base and its poverty. · In Kosovo, some local capacity existed, although mainly in informal networks competing to an extent with the UN and the formal institutions of provisional self-government, but lack of coordination in technical assistance resulted in conflicting advice and `turf' fights among expatriate experts. · In Sierra Leone and the Greater Great Lakes, the core issue was disarmament and reintegration of former rebel combatants--including the difficult problem of child soldiers. · In Afghanistan, the key objective is to reconstitute a functionally limited but effective state, and to reduce economic distance by re-establishing security and restoring transportation and communication links. While oversimplification is to be avoided, an early identification of the core features of each post-conflict situation is mandatory for an effective reconstruction strategy, and for the design of the financing and aid management architecture that is to embody that strategy. Trust funds and aid management organization By all evidence, the World Bank has been generally effective and very timely in assisting post-conflict reconstruction; the trust funds and other financing arrangements have not only been grounded on the experience of earlier cases, but have also been innovative where needed; and much successful effort has gone into facilitating the involvement of the recipient government and local partners. In particular, the quality, efficiency and speed of Bank work in the early post-conflict days have been nothing short of remarkable. The Bank has also endeavored to consult with and elicit the cooperation of other major donor agencies, although with mixed results--mainly because of higher-level political disputes. ii However, the Bank (and the donor community as a whole) has occasionally been weak and/or insufficiently clear on the need for the recipient government and other donors to abide by certain basic policy standards. Also, the dynamics of the development business and the short-term urgencies typical of post-conflict situations have often been allowed to prevail over long-term institutional development objectives. In the future, the Bank might consider being more firm on the essentials and more flexible on everything else. Overall, there is a consensus that only the World Bank can provide the experience, drive, and diversity of knowledge, and credibility required for effective leadership of international assistance to post-conflict civilian reconstruction. With that consensus comes a responsibility for the Bank to listen, consult, and facilitate. The main lessons of experience on multi-donor trust funds (MDTFs) and on the recipient government's aid management in post-conflict situations are summarized below. Among structural and design issues, the main conclusions are: · An MDTF must fulfill both a fiduciary and an executive function. The legal, accounting, disbursement and reporting provisions required for the fiduciary function are now well defined. The effective exercise of the executive function of an MDTF requires, in addition, an agreed program of reconstruction activities; ownership by government and major potential donors; a workable interface with the local aid management agency; fullest possible transparency; and openings for citizens' voice. · A realistic, comprehensive and public government budget, consistent with the reconstruction program, is essential. · There must be incentives for individual donors to join an MDTF, including an MDTF design that gives them comfort that their aid goes for priority purposes while precluding earmarking of the aid. While all contributions are commingled into a common pool, donor preferences can be explicitly acknowledged, and expenditures in the broad categories should be regularly reported. · MDTF governance arrangements must provide for systematic consultation with and reporting to the contributing donors. In this regard, clarity of arrangements and cohesion among the main donors are more important than organizational simplicity per se. · The coverage of an umbrella MDTF can potentially encompass all civilian expenditure except very large flagship investment projects, strictly humanitarian aid, or security-related programs such as de- mining--separate dedicated trust funds can be useful mechanisms for those activities. The main advantage of an umbrella fund is the possibility of a closer linkage with the recipient country's budget, and hence more robust dialogue on fiscal and development policy. What is to be avoided is fragmentation of funding vehicles, especially between financing of recurrent costs and investments. · On balance, it is preferable for a Bank-administered MDTF not to finance police or prisons. Among the organizational and procedural issues, the main ones are: · Time is of the essence for aid interventions in post-conflict situations. A pragmatic compromise is needed between the extremes of waiting until all contributions are deposited, and of starting MDTF operations as soon as the first pledges are made. · No compromise can be made, however, with the need to put in place measures to minimize corruption and leakages before the MDTF enters into any commitment. · Non-project technical assistance (TA for institutional development and capacity building) can be financed either as a component of an umbrella MDTF or by a separate trust fund. In either case, the key requirements are an agreed framework of priorities, the closest possible involvement of the local counterparts, and tight monitoring and quality control. Non-project TA activities should be linked with a host-government capacity-building program, and every single TA contract should include a training element. At a minimum, TA should take special care not to aggravate local capacity problems by introducing overly complex systems or methods unsuited to local conditions. · While an agreed reconstruction program and national budget are fundamental, an unallocated financial cushion should be maintained in the MDTF for urgent expenditure needs as they arise. · Recurrent costs are well suited for financing through an MDTF, but close monitoring of the broad expenditure categories (on the basis of clear budgetary understandings) is essential. Also, the iii payments mechanism must function well--even if it must be initially subcontracted. On balance, in post-conflict situations it is advisable to hire a consulting firm to act as agent for the MDTF to verify the eligibility and correctness of withdrawal applications, and carry out spot checks of the validity of transactions. · The Bank should be prepared to halt disbursements in the event of serious and uncorrected deviation from the agreed policies and expenditure composition, and/or substantial corruption. The aid management agency (AMA) of the recipient government is the bridge from donors to government and the main source of initial client ownership. As noted, the fundamental strategic challenge of post- conflict reconstruction is the reconciliation between immediate urgencies and longer-term sustainability. In most cases, the government is extremely weak during the immediate post-conflict period. Thus, the agency must not only interface with donors and regulate aid traffic, as in steady-state situations, but also help formulate the reconstruction program, serve as proto-government, and implement directly a number of activities. There is accordingly a "sunset dilemma" in post-conflict reconstruction: an aid management agency is created because the formal government structures do not yet exist or do not have the capacity to implement reconstruction, and also because donors require transparent and reasonably corruption-free financial management. Over time, as the regular governmental institutions grow, competition emerges between the governmental structures and the parallel aid management agency. Instead of a smooth handover, the paralle l tracks tend to persist, partly because the AMA has built up greater implementation capacity and partly because accountability and financial transparency remain a must for donors. And so, the AMA stays active longer than envisaged, competing with regular government ministries for resources and authority. The key lesson of experience is thus that government and donors should agree from the outset on an exit strategy for the aid management agency; link that strategy to appropriate conditionality vis-à-vis emerging government structures; and at the same time assist in building institutional capacity in the regular organs of government. Proliferation of weak or corrupt government ministries is not a sound alternative to a technocratic monopoly of decision-making by the AMA. Other lessons of experience with the structure and operations of aid management agencies in post-conflict cases are discussed in the text. The reader is also referred to the text for a discussion of the allocation of MDTF responsibilities among the Bank, the aid management agency, and the monitoring agent for the trust fund, as well as a suggested sequence of stages of Bank involvement in reconstruction needs assessment, aid coordination and MDTF establishment--respectively during the in-conflict, pre-post-conflict, and immediate post-conflict periods. General issues and recommendations A general prerequisite for post-conflict reconstruction and effective financing is to establish and maintain public order and security. The best needs assessment, reconstruction program, ownership, staffing, financing arrangements, resource availability, policy framework, and implementation capacity are worth little if program implementation and economic activity are constrained by lack of physical security. Among the other contextual and coordination requirements emerging clearly from experience (in addition to the need for a comprehensive and realistic government budget, noted earlier) the most important are: · The need for a definite political resolution of the conflict and clear mandate for the Bank. · Genuine partnership among the key international institutions--the Bank should insist on the "script", not on the "cast", i.e., actively provide for meaningful roles for other donors and liberally share the credit for accomplishments, but without compromising on the basic criteria for effective assistance. For their part, once a consensus is reached all donors should cooperate with the norms and programs--whether or not they choose to contribute to the MDTF. The frequently used "comparative iv advantage" terminology is especially misleading in a relationship that should be characterized by cooperation and not by trading. · Although the priority in post-conflict reconstruction is on urgent short-term needs, neglect of basic economic policy issues will preclude the achievement of sustainable results. · Aid must encourage government capacity building and facilitate local initiative as much as possible. · A large diaspora can be a human and financial asset, but a balance between "outsider nationals" and "insider nationals" is important for sustainability. External agencies should be mindful of their natural bias in favor of the "outsiders" and try to correct it. · Proactive communication is important to deflate early unrealistic expectations and the ensuing erosion of public support. · A tight timetable of political and economic progress, with appropriate benchmarks, is needed for momentum and focus. · The divergence between the political timetable and the reconstruction timetable is normal, but needs to be recognized in order to be bridged. In cases when the UN exercises civilian authority and serves as de facto government, it is especially necessary for the Bank and the UN to reach early and clear understandings of each other's roles and responsibilities. · The international experience yields no conclusion as to the relative advantage of tackling hard political issues early in order to facilitate the reconstruction efforts, or after a period of consolidation in order to permit some rebuilding of trust. However, it must neither be assumed that hard issues must necessarily be left until later, nor to forget that at some stage they do need to be addressed. · Setting public management standards and sound policies is key for sustainable fiscal and institutional development, but in high-profile cases efforts to do so can be frustrated by external pressures to keep the money flowing. Reasoned firmness by the Bank, through judicious use of the MDTF resources, has a chance to prevent excessive expansion of governmental structures and erosion of standards of transparency, integrity, accountability and the rule of law--but only if there is a general consensus on policy direction and the donor community coalesces around those standards. · Because post-conflict reconstruction is inherently a top down affair, caution must be exercised lest existing NGO activities and local structures are inadvertently suffocated by the reconstruction assistance. Beyond protecting what exists, the Bank should aggressively encourage the recipient government (and the donors) to incorporate into the reconstruction program the contribution of local communities, and of the local and international NGOs that have been laboring in conflict areas for years. The last point leads back to the initial remark. The potential contribution of NGOs and local structures goes much beyond assisting in implementation, or even acting directly as implementing agencies. Some of the most effective components of reconstruction programs in the past have relied on empowerment of local communities and partnerships with NGOs. Moreover, capacity-building at the local level is a necessary condition for the evolution of the rule of law, accountability, and transparency, and the local structures can be important as well to underpin the gradual rebuilding of the social capital destroyed by the conflict. FINANCING AND AIDMANAGEMENT ARRANGEMENTS IN POST-CONFLICT SITUATIONS "Better to let them do it imperfectly than to do it perfectly yourself, for it is their country, their way, and your time is short." T.E. Lawrence I. INTRODUCTION Severe conflict, especially its most virulent ethnic form, destroys much more than buildings and power plants. It short-circuits the rules that keep human interaction constructive and predictable, targets primarily the organizations and individuals who administer those rules, and wipes out most positive forms of social capital. Post-conflict reconstruction is first and foremost an institutional challenge. Failure to meet that challenge dooms the effectiveness of aid. Hence, while this paper is primarily concerned with the modalities of funding and aid management for post-conflict reconstruction, the single most important lesson of experience is the imperative of robust linkages between the aid and the rebuilding of local institutions, and the core challenge is the balancing of immediate reconstruction priorities with long-term institutional development in the recipient entity. The Bank's first loan was a post-conflict loan, to France for infrastructure reconstruction. Since, then, post-conflict reconstruction has been a major focus of Bank work and, by rough count, well over a hundred post-conflict trust funds have been created at the Bank.1 Of these, the large majority has been small and targeted to special programs or projects (most were financed from Bank income without donor contributions). Clearly, not all post-conflict funding and aid management arrangements could be examined in detail through this paper. Selectivity was needed, and this study rests accordingly on the review of the major recent cases of what may be described as "nation-building" post-conflict situations.2 Among these cases, West Bank and Gaza, Bosnia and Herzegovina, East Timor, and Afghanistan are assessed in detail. In each case, the study focuses on the experience with the design, governance, and operations of multi-donor trust funds (MDTF); donor coordination; and arrangements for aid management by the recipient government. The paper is thus limited in two important ways. First, it does not cover the crucial work done in recent years to program and finance disarmament, demobilization and reintegration (DDR)--particularly in Africa.3 This is due to time and space limitations as well as to the different nature of DDR issues from reconstruction programs in a `nation-building' context. Second, it does not address the contribution of project and sectoral post-conflict assistance. Each of these topics deserves in-depth analysis in its own right. Focusing on the high-profile nation-building cases does not negate the efforts of those laboring in post-conflict settings (non-governmental actors included) to address specific micro- or meso-level issues. On the contrary, without those efforts, the macro-level financing and aid management architecture would rest on a foundation of sand. 1It is difficult to identify all such funds from the existing database. With the exception of activities covered by the Post-Conflict Fund (PCF)--itself largely financed from the Development Grant Facility (DGF)--the titles of many trust funds do not indicate that they were established to finance post-conflict reconstruction. 2 Mainly Afghanistan, Bosnia, Cambodia, Timor-Leste, Kosovo, Sri Lanka, and West Bank and Gaza. Other cases also reviewed for this paper include El Salvador, Haiti, Guatemala, Mozambique, and Uganda. 3 Nevertheless, the experience with some DDR trust funds and other post-conflict assistance has also underpinned the research for this paper, and is to some extent reflected in its conclusions. The major cases of MDTFs for DDR are Sierra Leone and the Greater Great Lakes (for Central Africa). The latter is the only contemporary example of a multi-recipient trust fund. (The Marshall Plan was the first.) Because of its fragmented structure, however, for all intents and purposes Bosnia and Herzegovina is also a multi-recipient case. Conversely, while Kosovo remains technically a province of Serbia, it constitutes in effect a separate entity. All other cases of post-conflict assistance are to national governments. 2 Note at the outset that aid effectiveness is conceptually and practically a continuum. Experience shows that while a well-designed funding arrangement greatly facilitates good aid coordination and funding effectiveness, it is neither a sufficient nor a necessary condition. Trust funds financed only from Bank surplus can perform an aid harmonization function similar to that accomplished through an umbrella MDTF, if the financing from other donors is closely coordinated, and is programmed to support an agreed and coherent program of reconstruction reflected in a comprehensive and realistic budget. The necessary synergy between the aid mechanisms and the recipient's own institutional arrangements can also be achieved through separate but convergent funding arrangements. Conversely, a single umbrella MDTF will accomplish little for aid effectiveness if its design is faulty, or its provisions are disregarded by donors, or its sectoral and micro underpinnings are weak, or the security environment unstable. The next chapter examines in detail each of the four major cases--in chronological rather than alphabetical order, as in every instance the first step by the Bank staff team has been to review the experience of earlier cases and tap the memory of those concerned. Each case is structured in similar fashion: the broad political economy context; the mandate for and key milestones of the aid effort; donor coordination, funding mechanisms and aid management arrangements; and major accomplishments and problems. For the readers' convenient reference, each case concludes with a brief chronology of events immediately preceding and following the settlement of the conflict. Chapter III recapitulates the features and modalities of establishment of MDTFs and of the recipient government's aid management agency, concluding with a recommended sequence of steps toward establishing an MDTF. The concluding Chapter IV brings out more general findings and recommendations germane to the issue of funding and aid management arrangements in post-conflict situations--respectively vis-à-vis other donors, the recipient government; NGOs and civil society; and internal Bank processes. A partial list of persons contacted is annexed. The key operational documents are not included in this paper for reasons of length and excessive detail, but are available on request from the Conflict Prevention and Reconstruction Unit of the World Bank. These are: the grant agreement, a sample donor agreement, and the monitoring agent's terms of reference for the most recent MDTF (the Afghanistan Reconstruction Trust Fund); the grant agreement for the TA Trust Fund for West Bank and Gaza; and a "model" organizational chart for an aid management agency. Each section is intended to be self-standing, and some slight repetition between sections is therefore inevitable. The interested readers are advised to not necessarily go through the entire paper, but to take a look at the Executive Summary and then pick and choose from the Table of Contents the case study or sections most germane to their interests. The paper is structured to make this possible. 3 2. CASES WEST BANK AND GAZA4 The Authorizing Environment The Context This is obviously not the place to attempt to recount the tortured history of Israeli-Palestinian relations, nor to evaluate the reasons for the collapse of the Oslo process, or the tragic events of the last two years. It is essential to underline, however, that both the atmosphere and the factual situation were very different during the early years of operations of the Bank-administered MDTF for the West Bank and Gaza (WBG). The political agreements were being implemented step by step, even if imperfectly; mutual distrust was slowly dissipating; Israeli-Palestinian cooperation was expanding and in some areas becoming routine (including the startling spectacle of joint security patrols); and results on the ground were just beginning to undo some of the damage suffered by the WBG economy. All this began to unravel after Rabin's assassination and the coming to power of the new government in Israel. In late 1992, the Israeli Government agreed to start direct discussions with the Palestinians in Madrid--on the basis of a compromise solution whereby the Palestinian delegation would be sanctioned by the PLO but would not officially represent the PLO nor include any of its leaders. However, the Madrid negotiations occupied most of 1993 in formalistic exchanges intended in effect to avoid confronting any real issue. After several fruitless months, a separate informal demarche by Israeli academics and Palestinian personalities led to discussions hosted by the Norwegian Government in Oslo. These discussions, held in complete secrecy, involved progressively higher and higher political levels, and eventually culminated in the Oslo Accords, sanctioned by the famous Arafat-Rabin September handshake on the White House lawn.5 There are major differences between WBG and other post-conflict situations. The physical infrastructure in the Occupied Territories6 was dilapidated but somewhat serviceable. The quality of human resources was high and the structures of civil administration were workable--only the top levels were staffed by Israeli personnel, with most of the work done by Palestinians. The economic situation was depressed but largely as a temporary result of the return of a large number of Palestinian workers after the 1991 Gulf War rather than structural reasons, and prospects looked good. Nor were there internal cleavages--apart from a sharp ideological divide between the PLO and Hamas and other groups, and some friction between "insiders" and "outsiders".7 Ownership was and remained high.8 The key distinguishing features of 4 This section draws heavily from direct experience and from personal communications with other Bank staff listed at the end of the paper. Especially useful sources were the account of the operations of the Holst trust fund (Holst Peace Fund--Supporting Development in the West Bank and Gaza, World Bank, 2002); the relevant sections of The World Bank's Experience with Post-Conflict Reconstruction (OED, Washington, 1998); and the OED report West Bank and Gaza--An Evaluation of Bank Assistance (Report No.23820, March 7, 2002). The OED report, however, contains errors of fact and some of its qualitative conclusions are highly debatable or altogether off the mark. 5In light of all the criticism unfairly heaped on the Oslo process, it is ironic that the recently-appointed Prime Minister Mahmoud Abbas ("Abu Mazen"), was the principal conduit between Arafat and the Palestinian delegation in Oslo--which was led by current legislative assembly chairman Ahmed Qurie (Abu Ala). As of May 2003, the chief executive and the leader of the Palestinian legislature are the two men most closely identified with the Oslo Agreement. 6The Bank's compromise term between the Israeli preference for "Territories" and the Palestinian preference for "Occupied Palestinian Territories". 7The insiders saw themselves as resisters of Israeli occupation and were resentful of the "Tunis" outsiders taking over; the PLO hierarchy, for its part, was not fully trusting of those who had lived under occupation. An accommodation eventually did take place. A similar but more acute problem is affecting Afghan reconstruction today. 4 Palestine to which the funding and programming of post-conflict reconstruction had to adapt were the constraints posed by Israel, the excessive centralization of decision-making in the Palestinian Authority, and the patchwork nature of the territories under Palestinian control. The Mandate for International Aid and Key Milestones The Oslo accords of September 1993 came as a surprise to everyone. The Bank had thus no opportunity to carry out any of the advance planning and consensus-building activities that were to facilitate the funding and implementation of reconstruction programs in many subsequent post-conflict cases. However, the Madrid-organizing group had earlier requested the Bank to carry out a comprehensive assessment of the development problems and prospects in the occupied Palestinian territories. By fortunate coincidence, the Bank's six-volume report had been completed in August 1993, and was published to general praise just two weeks after the conclusion of the Oslo Accords.9 The Bank thus found itself in the unusual position of having not only the best information on the situation on the ground, but also excellent contacts with both Palestinian and Israeli interlocutors and a staff team with strong leadership already in place. Moreover, although politically the Madrid negotiations went nowhere, they did provide the Palestinian side with the opportunity to assemble a diverse and experienced group of individuals. Thus, just a few days after the White House handshake, a large Bank mission was already in Jerusalem, interacting with a large, highly motivated and very competent Palestinian group of interlocutors, and maintaining constructive contacts with the Israeli authorities. The mission conducted a full assessment, jointly with the Palestinian interlocutors, of infrastructure and social needs in the economy of the West Bank and Gaza, and formulated proposals for the establishment of a Palestinian aid management and implementation agency, as well as for international funding mechanisms commensurate to the challenge. The Emergency Assistance Program produced as a result of the October mission was presented at an interim donor meeting in Paris in December 1993, and at a subsequent one in January 1994, which yielded pledges of $2.1 billion for 1994-98. In the meantime, the Bank's Legal Department had produced a pioneering legal opinion opening the door to Bank assistance to WBG through administration of trust funds.10 This opinion, taking its lead partly from the 1991 fund for technical assistance to the Soviet Union,11 provided a lucid argument and set the precedent for Bank administration of multi-donor trust funds in all subsequent post-conflict situations. In essence, the argument was that although Palestine was neither a Bank member nor (at the time) a prospective state, assistance to it would meet the legal requirement of Bank assistance being for the "benefit of Bank members", because several countries members of the Bank would benefit from peace in the Middle East. As a result, the Board approved in October a multi-donor Technical Assistance Trust Fund (TATF) and in November a Bank-financed Trust Fund for Gaza (later expanded to the West Bank, the TFGWB). Also, the Palestinian Authority, successor to the Palestine Liberation Organization, had created at the end of October its own aid management agency--the Palestinian Council for Development and Reconstruction, PECDAR--largely based on recommendations of the Bank mission. The last 8 While donor coordination problems were substantial, the conclusion in the 2002 OED report that Palestinian ownership was weak is flat wrong, and so are many of the assertions on which it is based. 9 Developing the Occupied Territories--An Investment in Peace, (ECA/MENA Technical Department, World Bank, Washington, October 1993.) The coincident timing of publication of the report led some observers to wonder at the amazing speed and efficiency of the Bank, and other observers with a more conspiratorial turn of mind to infer that the Bank had somehow been apprised of the secret Oslo negotiations. 10 See Legal Memorandum of September 30, 1993. An elaboration by Ibrahim Shihata, Hadi Abushakra, and Hans- Jurgen Gruss was published later as "Legal Aspects of the World Bank's Assistance to the West Bank and Gaza Strip" (The Palestine Yearbook of International Law, Vol. II, 1992/94.) 11For $30 million from Bank surplus; see "Proposed Transfer from Surplus to Fund Technical Assistance to the USSR", Doc R91-184 of July 30, 1991. 5 building block for reconstruction assistance was put in place when the Board approved in January 1994 the "start-up costs" trust fund--later renamed Holst Fund to honor the memory of the Norwegian Foreign Minister Johan Jorgen Holst, who had exercised a crucial role in the peace negotiations. Donor Coordination, Reconstruction Funding and Aid Management Donor Coordination The surprise of the sudden Oslo Accords and the breakneck speed of the subsequent events made it impossible for the Bank to enlist other donors in the October mission and to carry out the other normal preliminary contacts with its development partners. Through discussions concurrent with the October mission, however, a 12-donor Ad Hoc Liaison Committee was created. Chaired by Norway and with the Bank as Secretariat, the AHLC functioned as the main policy coordination mechanism for the reconstruction and development program in WBG. In addition, the usual Bank-chaired Consultative Group (CG) meetings took place. With the shift of much of the operational decision-making to the field, the AHLC created in November 1994 a local Joint Liaison Committee. The JLC, chaired by the Palestinian Minister of Planning and International Cooperation, served mainly to monitor the implementation of the Tripartite Action Plan agreed between the Palestinians, Israel and donors after the first CG in September 1994. The customary aid coordination committee composed of local representatives of donors dealt with operational coordination in the field. In the case of WBG, the local committee included representatives of Israel and the Palestinian authority, and was co-chaired by the local representatives of Norway, the UN and the Bank. Finally, in 1995, the aid coordination committee established 12 sector-working groups, many of which were ineffective (as later in Bosnia). This structure was replaced in 1999 by a simpler set of four groups corresponding to the main sectors in the Palestinian Development Plan: infrastructure, productive sectors, social sectors, and institution building. Confronted with a most unusual set of donor coordination challenges, the Bank responded well and with sensitive handling of politically charged situations. After the extraordinary rush of work and accomplishments of the last 10 weeks of 1993, workable consultation arrangements were quickly put in place and, by and large, donors became satisfied with the Bank's work as AHLC Secretariat, cooperation in the field, and response to the periodic crises that were the hallmark of life in Palestine.12 Thus, for example, in response to the failure of the September 1994 CG from a Palestinian-Israeli dispute over proposed Palestinian investments in Jerusalem, the Bank facilitated the formation of the first Tripartite Action Plan, by which donors would make firm commitments in return for coordinated Israeli and Palestinian actions to deal with tax and expenditure issues. The Technical Assistance Trust Fund (TATF)13 The TATF, established in October 1993, received contributions totaling about $23 million from 12 bilateral donors. The original closing date of June 1997 was extended to 2001. The TATF was designed carefully to finance non-project related TA, and thus went mainly for institution building. Partly to be responsive to Palestinian wishes, it included a wide variety of activities in 15 sectors. The TATF was invaluable in some respects, notably, in setting the general framework, financing a number of feasibility studies that led to donor-funded projects, and supporting some of the institution-building interventions. However, it also proved to be and unwieldy instrument and costly to administer (with almost 10% in staff 12 As confirmed by non-Bank sources. 13 See the TATF Grant Agreement in Annex 1. 6 costs.)14 In general, the effectiveness of TA projects depends largely on: (i) flexibility, within broadly defined criteria and priorities; and (ii) supervision continuity and resources.15 The latter existed in the TATF, with competent and dedicated Bank management; but the former did not. The TATF experience indicates both the need for a reasonably comprehensive framework and the risk of over-specifying it in advance. The experience does not indicate, however, whether it would be preferable to have a separate TA multi-donor trust fund or to include instead a non-project TA component into an umbrella MDTF--as was done later in Afghanistan.In either case, as elaborated later, good coordination and explicit links to local capacity-building are essential. The Trust Fund for Gaza and the West Bank (TFGWB)16 The Trust Fund for Gaza and the West Bank (TFGWB), established in November 1993, received a total of $380 million from Bank surplus, in four replenishments between November 1993 and July 1999. It was the main vehicle in the early days to finance emergency projects--the first two of which were the Emergency Rehabilitation Project and the Education and Health Rehabilitation Project, together accounting for the first $50 million transfer. The importance for the momentum of the peace process of quick wins and visible results on the ground was understood by all sides from the very beginning. However, when private investment proved much weaker than envisaged, after the replenishment of late 1995 the TFGWB placed even greater emphasis on employment generation and community development, with a focus on small-scale infrastructure. This emphasis was to become a standard feature of later assistance to post-conflict reconstruction. The Holst Fund17 As the Holst Fund was the first major umbrella MDTF in a post-conflict situation, had a major positive impact, and became a model for subsequent initiatives, it is worth describing in greater detail. Genesis and evolution The proposal for an MDTF to finance the start-up costs of the new Palestinian Authority (PA) originated from a member of the October 1993 assessment mission. After fleshing out the proposal, and obtaining a preliminary positive legal opinion, the proposal found its way to senior Bank management. It was originally turned down, because it would set a new precedent with reputational and fiduciary risks for the Bank (as trustee). Unlike trust funds for investment purposes, which are subject to normal Bank safeguards, procurement and disbursement rules, the fungibility of aid entails special risks in financing recurrent costs. In addition, there was no precedent for quick-disbursing loans for purposes other than balance-of-payments financing. At the time, the now-accepted rationale for adjustment lending as budget support--without necessarily a foreign exchange need--was not yet in place. Moreover, the WBG situation certainly did not appear to be amenable to conventional conditionality--a point discussed later. In any event, the argument for a new flexible mechanism to finance the infant PA prevailed. The proposal was eventually approved by President Preston, albeit with new safeguards; it was floated at the January 1994 donor meeting, which endorsed it; and the Grant Agreement was signed on April 19,1994. In July 1995, the Fund was broadened into a mechanism for overall budget support, and substantial additional pledges were received. The Holst Fund closed on August 31, 2001--having provided the 14 A major contributing difficulty was posed by the demands and management style of the PECDAR office in charge of TA, a reminder that much of the effectiveness of post-conflict reconstruction arrangements is a function of individual personalities, and has nothing to do with design or systemic factors. 15See OED's March 2002 report, op cit., p.29. 16The TFGWB provisions were standard as in other trust funds financed from Bank surplus, and are not annexed. 17See the Grant Agreement for Financial Assistance to Cover Start-Up Expenditures in the West Bank and Gaza (Document TF26074 of April 19, 1994), and its amendment of 20 July 1995. 7 financial backbone of the PA through the first two years of extreme uncertainty on revenue and expenditure needs. Governance and implementing agency The Holst Fund was established as a trust fund of IDA, constituted of the contributions from bilateral donors as well as $12 million transfer from Bank surplus and income from investing the contributions. As in all other MDTFs, IDA was the Administrator (trustee), responsible for managing and using the funds and for entering into contribution agreements with each donor. PECDAR (see below) was the main implementing agency for the PA. Contributions from donors were converted into US dollars and commingled in a pooled Holst Fund trust account. Allocation decisions were made by the Bank, in consultation with PECDAR, based on the disbursement categories specified in the grant agreement. Among the precedents established by the Fund was that, while the funds were commingled, the Bank would keep a record of each donor's "preferred" contributions, and report to the donor community on aggregate expenditure by sectors--without allowing any earmarking as such. Administrative costs totaled about 4% of funds contributed, with fees of 2% each for the Bank and the Agent (see below). Resources and uses The Holst Fund initially received total pledges of $19 million from eight donors. By the time it closed in August 2001, it had received contributions from 26 countries and disbursed almost $295 million. The three main components were: recurrent budget support, initially for the ministries of education and health; job creation, through the Emergency Employment Generation Program; and microprojects.18 Recurrent budget support was by far the largest. At first, the Holst Fund financed only start-up costs, i.e. supplies for education and health, and salaries for education, health, treasury, tax administration and telecommunications. In July 1995, donors provided fresh contributions, and the Fund was re-engineered into a mechanism for overall budget support, financing over the subsequent 18 months virtually the entirety of Palestinian central administration salaries (over 70% of the total), and a good share of operations and maintenance expenditure--plus lump sum transfers for municipalities and welfare payments. Recurrent budget support declined sharply after mid-1996, with the improvement in the PA's own revenue administration, and virtually ceased by end-1997. Overall, the Fund provided $222 million in recurrent budget support, almost all of it in its first two years. Administration and procedures By comparison to some later experiences, the six-month time lag between the agreement on setting up the Fund and the beginning of disbursements appears long. This is explained by the natural reluctance of some donors to contribute to a novel mechanism, and the imperative of assuring that strong fiduciary safeguards should be in place before any moneys changed hands. These procedures had to be formulated from scratch, mostly without the benefit of precedent, and had to go through internal clearances as well as discussion with donors.19 Taking advantage of the precedent set and the tools elaborated for the Holst Fund, trust funds in subsequent cases were established and began disbursing somewhat more quickly. 18The only controversial issue concerned the payments for the widows and orphans of killed Palestinian fighters (the so-called Martyrs Fund). On the one hand, these payments could be viewed as creating a "moral hazard" insofar as they could be construed as encouraging violent activities. (But note that, while there would later be Hamas attacks against Israeli civilians during the 1996 electoral campaign, "suicide bombings" were unknown at the time, and began only in the second year of the second intifada.) On the other hand, these payments were also support for indigent families. In the event, the PA was persuaded not to use Fund resources for this purpose, Norway picked up the bill through its bilateral program, and the line-item in the Holst Fund remained blank. 19 A much shorter time lag was achieved in DDR cases and in MDTFs for specific projects or sectors. 8 The guiding principle of Holst Fund administration was maximum responsiveness consistent with accountability and transparency, as the fiduciary risks associated with financing salaries and other recurrent costs are greater than for normal investment operations. Among the standard safeguards were requirements that PECDAR, as implementing agency, maintain accounts and records consistent with sound accounting practices; retain an independent auditor acceptable to the Bank, to develop internal controls and submit an annual audit report within six months of the end of the year; and provide monthly progress reports to the Bank. PECDAR's performance in these respects was satisfactory. Procurement procedures were consistent with Bank rules but provided for some additional flexibility and discretion. Thus, purchases for more than $200,000 could be made by international shopping, and those for smaller amounts by national shopping. Services for the Gaza cleanup operation could be directly contracted if procedures were judged acceptable. As usual, consultants were selected in accordance with Bank guidelines. In any event, most of the withdrawals from the fund were for salaries, and the financial control function in that respect was exercised by the Holst Agent. Disbursements and the role of the agent In addition to the nature of the expenditures, the absence of established financial authorities and of a sound financial management and control system in Palestine called for special arrangements. The solution was to hire, through competitive bidding, an international firm as Agent to monitor financial management by PECDAR as well as help develop PECDAR's own financial management capacity. The Agent, with offices in both Gaza and the West Bank, monitored day-to-day transactions and screened withdrawal applications. Funds for eligible expenditures were provided from the Holst special account either as advances or reimbursements, upon applications for withdrawals signed by the President of the PA (Arafat) and countersigned by the Managing Director of PECDAR (at first, Abu Alaa, the current leader of the legislative assembly). The withdrawal applications were sent to the Agent, who prepared a report for each application, validating the expenditure, and submitted the application to the Bank's Loan Department for processing and payment. The special risks of the situation were also addressed by the unusual requirement that the Holst Fund task manager review and clear all withdrawal applications prior to submission to LOA.20 Aid Management and Implementation Arrangements--PECDAR Not only were there no Palestinian governmental institutions in late 1993, but none would be permitted by the Israeli Government except within the strict context of an agreed roadmap within the Oslo peace process. The situation after the White House handshake thus called for a Palestinian agency that could serve three purposes: aid management, project implementation, and proto-government. The quality of human resources was high and did not pose a constraint. The dilemma was instead how to put in place a strong capacity to fulfill these purposes without thereby pre-empting longer-term development of regular institutions of government. The strategic solution was to avoid an institutional enclave while creating a "nucleus of efficiency"21 which could: (i) regulate and coordinate the flow of assistance, on a semi- permanent basis; (ii) implement projects directly, in a first phase, and then serve as implementation services provider for a regular Palestinian government; and (iii) incubate the nuclei of sectoral ministries, until such time as the progress of the peace process would allow. Intrinsic to the nucleus of efficiency concept was the expectation, embodied in an explicit sunset clause, that the agency's functions and staff would progressively be spun off to regular organs of government, and that the agency itself would go out of business in three years, by the end of 1996. 20 This special requirement was dispensed with in the later case of the Afghanistan Reconstruction Trust Fund. 21 For an elaboration of the "nucleus of efficiency" concept, see Schiavo-Campo, 1994. 9 Accordingly, at the end of the October mission, agreement was reached to establish PECDAR, with a clear distinction between the policy-guidance function (exercised by the Chairman of the PA and the PECDAR Board of Governors) and the managerial and fiduciary function (exercised by the Managing Director of PECDAR). PECDAR had six functional offices. The office of economic analysis was the nucleus of an eventual ministry of planning; the office of program formulation was the nucleus of an eventual ministry of finance; the aid management and technical assistance offices coordinated project aid and non-project TA, respectively; the NGO office aimed at facilitating an NGO role in project implementation but also at bringing under unified monitoring a variety of disparate organizations with different agendas; and the project management office was the operational heart of the organization. The Bank also formulated a full procedural, financial management and personnel handbook, which was adopted by PECDAR and generally followed as the basis for all of its activities.22 The consensus of all sides is that PECDAR served very well, and was efficient, competent and transparent in its use of funds.23 However, PECDAR's TA office was nowhere as constructive as the rest of the organization (a main reason for the mixed record of the TA Trust Fund), and the NGO office controlled NGOs more than it facilitated or monitored their operations.24 Strategically, too, the original conception was not fully implemented: the economics and program offices did go out of business promptly when the Ministries of Planning and Finance were established, and the aid management function was (partly) transferred to the Ministry of Planning, as envisaged. But PECDAR continued to manage projects directly instead of progressively devolving responsibility to the competent line ministries. As a result, instead of putting itself out of business by end-1996 as planned, as of May 2003 PECDAR is still implementing donor-funded projects. The reasons are explained below. Accomplishments and Problems By dint of circumstance, the Bank found itself in a role more central in WBG than in any other major post-conflict situation, before or since, but managed to steer international support for Palestinian reconstruction and institutions around the difficult political shoals, without losing credibility and earning the trust of external and internal parties. Considering the intensity of mutual suspicion between Israeli and Palestinians and the mixed agendas of other major donors, this is a noteworthy achievement that came at a very opportune time--when criticism of the Bank and of the value of aid in general was mounting.25 Accomplishments and Major Success Factors Capable counterparts and high ownership A first major reason for this success was the presence of competent and highly motivated counterparts, beginning with the Palestinian committees assembled to work with the needs assessment mission of October 1993. As noted earlier, ownership and human resources were never a problem in WBG. 22 The reader interested in the structure of an aid management agency in a nation-building situation is referred to the earlier Palestinian example rather than the recent Afghanistan Aid Coordination Authority (AACA), as the organizational arrangements of the latter have not been spelled out in comparable detail and remain in a state of flux. 23As concluded by the OED March 2002 report. However, the report badly understates PECDAR's role as just an implementing agency for aid-assisted projects, and states incorrectly that the model for PECDAR was the Lebanese Council for Development and Reconstruction (CDR). Not only were the two agencies developed independently, but they had very little in common except for their name--which was chosen by Arafat to take public relations advantage of the "go-getter" reputation of Lebanon's CDR in its first couple of years. 24It must be noted as part of the complex realities of Palestine that several of the more active social service NGOs were associated with Hamas (an important reason for Hamas' popularity among poorer Palestinians), and closer PA monitoring of their activities was viewed as part of the overall internal political dynamics. 25Through the "Fifty Years is Enough" campaign and from other sources. 10 Excessive centralization of decision-making and inadequate external influence on transparency and participation were a different matter, as discussed below. Institutional successes Institutional development was at the top of the Bank agenda from the first October mission. Indeed, as the notion of an independent Palestinian state was still anathema in Israel, only the firmness of that mission succeeded in overcoming the initial strong reluctance of the Israeli side to allow--let alone encourage--the development of autonomous Palestinian public institutions. And develop they did, largely through the implementation of the reconstruction and development program itself. The institutional scorecard is mixed, however. The brightest spots were the creation of an efficient and business-like aid management and implementation agency; capacity-building in public health and education; and support for local government structures and service delivery. The failures are noted in the next section. Robust but flexible procedures The device of employing an agent for the Holst Fund worked well. Accounting, disbursements, and reporting to both the Bank and the donors were carried out on a timely and efficient manner, and with controls adequate to prevent most fraud and misallocation. An enabling Bank environment and external support The Bank team also had the right skills mix, and enjoyed an excellent working relation with counterparts as well as easy and frequent access to the top Palestinian leadership. Moreover, Bank senior management created the proverbial enabling environment that permitted staff to take certain calculated risks and explore innovative and flexible modalities of assistance. These assets were both cause and effect of the trust placed by the international community in the Bank, and external evaluations of Bank assistance were uniformly positive, including from the US General Accounting Office. Finally, the political importance of fostering peace in the Middle East gave the Bank an invaluable underpinning of support from the major international actors, including the US. However, this support cut two ways, and weakened the Bank's position in respect to some key requirements of aid effectiveness, as discussed next. Problems and Issues Institutional failures After the creation of PECDAR and, later, of the core ministries (Finance and Planning) and of essential line ministries such as education and health, the number of ministries in the PA proliferated, and so did government employment. Employment in the central administration grew from about 23,000 at the time of the Oslo Accords (in the so-called Israeli Civil Administration, which was staffed mostly by Palestinians) to well over 100,000 by the end of the decade, or about 4% of the population--high even by the bloated norms of government employment in the Middle East. While as a result government salaries absorbed excessive resources, neither administrative capacity nor integrity kept pace with this expansion--rather the reverse. As noted earlier, the limited capacity of most sector ministries prevented PECDAR from shedding functions to the regular organs of government, as had been envisaged. In turn, PECDAR's own efficiency, paradoxically, weakened donor incentives to demand better performance from the regular ministries. Such over expansion of government machinery is hardly unprecedented in the Middle East and might be explained (while not justified), by political pressures on the PA and the sorry state of the labor market after 1996. However, these considerations do not absolve the Bank and the international community from the responsibility to weigh in more forcefully on the side of fiscal responsibility, administrative 11 effectiveness, and public integrity. It is unwarranted to say, as some observers have done, that the Holst Fund enabled this over expansion. By the end of 1996, when the Holst Fund had virtually ceased to finance PA recurrent costs, PA employment was around 68,000--at 2.5% of population still an acceptable level. However, it is warranted to point to the failure to use the Holst Fund as a lever to slow down the early administrative growth, encourage a medium-term perspective and, most importantly, induce the PA to move in the direction of transparent and meritocratic recruitment and personnel management. The same can be said of the failure to encourage a more participatory budgeting process and the open legislative scrutiny of the budget that is an essential ingredient of sound public sector management. This is not hindsight. Many Palestinian decision-makers had voiced these concerns from the very start of the process, and some Bank staff did bring the troublesome developments to the table, and argued for a more muscular use of the tools of external assistance in the interest of long-run institutional development in Palestine, but without any success. This leads to next point. The limits of conditionality The political importance of facilitating peace in the Middle East by supporting economic recovery and financing the PA posed a severe constraint on the Bank's ability to insist on certain policy conditions. As mentioned, Bank staff were well aware of the need for setting early standards--especially in public administration and economic management--and insisting on their enforcement. Indeed, in the early months, this awareness did translate into clarity, firmness and results. Thus, supported by Norway (the host of the peace process) and the US, the Bank stood firm on the requirement that PECDAR had to operate on business principles and free from political patronage and interference. The dialogue with the PLO on this subject was certainly not easy, but support from senior Bank management was strong, and the outcome was a workable distinction between policy guidance by the leadership and the managerial autonomy of the organization-- a key ingredient in PECDAR's subsequent success. Shortly thereafter, however, Bank attempts to reach clear policy agreements with the two parties and stand by them--in the interests of sustainable fiscal and institutional development--were frustrated by external political pressures to prevent any interruption in the flow of financial support. Nowhere was this more evident, nor more damaging in the long run, than in the Bank's failure to condition continuation of its assistance on an acceptable PA budget reflecting key economic and social priorities--publicly debated and approved by the Palestinian legislature--and on limiting the proliferation of governmental structures and employment. As noted, pressures in those directions were not unique to the PA. Indeed, they are typical of countries in transition and of new governments looking for short-term support. Nor can one underestimate the huge impact of the uncertainties and obstacles created from the Israeli side, especially after 1995. In other situations, however, the Bank was able to stand firm on basic policies. In the high profile, pressure-cooker circumstances of WBG, the Bank could not succeed in doing so, and the short- term urgencies won out over the longer-term aims.26 West Bank and Gaza: A Chronology 1992-93: First negotiations between Israel and the PLO, in Madrid--albeit indirectly, via the device of a Palestinian delegation sanctioned by the PLO but without PLO members. The negotiations proceed in formalistic fashion and without results. 1993: Sparked by private efforts of Israeli and Palestinian academics, and hosted by the Norwegian Government, secret direct talks take place in Oslo, eventually producing an agreement on a road map for peace, entailing recognition of Israel and self-determination for the Palestinians in WBG. 26This outcome was perhaps to be expected, but is no less troublesome. It is sobering to reflect on what might have developed in terms of public support for the PA, its image and credibility (and thus capacity to take risks, as in 2000), if it had been induced from the early days to develop a transparent, responsive, and efficient administration. An open and critical debate on the annual budget would have been the best single tool to produce such an outcome, and the Holst Fund was available as a stout operational handle. 12 Sept. 1993: The Oslo Agreement is sanctioned by the handshake on White House lawn between Yitzhak Rabin and Yassir Arafat. Oct.: A Bank mission in WBG works together with a large Palestinian delegation (liaising with Israeli authorities), to produce a comprehensive needs assessment and Emergency Assistance Program; PECDAR is established; an Ad Hoc Liaison Committee (AHLC) of donors is set-up, chaired by Norway and with the Bank as Secretariat; the Board approves the multi-donor TATF. Nov.: Board approves the Bank-financed Trust Fund for Gaza and West Bank (TFGWB). Jan. 1994: Major donor meeting in Paris endorses the Emergency Assistance Program and pledges $2.1 billion for its implementation; Bank Board approves start-up costs multi-donor trust fund, renamed Holst Fund in honor of the late Norwegian Foreign Minister. May: Israeli-Palestinian Cairo Agreement. July: Israel-Jordan Peace Treaty signed. Sept: First formal CG meeting for WBG ends in disarray; as a result, however, the Joint Tripartite Action Plan is formulated, and forms the basis for subsequent agreements on aid and institutional development. July 1995: Holst Fund is broadened into a comprehensive recurrent cost trust fund. Sept. 1995: Interim Israel-Palestinian Accords. Nov 1995: Rabin is assassinated. 1996: Likud wins Israeli elections after a campaign marred by Hamas attacks on civilians in Israel; Netanyahu becomes Prime Minister; Oslo peace process de facto stops. BOSNIA AND HERZEGOVINA27 The Authorizing Environment The Context The Balkan tragedy is too well known to be recounted here in any detail. It is fundamental to realize, however, that the fashionable explanation of "ancestral ethnic hatreds"--for so long used as an alibi for non-intervention--is wholly inconsistent with historical realities. While transitory resentment was generated in the 15th and 16th centuries against the inhabitants of the peninsula who chose to convert to Islam after the Ottoman conquest,28 the emergence of national sub-identities among South Slavs dates only from the latter 19th century, concurrent with the rise of nationalism throughout Europe. Moreover, while Muslims, Catholics Croats and Orthodox Serbs were a majority in different regions of the Balkans, substantial population movements occurred and intermarriage was common, especially in Bosnia. Indeed, in contemporary times, Bosnia was the only constituent republic of the former Yugoslavia that was not nationality-based. In reality, the mythical ancestral ethnic hatreds were the contemporary product of a deliberate policy of separation and fear-mongering pursued by political leaders (primarily, of course, Serbia's Milosevic and Croatia's Tudjman, but to a much lesser degree by Bosnia's Izebegovic as well) for personal power agendas. 27 This section draws heavily from personal communications with key Bank staff, listed in the annex, and the excellent Bosnia and Herzegovina: Post-Conflict Reconstruction (Country Case Study Series, OED, World Bank, Washington, 2000). Also useful were Susan Woodward's Bosnia and Herzegovina: Priorities for Recovery and Growth (Washington, Brookings, 1995); the World Bank's Bosnia and Herzegovina: Towards Economic Recovery (Discussion Paper No.1, Report 15668, 1996); and the relevant sections of The World Bank's Experience with Post- Conflict Reconstruction (OED, Washington, 1998.) Before the break-up of Yugoslavia, Herzegovina was only a province of the republic of Bosnia. After the war, the Croat-Bosniac hostility led to renaming the new state Bosnia and Herzegovina, normally abbreviated as BiH. 28 While in those days non-Muslims were generally treated far better under Ottoman rule than religious minorities in Christian countries, they were subject nevertheless to significant discrimination in taxation and employment. 13 Unfortunately, by the time of the Dayton (Ohio) Accords of November 1995 that ended the Bosnian conflict, those ethnic hatreds had been made all too real by 12 years of manufactured paranoia after Tito's death in 1980, and the 44 months of brutal conflict among Croats, Serbs and Bosniacs (Muslim Bosnians) after Bosnia's declaration of independence in February 1992. The Dayton agreement thus had to balance the rejection of the national fundamentalism that had produced ethnic cleansing with the reality that, as a result of the conflict, the three ethnicities had become heavily concentrated in different areas. The new structure of Bosnia and Herzegovina (BiH) reflects this balancing act. The extremely weak central state combines two entities: the Serb Republika Srpska (RS) and the Bosniac-Croat Federation.29 However, the Federation itself was (and is) sharply divided between the six Bosniac cantons (provinces) and the two Croat cantons--cleavage best exemplified in the destruction of the 16th century bridge between West and East Mostar.30 Moreover, substantial responsibilities (but not necessarily resources) are devolved to municipalities--in the Federation and the RS. This extraordinary political and economic fragmentation conditioned virtually every aspect of post-conflict reconstruction and the assistance to finance it, including through trust funds. Some observers believe that, for the sake of precluding a resurgence of the conflict, Dayton went too far in the direction of accepting the reconfiguration generated by the war. In this view, the Dayton provisions precluded from the very start possible openings for inter-entity (and interethnic) cooperation. However, it is important to recall that, until the fall of Milosevic, the Republika Srpska looked to Belgrade for support and had neither interest nor incentive to work together with the Federation or indeed with the BiH state. The same was true of the Croat cantons in the Federation, which until Tudjman's death received substantial aid from Croatia and were actively dissuaded from cooperating with the Bosniacs.31 It is fair to say that under any organizational architecture of the state the reconstruction effort and the funding arrangements would still have had to contend with the reality of extreme interethnic hostility and administrative fragmentation between entities, among (and within) cantons, and vis-à-vis municipalities. That reality had its roots partly in the suspicion that a viable central government would reinstate the Yugoslav tradition of cross-subsidies providing greater benefits to the less well-off majority group. Aggravated by the war, such suspicion reached such a level that even areas objectively poorer feared that they would lose from a redistribution system. Thus, assigning most expenditure and revenue to the sub- state levels was the only arrangement that could receive agreement from all of the involved parties.32 Under the circumstances, and taking the long view, the challenge was handled well by the international community. Partly because of the reconstruction effort itself, as of 2003 the various parties in BiH were neither fighting nor still looking outward for a solution of their difficulties. The disappearance from the scene of the two major external problems--Milosevic and Tudjiman--and the recognition of common economic interests might yet produce in the near future a beginning of cooperation and a rebuilding of inter-ethnic trust.33 In his context, it is essential to recognize that economic reforms taken as granted elsewhere owing to their aggregate efficiency benefits must, in this part of the world, be shaped by 29The economic authority of the center is limited to regulation of inter-entity transport and communications, external debt, and foreign trade and customs policy. However, because tax and customs administration is reserved to the Serb and Bosniac-Croat entities, the central state authority in foreign trade policy is partly theoretical. 30The remaining two of the ten cantons of the Federation are ethnically mixed. 31Most striking was the contrast in 1996 between the dilapidated offices of the official (Bosniac) Governor of Mostar, and the brand new facilities and equipment found across the river in the Croat area of the city , many still in cartons with Zagreb stamps. 32I am indebted to Christine Wallich for this argument, which is part of her forthcoming paper with William Fox "Fiscal Federalism in Bosnia-Herzegovina: Subsidiarity and Solidarity in a Three-Nation State". 33The interethnic dialogue deliberately fostered through the preparation of the recent Bosnia: Public Expenditure and Institutional Review (World Bank, 2002) could be a step in that direction. 14 distributional imperatives.34 The pursuit of inclusiveness in the reform process may be far more important than their proximate outcome. The strategic principle for future Bosnian recovery and development can thus be formulated as "infant social capital" protection. In any event, the core feature shaping programming and funding of post-conflict reconstruction in BiH was fragmentation, and the main challenge was to recognize the reality of dysfunctional fragmentation without thereby strengthening it. The Mandate for International Aid and Key Milestones Perhaps the most distinctive feature of the process of early planning for reconstruction assistance in BiH was the constructive role played by the Dutch Executive Director of the World Bank. The personal interest and commitment of this ED facilitated the creation of the Bosnia Working Group at the Bank, and the subsequent start of planning activities for post-conflict reconstruction shortly after the Bosniac-Croat ceasefire of March 1994--even as the Serb-Bosniac conflict continued unabated. The result was threefold. First, Dutch grants were made available (later followed by a $10 million Dutch Trust Fund) to finance project preparation and pilot projects--crucial at a time when no legal way existed for the Bank to finance such activities. Second, the Bank's early planning for reconstruction caused it to be invited to participate in the political settlement negotiations--unprecedented then and since. In turn, Bank attendance to delicate political negotiations--although not as an equal party--strengthened the essential link between the political and the economic. Third, the head start gained in 1995 made it possible for the Bank to hit the ground running after the signing of the agreement on December 14, leading a needs assessment mission jointly with the IMF, EBRD, EC and USAID, and co-chairing (with the EC) the First Donor Conference for Bosnia in December 20-21--barely five weeks after the end of the conflict. Donor Coordination, Reconstruction Funding, and Aid Management Donor Coordination The overall guidance of the reconstruction process rested, at the apex, on a Peace Implementation Council (PIC), comprising the parties to the Dayton agreement; a PIC Steering Board comprising a subgroup of the parties; and the Office of the High Representative (OHR), which had final authority (confirmed by the UN Security Council) over all civilian aspects of the agreement. Concerning the overall aid program in BiH, the leadership was shared between the Bank and the EU, which cooperated with the Bank in organizing and co-chairing the donor conferences. The mobilization of assistance for BiH reconstruction was successful, but donor cooperation much less so, despite lip service and generally good personal relationships. A particular source of tension was that OHR and the EU had political mandates in BiH, whereas the Bank only an economic mandate. Thus, the OHR sometimes wanted to condition aid on political steps (peace conditionality), or design sector programs to achieve political objectives that were beyond to the Bank's mandate (as distinct from promoting creation of social capital, which the Bank can and did support).35 As explained later, in light of the diversity of donor interests and procedures, an attempt was made by the Bank to improve the coherence of assistance by designing a full set of sectoral interventions (projects), consistent with an overall blueprint for reconstruction, to serve as receptacles for donor resources. Each 34 Reportedly, this reality was not recognized in the formulation of the IMF policies vis -à-vis the former Yugoslav republics. It is doubtful that this factor aggravated inter-group hostilities, but it certainly did not make for effective policy implementation 35 From the OHR viewpoint, achievement of the political objectives was indispensable to the sustainability of development, and the Bank position was seen as narrow. A somewhat similar tension existed in Kosovo, between the Bank and the UN Mission in Kosovo (UNMIK), which was acting as de facto local government despite the formal existence of the Kosovar Provisional Institutions of Self-Government. 15 sectoral project was managed by a PIU working with the relevant counterpart BiH agency.36 Some donors did channel their assistance through the PIUs, but many did not. The overall coordination for each sector was to be provided by 11 sector task forces, with each task force chaired by a different lead donor.37 Because the Bank chaired only the economic policy task force (co- chaired by the IMF), it was unable to exercise the same facilitating and coordinating role at sector level that it had exercised in organizing donor conferences and mobilizing funds. The 11 sectoral task forces were instead supposed to be coordinated by the Economic Task Force (ETF), chaired by OHR--the highest international political authority in Bosnia. The role of the ETF in aid coordination, however, remained loose and ambiguous, and so did the relationship between the timing of the political agenda and that of the reconstruction program. The Bank intentionally chaired only one task force in order to avoid appearing heavy-handed. However, its representatives on the task forces were active and, in some cases, the de facto shadow chair. Still, these complicated arrangements were not conducive to clear responsibility and accountability, and gave bilateral donors neither a positive nor a negative incentive to act in concert. Moreover, some of the sector lead donors proved unwilling or unable to exercise a robust coordinating role. Not surprisingly, several donors opted to act independently--in project selection, choice of local interlocutor, and procurement and fiduciary procedures. Good results were achieved in some sectors, less so in others. Important externalities were not addressed; bottlenecks emerged from uncoordinated sequencing; and, partly as a result, Bosnian authorities' ownership of the aid effort was weak.38 As often the case, the fragmentation of donor efforts mirrored the fragmentation in the recipient government. The partial saving grace was threefold. First, it appears that Bank relations with NGOs were fairly good--partly due to outreach and local initiative elements in Bank projects. Second, security on the ground was assured by SFOR (the international military stabilization force) and its cooperation with the Bank was unusually close. The stable and predictable security climate was undoubtedly a key factor in the good implementation record of Bank (and other) projects--just as security problems have recently hampered aid operations in Afghanistan. Finally, substantial capacity building was achieved in some sectors through the PIUs themselves, which were largely staffed by locals. The Trust Fund for Bosnia and Herzegovina Following the First Donor Conference, and the creation of the $10 million Dutch Trust Fund, the Board approved in February 1996 the Trust Fund for BiH (TFBH), for $150 million, of which $25 million grant and $125 million on IDA terms. Except for a small recurrent cost component, the TFBH financed only investments. Administered by the Bank, it was designed as a potential MDTF, and could accept contributions from other donors on terms to be agreed between the Bank and each contributor. However, the creation of the TFBH was not subject to the background discussions and negotiations that have preceded other MDTFs. Bank staff and country management did try to persuade donors to contribute to a common fund, but convinced only three small bilaterals. The main factor for this reluctance was in all probability the unusual weight of political and security elements in most donors' agendas. 36 Serious consideration had been given to implementing projects through an agent--as in the Holst Fund and later the Afghanistan Reconstruction Trust Fund. The PIU route was chosen mainly to build local capacity. 37 Four task forces were chaired by the International Management Group, one each by the ILO, USAID and the then-ODA, three by UN agencies, and one by the Bank. 38 This sectoral approach to coordination was reminiscent of the Sector Investment Program approach which was briefly popular in the late 1990s as second-best substitute for a comprehensive public investment program. 16 The Bank's fallback position was to try to inject a measure of coherence in the assistance program by a sectoral approach to co-financing, setting up a multiplicity of trust funds in major sectors (eventually totaling over 80, including multiple TFs from individual donors). Although transaction costs were especially heavy for the Bank staff concerned,39 the notion that a "flag" trust fund for a specific purpose in each of the major sectors could serve as co-financing magnet for donors interested in that purpose and sector proved generally valid. Whether directly associated with TFBH-financed activities, or as coordinated funding emerging from the frequent contacts between the Bank and other donors, co- financing was substantial. Thus, for example, of the $160 million Emergency Recovery Project, $45 million came from the TFBH and $115 million from the EU, Netherlands, Canada, and other bilaterals. And the EBRD provided parallel financing for the Emergency Transport Reconstruction Project, in the form of aid to rehabilitate and reopen Sarajevo airport. The beneficiary of the TFBH was very broadly defined as either BiH or "appropriate public and private entities", and the purpose was "emergency reconstruction". The largest single investment was $45 million for the Emergency Recovery Project, which covered a variety of urgent interventions, including a positive list of imports, lines of credit, support for core government institutions, and a social fund for transitional cash assistance to the poorest households. The remaining $115 million were invested in more conventional projects (insofar as any assistance to BiH in the early years could be called conventional).40 All TFBH funds were to be committed within two years. It is remarkable--and again a testament to the importance of early planning--that the seven major projects, accounting for virtually the entirety of TFBH funds, were prepared and approved within just 4 months from the establishment of the fund. Procurement procedures within each project, proved workable, and disbursements were rapid.41 Moreover, the PIUs established for each project were reportedly effective. As a result, OED found the projects to have been successful, especially in reaching intended beneficiaries. A major contributing factor was--as in other post-conflict cases--the element of outreach and reliance on local initiative. Through the Emergency Recovery Project the TFBH in practice also financed a small amount of recurrent costs (including a salary top-up, on a declining basis, for state and entity employees.) However, quick- disbursing balance-of-payments/budget support was provided mainly through a $90 million IDA Transition Assistance Credit approved six months later. Some observers believe that it would have been far more timely and effective for institution building to provide adequate budget support earlier through the TFBH, using the start-up cost precedent set by the Holst Fund for WBG. Whether the argument is valid or not, it is clear that the success of the emergency interventions was not matched by comparable success in institution-building and overall sustainability. Once again, however, the extraordinary internal animosities and suspicions in BiH make it difficult to imagine a different outcome in the early years after cessation of such brutal conflict. Aid Management and Implementation Arrangements Given the logic of BiH political architecture, strong aid management from the center was out of the question. Indeed, there were essentially three parallel aid administration and implementation systems, one for each nationality, going through three de facto separate budget systems, and three separate debt and cash management systems and accounts. 39The administrative costs to the Bank were covered by the trust fund fees. What is at issue here is the burden on individual staff working in the Bosnia program. 40Emergency Transport Reconstruction, Emergency Farm Reconstruction, Water, Sanitation and Solid Waste, Emergency District Heating Rehabilitation, Emergency Education Reconstruction, and War Victims Rehabilitation, accounted for the balance of the funds transferred to the TFBH. 41 This was not the case of a number of separate donors' projects. 17 Nevertheless, an aid management unit was formed in the state Ministry of Foreign Affairs, and did a narrow but reasonable job in constructing an aid database and recording aid flows. Two similar units existed in the Federation and RS, but essentially only to liaise with the central unit in the collection and exchange of data. As noted, none of these units exercised an operational role. The sector PIUs filled part of the void, with the local staff in some measure facilitating aid traffic as well as implementing projects in their sector, but the only fair assessment of aid management in BiH is that it did not exist. Accomplishments and Problems It is instructive to note that the institutional fragmentation in BiH was to some extent mirrored in the aid coordination arrangements. However, just as it is impossible to be sure that the Dayton accords went too far in recognizing the reality of ethnic division, it is difficult and unfair to speculate whether aid could have been better harmonized if only the trust fund and aid coordination arrangements had been different. As mentioned, Bank staff did try, unsuccessfully, to persuade donors to come into the single trust fund tent, but it is likely that the diversity of interests of the various donors and of the constituent parts of the Bosnian state would have rendered ineffective the best financing mechanism. Nonetheless, in hindsight, it is also conceivable that stronger support from top Bank management and persistence in channeling much more external aid through the (potentially multi-donor) TFBH--with due regard for donor preferences--could have reduced somewhat the problems of coordination. The jury is still out, too, on the effectiveness of the sector task force arrangements. Persons intimately familiar with the process argue that the sector task forces provided far greater information exchange and sectoral coherence than in most settings, even non-conflict settings. The counterfactual of a much more disorderly and confusing scenario in the absence of the task forces is persuasive, considering the diverse political agendas of donors. (For example, some observers argued that Germany's objective to repatriate Bosnian refugees and reduce the budgetary burden of asylum seekers frequently collided with what other donors thought was a prior need for adequate preparation in housing and other areas.) Others believe, however, that--as in the earlier WBG case--the number of task forces was excessive and made it difficult to address externalities and proper sequencing--even when the task force provided effective donor coordination in its sector. In the event, three things are reasonably clear. First, the projects financed through the TFBH and IDA's own projects had an excellent implementation record. Second, sectoral PIUs can serve as a good vehicle for capacity building, provided that they are largely staffed by locals. Finally, the sectoral trust fund approach can be a workable second-best option when circumstances do not permit formation of an umbrella MDTF. Through its preparation of an entire portfolio of projects covering all the main sectors, the Bank provided the rationale, policy approach and project design framework that permitted other donors to put their money into implementing a coherent overall blueprint for reconstruction. Bosnia and Herzegovina: A Chronology Feb. 1992: Bosnia secedes from FRY and declares independence as Bosnia and Herzegovina, after a referendum approved by Bosnian Muslims and Croats but boycotted by Bosnian Serbs. Feb. 1992-Mar. 1994: Armed conflict among Bosnian Muslims, Croats and Serbs. Mar. 1994: Ceasefire between Bosnian Muslims (Bosniacs) and Croats, based on concept of loose federation. Conflict between Serbs and Bosniacs continues. Late 1994: Bosnia Working Group established in the Bank, with support from the Dutch ED, to prepare strategy for Bank assistance. 1995: With first planning mission in Jan. 1995, Bank begins assessing reconstruction needs and leads donor coordination, especially with EBRD, Soros Foundation, and UN agencies active on the ground. A $10 million Dutch Trust Fund for project preparation and pilot projects is established. 18 Oct. 1995: Informal discussion at Annual Meetings endorse Bank role in leading the preparation of reconstruction program. Nov. 1995: Bank invited to Dayton negotiations. Dayton Accord ends conflict and establishes state of Bosnia and Herzegovina (BiH) comprising the Bosniac-Croat Federation and the Serb republic (Republika Srpska) as largely autonomous entities. Nov. 1995: Bank leads joint mission with IMF, EC, IBRD and USAID, finalizing reconstruction needs assessment and strategy for recovery Dec. 1995: First donor conference, co-chaired by Bank and EC, to present needs assessment and mobilize pledges totaling about $600 million. Feb. 1996: Board approves Bank Trust Fund for BiH (TFBH), with $150 million from Bank surplus. No MDTF is established, but other donor contributions to the TFBH are welcome. Mar. 1996: Arrears clearance plan approved (on Bangladesh precedent). April 1, 1996: BiH becomes member of World Bank Group, retroactive to Feb 25, 1993. EAST TIMOR42 The Authorizing Environment The Context In 1978, Indonesia occupied and annexed East Timor, a poor and small Portuguese colony with a population of around 500,000. The annexation was never recognized by the United Nations and for the subsequent 20 years the Timorese Front for National Liberation (FRETILIN) conducted sporadic but persistent resistance to the Indonesian occupation. In May 1999, shortly after the fall of the Soeharto regime, the new Indonesian Government signed an agreement with Portugal (the New York Agreement), requesting the UN to supervise a referendum in East Timor on whether to accept special autonomy within Indonesia. If the referendum were to reject special autonomy in favor of independence, the agreement called for Indonesia to take all necessary steps to terminate its links with East Timor, thereby returning the territory to its pre-annexation status, and authority would be turned over to the UN in preparation for East Timor independence. The referendum held on August 30, 1999, rejected autonomy in favor of independence by a 4-1 margin. As soon as the results were announced, organized militias opposed to independence went on a rampage, with at least the tacit consent of the Indonesian troops on the island following guidance from the top military leadership. Thousands were murdered; more than half of the population was displaced internally; a further 150,000 were forcibly displaced into the Indonesian territory of West Timor; and physical damage was severe, with over 70% of public buildings on the island destroyed. After two weeks of unremitting violence, the UN Security Council authorized a multinational force in East Timor, and in October the Indonesian Assembly declared void the 1978 annexation--opening the door to the establishment of the UN Transitional Administration for East Timor (UNTAET), including a peacekeeping contingent. Two and a half years later, East Timor became independent on May 20, 2002. As in Kosovo and Bosnia, ethnic resentment between the majority inhabitants and the Indonesian settlers had been simmering; as in Afghanistan, physical destruction was severe; as in WB Gaza, political and administrative authority had been exercised by an occupying power. The similarities end here. While the 42This section relies in part on personal communications from the individuals in the annexed list. Especially useful was the comprehensive and nuanced assessment by Klaus Rohland and Sarah Cliffe (The East Timor Reconstruction Program: Successes, Problems and Tradeoffs, CPR Working Paper No.2, November 2002), and the Report of the Trustee for the Trust Fund for East Timor prepared for the December 2002 Donors' Council Meeting. Other sources are listed in the bibliography. 19 violence was of long duration (over one quarter of the Timorese population was killed between 1975 and the early 1980s, and large scale human rights abuses occurred during the 25 years of occupation), Timorese resentment was directed against the military and civilian Indonesian presence rather than rooted in internal ethnic divisions; unlike Palestine, East Timor did not possess a broad base of skilled and sophisticated human capital; and physical infrastructure had not been as developed as in Bosnia and Kosovo. External assistance in East Timor was influenced from the start by the virtual absence of any administrative capacity in the territory, the extremely weak human resource base and its poverty.43 Regrettably, as discussed below, the wrong lesson was drawn from the lack of local administrative capacity. The Mandate for International Aid and Key Milestones Unlike in Bosnia, the Bank was not a party to the political deliberations that shaped post-conflict arrangements. It attended neither in the Indonesia-Portugal New York Agreement, nor in the UN deliberations that put in place UNTAET after the September violence. However, prior to the referendum, the Bank worked closely with the UN Department of Political Affairs on economic scenarios for East Timor. The 1999 Annual Meetings provided the occasion and the venue for a first donor meeting, the Friends of East Timor meeting. Tracing the pattern already established in previous post-conflict cases, that meeting mandated the Bank to organize a joint assessment mission, which took place in October- November 1999. Unlike some other early assessments of reconstruction needs, that mission was joint in the real sense of incorporating the views and participation of a variety of donors and of Timorese counterparts. Immediately following the end of the violence and restoration of security through the establishment of UNTAET, a major donor meeting was organized in Tokyo in December, co-chaired by the Bank and UNTAET, to review the results of the assessment mission and mobilize commensurate assistance. That meeting was very successful in mobilizing pledges of assistance--almost $500 million. In the event, actual aid commitments exceeded that amount. At over $200 per capita, aid to East Timor was comparable to the high levels received by BiH and WBG. Donor Coordination, Reconstruction Funding, and Aid Management Donor Coordination Under its assistance program for Indonesia, the Bank had financed some activities in East Timor as part of projects covering the whole of Indonesia or its eastern region. Other donors had done so as well. Consequently, through the frequent donor meetings on Indonesia, some information on development problems and activities in East Timor had sporadically been exchanged. Donor coordination did not have to begin on the virtual tabula rasa that has characterized some other post-conflict situations. In the expectation that the referendum called for by the New York Agreement would produce major changes in status, in early 1999 the Bank shifted responsibility for East Timor from the Indonesia desk to the Pacific Country Department.44 Staff initiated a variety of studies and preparatory activities, including low-profile contacts with the Timorese leadership and, as noted, exchanges with the UN political bureaus, but was unsuccessful in eliciting active interest by other donors--who generally took a wait-and-see attitude. Although the violent spasm of September 1999 required a rethinking of priorities and costing, it 43After the orgy of destruction in September 1999, East Timor's per capita income can be roughly estimated at under $300. The economy was (and is) predominantly agricultural, with 90% of the population living in rural areas, and extremely low social indicators--56-year life expectancy, 41% literacy, less than 30% access to drinking water. Moreover, net transfers from Indonesia came to an end. These transfers, of about $100 million a year, financed very high public expenditures--largely for personnel costs, which accounted for almost 90% of current expenditure. 44ADB did the same a few months later, transferring responsibility for East Timor to its Office of Pacific Operations. 20 served to awaken keen donor interest, and the "dialogue infrastructure" built up in the previous months was important for quick action after the restoration of order. A key characteristic of the East Timor case is the special role of the UN. Elsewhere, the UN role has been primarily to prevent a resurgence of hostilities after the conflict was ended by international military intervention. By contrast, East Timor was not independent, and the UN thus became responsible for all administrative authority--as in Kosovo. However, local capacity in East Timor was even weaker than in Kosovo (where the core issue was to cope with and bring to the surface the parallel informal networks of service delivery and rent-seeking). Moreover, unlike Kosovo, the internal ethnic tensions had largely dissipated after the departure of the militias and the unfortunate resettlement of many of the former settlers. With the high degree of security, absence of serious internal cleavages, and near-total local administrative vacuum, UNTAET found itself enabled to administer the territory while supporting the creation of local public administration. It proved far more successful in the first than in the second challenge. Also, lack of continuity between the earlier UN mission (UNAMET)--which had co- coordinated the joint assessment mission on reconstruction needs--and UNTAET contributed to difficulties in the re-establishment of public administration. The Trust Fund for East Timor (TFET)45 As stressed later, the overall coherence of assistance was damaged by the fragmentation of funding vehicles, and particularly the separation (with little coordination) between recurrent cost financing (through the UN CFET) and sector-based reconstruction financing (through the TFET). The TFET itself largely met its original objective of anchoring the core reconstruction program, albeit with some delays attributable to lack of flexibility in procurement procedures. Genesis and evolution The first immediate action by the Bank, at the time of the Tokyo donor meeting, was Board approval in December 1999 of a $10 million Trust Fund for East Timor (TFET), funded by a transfer from Bank surplus.46 The Tokyo meeting also recommended the expansion of the TFET into an MDTF. The next key step was Board approval of the expanded TFET, on February 24, 2000. As of end-2002, almost 95% of TFET funds had been committed. In May, on the eve of independence, the Government presented a National Development Plan, the first year of which is supported by the Transition Support Program, which is expected to pick up the relay from the TFET. Governance and implementing agencies. The TFET was established as a trust fund of IDA, constituted of the contributions from bilateral and other donors as well as transfers from Bank surplus and income from investing the contributions. As in all other MDTFs, IDA was the Trustee, responsible for managing and using the funds, and for entering into contribution agreements with each donor. The TFET, however, established a precedent in that the Asian Development Bank (ADB) was assigned a substantial role as implementing agency alongside the Bank - albeit without prejudice to the Bank's responsibility as Trustee. In practice, a clear division of responsibilities was agreed, with ADB as implementing agency for physical infrastructure reconstruction and the Bank responsible for the other components of the program. A TFET Donors Council, composed of a representative from each contributor, met semi-annually to approve the work program (activities, funding, beneficiaries, and implementing agencies) submitted by the Government or, prior to 45 See for details Proposed Amendment to the Trust Fund Resolution No.99-5. For the operation of the TFET during the first two years, see the Report of the Trustee, Virtual Donors' Council Meeting of December 2002. The Report is not included in this paper, but is available on request from the author. 46 See the Board Paper World Bank Assistance to East Timor, Doc No. R99-219 of Nov 19, 1999. 21 independence, the highest Timorese consultative or executive body. This was the simplest governance arrangement of Bank-administered MDTFs in major recent post-conflict cases. Resources and uses Initial pledges to the TFET totaled just over $140 million, including the $10 million from Bank surplus. Eventually, over $176 million were contributed. The largest contributors were Portugal and the EC (with $50 million each), Japan ($28 million), Australia ($13 million), the UK and the Bank--each with about $10 million. Seven other countries accounted for the remainder. Over 20 grant agreements were negotiated, the largest of which were for $30 million for emergency infrastructure (with ADB), and for health, education, water and sanitation, and agriculture (about $80 million in total). An additional $22 million was granted to finance a set of community empowerment activities (on which more later.) The speed of preparation of the early TFET-financed projects, at less than 4 months, was second only to the experience under the TFBH in BiH. The results on the ground were comparably fast. Administration and procedures The Bank, as Trustee, had sole responsibility for administering the TFET, and received 1% of the total contributions to cover its costs.47 As administrator, the Bank was responsible for keeping records and accounts, commitments, receipts and disbursements, and reporting annually to the Bank Board and the contributors. However, the implementing agency concerned had sole responsibility for the use of TFET resources for the projects that it implemented. Each grant made by the Bank had to be administered in accordance with Bank procedures, including for consultant selection and procurement. Similarly, grants made by ADB were to be administered in accordance with the applicable procedures of ADB, including its own consultant selection and procurement rules. Aid Management and Implementation Arrangements With no national government in place, aid management from the recipient's side was in the hands of UNTAET. A potential local partner did exist, in the form of the National Council for Timorese Resistance (CNRT), the umbrella organization of pro-independence groups. However, UNTAET treated the CNRT as a political party rather than as proto-government, and kept its distance-- possibly because of the UN experience in BiH and Kosovo with the risk of being dragged into inadvertent support for sectarian interests. Moreover, the UN found itself in the anomalous position of being both the de facto government and a major participant in the process of external assistance. Eventually, some Timorese capacity for aid coordination was put in place, but after a period of tension due to understandable Timorese perceptions that they were marginalized from participation in decision-making by the UN administration.48 47Significantly lower than the standard 2% Bank fee for administering trust funds. In addition, up to 5% could be retained to cover TFET-related costs incurred by the Bank and ADB in January and February 2000 (before the establishment of the expanded TFET). 48The same role duality was present in Kosovo where, despite the existence of the PISG as political authority representing the Kosovars, UNMIK in effect exercised most of the governmental functions. In contrast with East Timor, however, some capacity existed in Kosovo: the problem was that capacity was located mainly in parallel structures of informal authority, which sometimes interacted with criminal forms of organization. As of 2003, it appears that many of these structures --and some of their leaders--have legitimized themselves and have become part of Kosovo's more formal politics. 22 Accomplishments and Problems Accomplishments and Major Success Factors Security Underpinning every good result of the reconstruction program in East Timor was the high level of security achieved and maintained by the international military contingent. The security factor is taken for granted and forgotten in cases when order has been re-established quickly, but as Sierra Leone and Afghanistan show, insecurity is by far the most severe constraint to reconstruction efforts. Good sectoral project implementation The TFET and major bilateral programs, which in most sectors were closely coordinated, generally proceeded well. The first projects were prepared in near-record time and, with reliance on local communities and NGOs, gave quick results on the ground. This contributed to preventing the gap frequently found between humanitarian and reconstruction assistance. The later, larger reconstruction projects were implemented more slowly, but not unduly so by comparison to other experiences. Unified presentation of aid requirements and benchmarking The unified presentation of aid requirements and the budget helped both in mobilizing the needed assistance and keeping a minimum of coordination in aid efforts. Equally useful was the formulation of a set of benchmarks covering the political and administrative transition, the reconstruction process, and the budget. These benchmarks were reviewed at the time of the semi-annual donor meetings and facilitated monitoring of progress and maintaining some links between different activities. In hindsight, these benchmarks were far more effective in assuring some coordination of physical reconstruction activities than in facilitating the transition to autonomous structures of self-government. However, the Timorese leadership considered the benchmarks a useful tool for Cabinet management of government activities, and continued their use after independence. Capacity-building for leadership A program of capacity building for leadership, co-sponsored by Portugal, Australia and the Bank, was successful in bringing together leaders from different factions to consider key political and development challenges. Arguably, however, this exercise should have been conducted much earlier in the transition, and been supported by a more comprehensive program of capacity building for leadership through the subsequent years. Cooperation with ADB Also helpful to ensuring some linkage among reconstruction activities was the unusually good cooperation with DB. Not only was ADB an implementing agency under the TFET, but working-level cooperation was also excellent as well, including establishment of a joint field office with the Bank.49 49A non-Bank observer attributes this unusual Bank-ADB cooperation to the common problems experienced vis -à- vis the UNTAET. 23 Problems and Issues Uneven political transition The political transition was characterized by a long period of inaction followed by an extremely rapid "end game" just before independence. From late 1999 through much of 2001, the formal role of the Timorese was restricted to consultative bodies. A later establishment of a joint "cabinet" structure was widely perceived by the Timorese as a mere token. After mid-2001 the process sped up dramatically--in large part in response to growing Timorese dissatisfaction with their marginalization and concern to bring the transition to an end. Elections for a constituent assembly took place in August 2001; the Constitution was completed in three months; and was followed in short order by elections for the Presidency. Compounding the problems from the initial slow pace of the transition, this rush at the end led to a non- acceptance by many Timorese of the provisions of the Constitution and to tensions between the Presidency (which under the Constitution is non-executive) and the executive branch of government. Fragmentation of funding mechanisms Despite the presentation of a unified budget arising from the joint assessment mission, the overall coherence of assistance was damaged by the fragmentation among five different aid vehicles (in addition to humanitarian aid, which should be included in the budget presentation but is normally handled through separate funding mechanisms): (i) the TFET, which covered sectoral investments; (ii) the UN- administered CFET, covering mainly recurrent costs; (iii) the assessed contribution budget of UNTAET; (iv) projects financed by different UN agencies; and (v) projects financed by different bilateral donors. Perhaps the most troublesome aspect of this fragmentation was the sharp separation between recurrent cost financing (through the UN CFET) and reconstruction financing (mainly through the TFET). The unified presentation of the East Timor budget from the joint assessment mission was helpful, but was not sufficient for integrating investment and recurrent decision-making. Capacity-building? The problems arising from the fragmentation of the funding mechanisms could perhaps have been alleviated by closer coordination between the Bank and the UN. The more serious problem, however, lay in the low priority given in practice to the objective of capacity building. The weak local capacity and absence of formal Timorese governmental authority during the transition to independence should have led to deliberate efforts at mobilizing systematic local participation and enabling genuine local decision- making. These realities were instead allowed to engender a disregard of the need for local ownership. Most regular functions of government were exercised by UN-financed expatriate experts. While turf infighting among expatriate experts did not reach the extremes of Kosovo, the other cardinal sin typical of technical assistance was committed in aggravated form. With a very few individual exceptions, the expatriate experts focused almost entirely on doing the jobs, instead of also gradually bringing in, training, and empowering local individuals. In part, this resulted from severe communications difficulties, with international personnel finding it hard to adapt to local languages-- Tetum or Bahasa Indonesia (the de facto languages of most civil servants) and Portuguese (the official language). The problem was exacerbated by the UNTAET decision to recruit government employees "from the bottom up"--which meant that until late 2001 there were no Timorese civil servants in senior or middle management positions, and made it practically impossible for Timorese ministers to exercise real policy influence. For its part, the Bank from the beginning involved Timorese counterparts in the planning, design and implementation of projects, and some capacity-building was accomplished through a few of the management units for TFET projects. In the absence of a strategy for building capacity in the central administration, these efforts were inevitably ad hoc and uneven. 24 Overall, at independence the Timorese were not much better prepared to run their own affairs than they had been thirty months earlier. The physical dimension of reconstruction was fairly well addressed; the human dimension was not. Avoidance of key political and developmental issues. As in many other post-conflict cases, important development challenges with political implications were avoided. In East Timor, the main ones were the treatment of war veterans and the clandestine movement, property rights and land claims. International experience provides no clear message as to the relative advantage of tackling hard issues earlier in the reconstruction process rather than after a period of consolidation. While lack of resolution of key issues inevitably interferes with the effectiveness of reconstruction efforts, it is also true that a certain amount of trust capital needs to be rebuilt before these issues can be addressed successfully.50 Given the absence of serious internal cleavages, East Timor may be a case where early resolution of such basic issues would on balance have been desirable. Procurement rigidities The application of standard Bank and ADB procurement rules for goods and works was a source of some delays and complications in project management.51 The procedures for bidding, prior review, etc., were demanding and hardly suited to the emergency requirements of such a small country with a virtually non- existent domestic market. (The only major concession to flexibility was the opening to use procedures "acceptable to the Bank" for community participation and local activities.) Partly as a consequence, larger reconstruction projects were not implemented as rapidly as in the earlier post-conflict interventions. Concerning consulting services, however, it does not appear that the application of the Bank's guidelines hampered consultants' choice or restricted flexibility unduly. Better results could have been achieved by more pre-qualification, faster deployment of procurement staff to the field, better use of technology, and simplified project documentation and processing. However, in the formal institutional vacuum of East Timor, it is also possible that greater flexibility in procurement might have generated greater integrity risks. On balance, the right approach probably would have been to hire a procurement firm to handle all procurement under the TFET--as very little local procurement capacity could have been built anyway owing to the expatriate tilt of UNTAET administration. East Timor: A Chronology Aug. 1978: Indonesia annexes the Portuguese territory of East Timor. The annexation is not recognized by the UN. 1978-1999: Sporadic resistance to Indonesian occupation by the Timorese National Liberation Front (FRETILIN); some Indonesian migration into East Timor from West Timor and from Java itself. 1998: Soeharto government falls in Indonesia. May 1999: New York Agreement between Indonesia and Portugal, calling for a referendum in East Timor on autonomy or independence. Aug. 30: The referendum votes for independence by a large margin; organized militia violence erupts, and extensive human and physical destruction ensues. Sept.: UN Security Council authorizes multinational force to restore order; Friends of East Timor Donor Meeting in Washington. 50The reader familiar with the Middle East will recognize here a variant of the same dilemma which was and remains central to the Palestinian predicament. 51The difference in rules did not in itself cause confusion, as ADB activities were clearly limited to physical infrastructure 25 Oct: Indonesian Assembly voids the 1978 annexation, thereby returning East Timor to the status quo ante; UN Security Council authorizes creation of a UN Transitional Administration for East Timor (UNTAET), to exercise all authority in the territory, including a peacekeeping contingent; order is restored. Oct.-Nov: Joint assessment mission to formulate reconstruction priorities and arrive at understandings on financing arrangements. Dec. 1999: Tokyo donor meeting mobilizes a total of almost $500 million in reconstruction and humanitarian aid; Trust Fund for East Timor (TFET) approved by Bank Board, with $10 million from Bank surplus. Feb. 2000: Bank Board approves expansion of TFET into an MDTF, administered by the Bank but with major role for ADB, to finance sectoral investments. Concurrent Consolidated Fund for East Timor (CFET) is established by the UN to finance recurrent costs. 2000-2002: TFET receives more than had been pledged, and commits about $160 million. May 20, 2002: Timor Leste becomes independent. AFGHANISTAN52 The Authorizing Environment The Context After nine years of war, the Soviet Union withdrew from Afghanistan in 1988. Violence and instability continued, however, as the mujaheddin (freedom-fighter) resistance groups started to fight among themselves; a variety of warlords emerged, and physical damage from the conflict was even worse during 1988-94--especially in Kabul--than during the war against the Soviet occupation. In 1994, the Taliban established effective if grotesquely oppressive control over most of the country, with the exception of the northern territories. In 1996, the Al-Qaeda organization established itself in Afghanistan and progressively acquired substantial influence. On October 6, 2002, less than a month after the destruction of the World Trade Center, the war against the Taliban-Al Qaeda regime began, by Northern Alliance forces supported by a US air campaign and Special Forces on the ground. The Taliban were forced out of Kabul six weeks later and lost control over the remaining southern region by yearend. The current government came to power in January 2002--inheriting a country in political and physical shambles. Afghanistan--large, mountainous, tradition-bound--has historically resisted both external domination and central government attempts at superceding local customs with uniform rules. Indeed, these have often been two sides of the same coin, as external invasion has tended to be accompanied by efforts at imposing the invader's norms over the entire territory. The Soviet occupation was resisted as much for the attempt to modernize Afghan society as for its foreign-ness. Weakness of the center of government and tension between the center and the periphery have been and remain the twin constants of Afghan political architecture. The balancing and unifying elements have been a vigorous entrepreneurial and commercial tradition-- bridging the various parts of the country--and the large stock of social capital at local level. Concerning the former, the country's location at the historical crossroads of trans-Asian trade has given Afghans an openness to free market exchange and business savvy second to none in the developing world. Concerning the latter, the popular notion that Afghanistan in 2001 had become an institutional desert is a misconception. The Taliban had indeed destroyed formal institutions, especially at the center of 52 This section draws heavily from direct experience and from communications with other key Bank staff listed in the Annex and other persons. Among other sources, especially useful was the IMF report on Afghanistan: Strengthening and Revitalizing Fiscal Management, of February 2002 26 government, but the social structure and informal governance at village level was left untouched, remained almost as effective as it had been historically, and is the strongest potential asset of Afghan reconstruction and development. Against this background, the challenge of post-conflict reconstruction in Afghanistan is to reconstitute a functionally limited but effective state; decentralize without fragmenting; and reduce economic distance by re-establishing security and restoring transportation and communication links. It is much too early to judge whether this challenge will in the end be addressed adequately. As of May 2003, however, it appears that the authority of the reconstituted government has been limited geographically rather than functionally; centrifugal tendencies have been allowed to become stronger; ethnic differences (whose role in Afghanistan is typically exaggerated) have been accentuated; and the still pervasive lack of basic security outside the main cities has kept economic distance high, hampering internal trade as well as reconstruction efforts. The Mandate for International Aid and Key Milestones In most other post-conflict situations, a political settlement preceded the preparations for reconstruction assistance, although in many cases (e.g., Bosnia and East Timor) preliminary discussions and fact gathering began before a formal political settlement was reached. In Afghanistan, the sudden collapse of the Taliban led to a deliberate two-track process, on political transition and reconstruction assistance, respectively. Partly by luck but mostly by effort, the two tracks were admirably well timed and no discontinuity emerged. Negotiations on arrangements for the political transition had started in late November 2001 in Bonn among the various anti-Taliban Afghan factions, and an agreement was reached on December 5. The negotiations, organized and led by the UN High Representative for Afghanistan Lakhdar Brahimi, involved the Tajik-dominated Northern Alliance; elements of the Pashtun majority ethnic group; loyalists to former King Zahir Shah (also mainly Pashtun), and remnants of the last pre-Taliban mujaheddin government. The Bonn Agreement called for the establishment of an Interim Authority, chaired by Hamid Karzai and with participation of key figures from the Northern Alliance as well as a few respected Afghan personalities. Six months later, an emergency loya jirga (grand assembly) was to convene to elect a Transitional Authority for a two-year term. During that period, a new constitution would be drafted, along the lines of the pre-Soviet constitution but on a republican rather than monarchical basis. Following ratification of the draft constitution by a second loya jirga, elections would then be held in 2004 to elect a regular government under the provisions of the new constitution, thus completing the political transition. The Afghanistan Interim Authority (AIA) was established in January 2002, and the first emergency loya jirga was held on schedule in June, selecting the new transitional government with Karzai as President. On the parallel track of preparations for reconstruction, also at the end of November 2001, a major conference was held in Islamabad to prepare for the reconstruction effort. A needs assessment mission took place in phases during December (while hostilities were still continuing in parts of Afghanistan) and in January 2002, with participation by UNDP, ADB and the Islamic Development Bank (IsDB) and under the overall coordination and facilitation of the Bank; an interim donor meeting was held in Brussels in late December; and the full needs assessment was presented in late January to a donor meeting in Tokyo. Donor Coordination, Reconstruction Funding and Aid Management Donor Coordination The Islamabad conference in November 2001, organized jointly by the Bank, UNDP and ADB, included all major prospective development partners and set the tone for an unprecedented degree of cooperation among multilateral development institutions. (The IsDB was also invited to become a formal partner in 27 the process, albeit with a lower profile than the other agencies.) The needs assessment was complicated by the difficult logistics of Kabul right after the end of the hostilities, and by the absence of local partners within Afghanistan. However, the Office of the UN High Representative was very active and most cooperative, and efforts were made by the mission to identify knowledgeable and respected Afghan interlocutors, both internal and through the diaspora. Tolerably good integration of Afghan viewpoints was obtained. The needs assessment and reconstruction program were endorsed by the new Interim Authority as soon as it was established and before they was presented to the Tokyo donor meeting. A joint Bank-IMF mission then took place at the end of January 2002 to discuss the first budget and the institutional building blocks of macroeconomic management. The final event of the immediate post- conflict period was an April meeting of the bilateral donors in Kabul, to review the evolving situation and mobilize contributions for the Afghanistan Reconstruction Trust Fund. After the establishment of field offices in Kabul in mid-2002, the usual group of local donor representatives began meeting regularly to iron out coordination issues and relate to the government on a day-to-day basis. The Afghanistan Reconstruction Trust Fund (ARTF)53 Genesis and evolution Just prior to the Islamabad conference, the Bank Board had discussed a proposed Bank assistance strategy to Afghanistan, including the concept of an umbrella MDTF to begin operating beyond the immediate aftermath of the war. The immediate needs of the new Afghan Interim Authority would be met by a UNDP-administered Emergency Trust Fund. Fully consistent with the political transition timetable set out at Bonn, the UNDP TF was to expire after six months, i.e., with the convening of the emergency loya jirga to establish the two-year transitional government. The Brussels and Tokyo meetings endorsed the notion of an MDTF to succeed the UNDP fund (albeit with differences of opinion regarding its coverage). The Bank-administered ARTF was approved by the Board on March 27; the Grant Agreement was signed in May; and the Fund began operating in late May 2002--pretty much along the lines of the Holst Fund set-up in 1994 for assistance to WB, but with broader coverage, as described below. The ARTF is intended to provide support only for the initial years, during which the Government would establish the capacity to mobilize revenue to cover its recurrent budget. Meshing with the political timetable, a first period of transition until the permanent government is in place in 2004 is expected to be followed by a similar period of consolidation. The ARTF closing date is therefore two years after the election of a permanent government or end-June 2006, whichever comes first (with the possibility for an extension following a thorough review of the fund operations envisaged for December 2005). By that time, it is envisaged that domestic revenues should have recovered sufficiently for the Government to finance most or all of its recurrent expenditures. Developments through May 2003 make it highly doubtful that the formal and informal (warlords) provincial authorities will transfer sufficient revenue to the center to meet this timetable. Governance and implementing agencies The ARTF governance arrangements are similar to those established in previous cases but include an intermediate tier owing to the partnership with other multilateral agencies. Between the Bank as Administrator and the group of donors contributing to the ARTF, there is an ARTF Management Committee (MC), composed of the Bank, ADB, UNDP and IsDB. The MC meets monthly to review funding proposals, after they have been reviewed and endorsed by the Government. A Donor Committee--constituted of donors who have contributed at least $5 million per year plus two seats for 53 See the Annexes for the Board paper, ARTF agreement and sample donor agreement, and the Terms of Reference of the Monitoring Agent. 28 representatives of other contributing donors--meets quarterly to review overall ARTF performance, based on reports provided by the MC. Resources and uses As of end-April 2003, ARTF resources totaled $435 million, contributed by 23 donors. The coverage is as broad as can reasonably be expected. As the Holst Fund, the ARTF includes expenditures for non- military recurrent costs of the government--wages, pensions, operations and maintenance, and debt service. Unlike earlier cases, however, because of the special importance of security in Afghanistan, certain police-related expenditures are also financed. Police-related expenditures are subject to a convoluted arrangement by which the Government submits the request to UNDP, which is responsible for the appraisal and supervision of the expenditures, and UNDP then recommends that the MC approve the use of ARTF funds for these expenditures. It would be much simpler to achieve the desired arm's length distance by direct financing from UNDP, or perhaps by an annual line-item allocation from the trust fund, to be administered by UNDP with no further involvement of trust fund managers, or simply by "twinning" the police with a bilateral donor.54 The fund can finance investments as well as non-project TA--covering the services of expatriate experts and those of diaspora returnees (especially important in Afghanistan).55 Finally, the essential link to the reconstruction program and the national budget is strengthened by the provision that eligibility of expenditures requires also prior clearance by the Government and their conformity with the government budget, as agreed by the MC. Administration, procurement and disbursement56 As standard in an MDTF, the Administrator of the ARTF is the World Bank (in this case subject to guidance from the MC, as mentioned earlier). The normal Bank fee of 2% for administering trust funds was reduced in this case to 1.0-1.5% (on a sliding scale depending on ARTF size.) Procurement procedures, while generally consistent with Bank criteria, are comparatively flexible with respect to the use of limited international bidding, and have a relatively high threshold for international shopping. Consultants' procurement is in accordance with Bank guidelines but here too there is some flexibility, in the form of greater scope for sole-source contracting. The Government's procurement consultants, Crown Agents, consider that the added flexibility in procurement is commensurate with the difficult situation on the ground and the uncertainties in Afghanistan. Only experience will tell whether the procurement procedures in the ARTF will prove adequate to the circumstances (as in WBG) or a constraint to effective trust fund operations (as in East Timor). In any event, as of 2003 the operative constraint to procurement and program implementation in Afghanistan lies not in rigidity of procedures, but in the pervasive lack of security, which makes it difficult if not impossible for potential bidders to estimate their costs--including life insurance premia for their staff. Concerning disbursements, the risks associated with financing recurrent costs were addressed by hiring, through competitive bidding, an international firm as monitoring Agent. The Agent, with an office in Kabul, screens withdrawal applications to verify the eligibility of expenditure, and its correctness through a sampling of 25% of transactions. The withdrawal applications verified by the Agent are then submitted to LOA for processing and payment. The fee for the Agent services varies with the size of the fund, but is between 1-2%. (Together with the Bank fee, total administrative costs of the ARTF are thus between 2- 3.5%.) Reflecting Afghan realities, while the existence of governmental institutions makes it unnecessary 54It should be noted that using the ARTF as a conduit for police expenditure was done at the urging of several major donors. Following the issue of a favorable legal opinion, the Board approved this unprecedented feature. 55Large investment projects are normally financed directly by the donor(s) concerned, and the ARTF does not finance a variety of other post-conflict activities--notably, demining, resettlement of refugees or IDPs, emergency security, disarmament and demobilization, military and paramilitary activities. 56See ARTF grant agreement and sample donor agreement, available on request from the CPR Unit. 29 to rely on the Agent to monitor financial management, the ARTF Agent is expected to be much more involved in capacity building than was the case in previous post-conflict situations.57 Aid Management and Implementation Arrangements Although in serious disrepair, Afghanistan still had in 2001 the shell of central government institutions, broadly adequate formal procedures, and a small number of dedicated staff who had somehow managed to keep doing their job throughout exceptionally difficult times.58 Therefore, of the three potential purposes to be served by an aid management agency in a post-conflict situation--aid management, project implementation, and proto-government--only the first two were relevant. The "sunset dilemma" of assuring an effective transition without pre-empting development of regular government institutions does not apply in Afghanistan in the same form as it had in WBG or East Timor. The Afghanistan Aid Coordination Authority (AACA) was formally created by decree in March 2002 to regulate and coordinate the flow of assistance, and implement some projects directly while capacity is re- built in the line ministries. The Bank provided advice and some assistance for the creation of the AACA, but the receptivity to the advice was much weaker and the dialogue much less systematic and collegial than in the case of Palestine. Although there is a general expectation that the AACA will not be needed after 2004 (when the political transition is supposed to be completed), the situation is fluid, the agency mandate in practice is not yet entirely clear, and its future is murky. It is too early to hazard any judgment on AACA efficiency--let alone its effectiveness. So far, the agency has been functioning in comparatively ad hoc fashion. AACA was originally set up as a separate entity, which functioned in competition with the Ministry of Finance and interacted directly with the line ministries. After selection of the transition government in June 2002, the head of AACA became also the Minister of Finance. Under normal circumstances, this could be a healthy development, as the aid management function is best located in the Ministry of Finance in order to assure integration of aid resources into the overall budget.59 However, since AACA also has a direct role in selecting and implementing projects, the "sunset dilemma" of managing reconstruction without short-circuiting the growth of capacity and empowerment in the competent line ministries has come back to the fore in different form. Accomplishments and Problems After barely 18 months from the end of the war, the situation in postwar Afghanistan is still fluid and evolving. Some accomplishments are clear and certain concerns may be mentioned, but other hopes and misgivings are mostly speculative and it would be inappropriate to raise them here. Accomplishments and Major Success Factors Meshing the political and reconstruction timetables There was a great deal of skepticism right after the war of the likelihood that the Bonn timetable for political transition would be met, and that it would moreover dovetail with the progress of the reconstruction program and the setting up of the funding mechanisms. In retrospect, the Bonn timetable 57See the Terms of Reference of the Monitoring Agent, available on request from the CPR Unit 58Observers have expressed a concern that some of the key people have been shunted aside by returning Afghans, fueling the insider-outsider resentment and eliminating the last vestiges of institutional memory. 59 For a discussion of this and other principles of aid management, see Ch.8 in Schiavo-Campo, 1994. 30 has worked out so far precisely as envisaged,60 and the needs assessment was completed just in time to be reviewed and endorsed by the newly-installed AIA, and then presented to the Tokyo donor meeting. Indeed, the already fast schedule of preparation in earlier post-conflict cases was compressed further to just seven weeks--from initial Bank/ADB/UNDP agreement on the conduct of the needs assessment, through the joint assessment missions, to the aid-pledging meeting. While the start of ARTF operations took another four months, the delay posed no problem in light of the essential support provided in the meantime by the UNDP Emergency Trust Fund, which ceased operations at about the time ARTF began disbursing--which in turn just preceded the holding of the first loya jirga. Such convergence of the political, economic and financing tracks is nothing short of remarkable, and reflects favorably on all major actors. Making good on the rhetoric of partnership The timeliness of preparation of the reconstruction program is especially noteworthy in light of the unusual degree of participation by other multilateral agencies. Although there were the usual problems with some individual UN agencies, cooperation with the UN itself and with UNDP has been better than on some other occasions, and with ADB about as good as in the case of East Timor. In particular, the division of labor during the joint needs assessment worked out well, evidence that an open and cooperative attitude by the Bank can help assuage the suspicions of other agencies and reduce somewhat the weight of turf considerations. Learning from the past While adapted to Afghan realities, the design of the ARTF benefited from earlier experiences, and its coverage is the most comprehensive yet achieved in post-conflict situations, with virtually all major budget categories covered except for flagship investment projects (which by definition carry the flag of the donor concerned.) ARTF problems, mentioned later, do not arise from its design or early management. In large part, the superior design of the ARTF and its "selling" to other donors, were due to the direct involvement of Bank staff who had been instrumental in earlier cases in devising the organizational and legal arrangements for MDTFs and aid management by the recipient government. Problems and Issues Insecurity More than a year after the end of the war, insecurity remains widespread. Land movement is impossible through some areas, through others only during daylight, and not risk-free anywhere in Afghanistan. This is not the place for discussing the reason for the continuing lack of security, but it is fair to say that this state of affairs was not inevitable. In any case, technical and financia l assistance cannot possibly be effective if program implementation and general economic activity are constrained by lack of physical security. Whose recurrent costs are being financed? Most of the warlords who emerged after the Soviet departure had been suppressed during the Taliban years. When the MDTF was first being discussed, several bilateral donors expressed a willingness to contribute substantial amounts, but only if the new central government had sufficient control and monitoring capacity to assure that the funds would not indirectly end up financing the resurgence of warlords in Afghanistan. Unfortunately, this did not prove to be the case; the April donor meeting in 60However, criticism was levied that the results of the June 2002 loya jirga were in effect decided behind closed doors, and presented to the assembly as a done deal--compromising the credibility of the democratic process from the start. 31 Kabul had disappointing results; and the initial contributions to the ARTF fell short of what had been envisaged. No budget gap emerged in the first months, but precisely because the limited geographic reach of central government authority kept its expenditures down. Later, when expenditure increased, additional contributions came in and the ARTF has become the essential vehicle for supporting the Government budget. Unfortunately, central government revenue has fallen short of forecast, owing to the willful failure of local authorities and warlords to remit customs and other revenue due to the government, and prospects are unfavorable. The warlords' position has been strengthened and consolidated partly from the support they receive as part of the continuing struggle against remnants of the former regime. Such support, however, has badly weakened the central government, and has made it that much less likely for a self-sufficient and stable political setup to emerge.61 Even in the optimistic scenario that the central government will soon be able to assert its authority vis-à-vis the warlords, international attention may no longer be focused on Afghanistan--and neither will the aid money. Underuse of some MDTF components As noted, in addition to recurrent costs and non-project expatriate TA, the ARTF has an investments component, and a component to finance the services of Afghans resident abroad that wish to return and contribute to reconstruction. After initial delays, the investment component began to be used in late 2002, and as of April 2003 comprises eight operations for total commitments of $55 million, for priority technical assistance and emergency employment generation activities. There remains a constraint, however, from the inattention to bundling small investments into coherent expenditure programs, with systematic criteria for evaluation and selection. The responsibility to do so, of course, is the Government's, but the Bank should proactively encourage the evolution of appropriate criteria for the small investments that can provide the mortar for effective reconstruction. So far, the component to finance the temporary services of Afghan returnees has not been used. The main reason appears to be lack of internal consensus on key issues, including which ministry or agency would lead the activity, and how to define the compensation schedule. The latter issue (a perennial problem not just in post-conflict reconstruction but in TA in general) requires among other things that compensation for Afghan returnees should be aligned with compensation schedules in numerous other programs--UNDP, International Organization for Migration, bilateral donors. In addition, however, the component may be indirectly hampered by the insider-outsider problem, discussed next. Insider vs. outsider The existence of a large diaspora can be an important asset for reconstruction. The skills of returnees can make a major contribution to capacity building; and their capital and business experience can kick start a set of new economic activities. On the other side of the ledger is the risk of misunderstandings and resentment between insiders and outsiders. The Palestinian case was early evidence.62 The same risk is even more evident in Afghanistan, where some returnees have been out of the country for a generation and have no direct comprehension of local events during that time. Reportedly, resentment against "outsider" Afghans, derisively called "dog-washers" by the locals has been growing and has reached acute levels. Such an emerging cleavage calls for careful management of the relationship. It is not clear 61 An Afghan observer finds it amusing that some of these commanders actually believe they are "using" local warlords who possess decades of experience in turning inexperienced foreigners to their own ends. 62As noted earlier on WBG, insiders saw themselves as resisters of Israeli occupation and resented the PLO "Tunis" outsiders taking over, while the PLO hierarchy did not fully trust those who had lived under occupation. Nevertheless, in Palestine the resentment was kept down by the status of Arafat as "father of the nation" and the fact that a majority of high decision-making positions below the top leadership were and continued to be occupied by insiders. 32 that the Afghan returnees now in a position of power have done so in sufficiently sensitive manner; it is clear that the risk of ensuing non-cooperation or even of violent reaction has the potential to seriously disrupt further the reconstruction of Afghanistan. Afghanistan: A Chronology 1979: Soviet occupation of Afghanistan begins. 1979-88: War and resistance; Soviet troops leave in 1988. 1988-94: War among mujaheddin factions and rise of the warlords. 1994: The Taliban ("students") gain power and establish control through most of the country. 1996: Al-Qaeda takes up residence in Afghanistan, and slowly emerges as dominant influence on the Taliban regime. Sept. 11, 2001: Attacks on New York and Washington. Oct. 6: War is launched by American and "Northern Alliance" forces. Mid-Nov.: Taliban leave Kabul. End-Nov.: Negotiations on transitional political arrangements begin in Bonn under auspices of UN High Representative; Conference on Afghanistan Reconstruction takes place in Islamabad. Dec. 5: Bonn Agreement on political transition. Dec.-Jan.: Joint Bank-ADB-UNDP reconstruction needs' assessment missions; UNDP Emergency Trust Fund established (through June 2002). Jan. 2002: Afghanistan Interim Authority established. Late Jan.: Tokyo donor meeting reviews needs assessment and pledges aid. April: First donor meeting in Kabul. May: Afghanistan Reconstruction Trust Fund begins operations. June: First loya jirga (grand assembly) convenes, selects transitional government for the following two years. (Second loya jirga is scheduled for mid 2004 to ratify new constitution, following which elections for a permanent government are to be held.) 33 III. ESTABLISHING A MULTI-DONOR TRUST FUND AND AID MANAGEMENT IN A POST-CONFLICT SITUATION The documents most relevant to the creation of an MDTF are available on request from the CPR Unit. The most recent documents, on the ARTF, build on the earlier experiences and are therefore the most directly relevant.63 However, persons concerned are strongly advised to copy nothing, adapt everything. Disbursement and accounting procedures are no different for post-conflict trust funds from those applicable to Bank-administered trust funds in general (except that if they are financed solely out of Bank income naturally no administration fee is charged).64 Trust fund disbursement and accounting procedures can be accessed through the Bank intranet--see the References for detailed web address. This section attempts to gather the main lessons of experience specific to the structure and governance of MDTFs, and suggests a sequence of steps for the creation of MDTFs in post-conflict cases. Some repetition of material included in the case studies and in the next chapter on conclusions and recommendation is necessary for the sake of a self-contained section. Structural and Design Issues Strategic criteria and the fundamental importance of a unified budget An MDTF performs both a trustee function and an executive function. The trustee function is essential, but raises no major issues, since the legal, accounting, disbursement and reporting aspects are now well established and clearly defined. The issues revolve around the exercise of the executive function in MDTFs. The key strategic criteria for the effective exercise of the executive function are: · an agreed program of reconstruction activities; · a workable interface with the local aid management agency; · ownership by potential donors; · fullest possible transparency; and · openings for citizens' "voice". A reasonably comprehensive government budget, realistic and public, is a necessary and (almost) sufficient answer to meet all of these criteria, notwithstanding its apparent narrow "technical" nature. It is through a budget that a coherent program of economic and social activities should be reflected. It is through a budget that the local aid management agency and the MDTF can interact. It is through a link with the budget that donors to an MDTF can see better prospects for implementation of their priority activities. It is through a budget that the allocation of resources can be made clear to all parties. And it is through a sound budget formulation process that the practice of public hearings, open debate, and habits of compromise can best be introduced. Transparency and participation, always important, are vital in post-conflict situations, where every shadow is seen as a threat and every closed door is a presumed conspiracy.65 Provide incentives to the donors: The use of fungibility If no donor contributes, there is no MDTF. This tautology underlines the necessity to focus on the incentives for individual donors to join an MDTF, and this, in turn, calls for understanding their 63 Board papers and trust fund agreements relating to earlier MDTFs are available through the Bank website. 64There are two deviations from standard Bank practice. DDR MDTFs have been given a waiver permitting cash payments to individuals (ex-combatants). Also, OPS has recently approved TFs financing of 100% of local consultant and staff costs (instead of the standard percentage), provided that the contributing donors agree. 65 Recall that we are discussing nation-building, macro-type MDTFs. For project-centered TFs, or for MDTFs limited to financing DDR activities, the requirement of a comprehensive government budget is less stringent. If the funding is substantial, however, there is no reason not to use the suasion or leverage it provides to move the recipient government in the direction of improved overall budgeting and fiscal transparency. 34 possibilities and constraints. As noted, donors need a coherent reconstruction program and genuine consultation. They also need, however, an MDTF designed to make it possible to claim that their assistance goes for a purpose of which their constituencies can approve--while precluding strict earmarking, which would negate the very purpose of an MDTF by making it impossible to scrutinize the comparative worth of different programs, and in effect transferring authority over expenditure programming from the recipient government to the donors. The fungibility of money provides a way out. The key consideration is that the degree of fungibility is directly related to the level of aggregation. At one extreme, allowing donors to specify their preferred use of the funds within a highly disaggregated budgetary classification would be equivalent to strict earmarking. At the other extreme, indicating simply that the funds will finance the reconstruction program as a whole would not provide an incentive to donors to put their money into a common pool. Somewhere in the middle, there is no constraint on the effectiveness of an MDTF if donors are allowed to indicate their preference for broad categories of expenditure (such as primary education, basic health, or transport infrastructure), and receive regular reports on the broad breakdown of actual expenditures. Donor preferences would be explicitly acknowledged at the time of their contribution, but all contributions would be commingled into a common pool and the MDTF Administrator would retain the authority to move funds between categories of expenditure within specified limits. If the expenditure categories are broadly defined, each contributing donor is highly likely to find that an amount equal to at least his contribution had been in fact spent for his preferred category. Undoubtedly, a variety of other ways can raise the incentives for donors to accept and contribute to an MDTF. It could be useful to organize a Bank workshop to brainstorm on how to get more traction in the future to get donors to buy into a single MDTF. Coverage MDTF coverage cannot be as broad as that of the national budget. First, while all investment expenditure should be included in the budget, no single MDTF can or should cover very large self-standing investment projects. Donors are not likely to accept to fold their flagship project into a single trust fund, nor would this be necessary for good implementation--since large projects invariably embody specific implementation arrangements.66 Very large projects, however, may themselves be financed through a separate dedicated trust fund, either because the expenditure is too large for any single donor, or to spread risks. An umbrella MDTF is normally unsuitable to also cover financing for strictly humanitarian aid, or security-related programs such as de-mining. (Again, separate dedicated trust funds can be useful mechanisms for those activities.) All other expenditure categories are potential candidates for financing through an umbrella MDTF. The advantage of a comprehensive fund at the outset, as in Afghanistan, is the possibility of a closer linkage with the recipient country's budget, and thus capacity to foster better integration between recurrent costs and (some) investment expenditure. The advantage of a selective fund, focusing on whichever program presents clear and unmet financing needs, is the capacity to grow and change as the situation develops--as the Holst Fund--provided it is designed in accordion fashion to permit such expansion. Financing recurrent costs through an MDTF, however, permits reaching crucial understandings with the government on administrative machinery, employment and wages. The choice between comprehensiveness and selectivity is of necessity country- and situation-specific. What is to be avoided in any event is an excessively a sharp separation between an investment trust fund and a recurrent cost trust fund (as in East Timor). Such separation tends to produce budgetary 66 In some cases and with notable exceptions, it may still be possible, however, to try and harmonize procurement procedures, in order to relieve the recipient government of the burden of dealing with different donor procedures, as well as help build capacity in procurement. In any event, donors should be strongly dissuaded from destructive competition for local talent, distorting labor markets and weakening the very capacity required for effective assistance 35 fragmentation on the recipient side and thus policy incoherence. Faced with this prospect, it could be preferable to forgo an umbrella trust fund and follow a different route--whether a sector-based co- financing approach (as in Bosnia), or simply, to obtain explicit agreement of all major donors with a common strategy and the policies and spending priorities needed to implement it. (The government budget would in that case become that much more important as a focus for aid coordination.) On balance, there is a strong practical presumption in favor of one umbrella MDTF, which can encompass, at a maximum, the following categories of civilian expenditure: · start-up costs (fully transitional); · salaries and pensions (mostly transitional); · operations and maintenance (partly transitional); · investments (except large flagship projects); · non-project TA; and · debt service. A special coverage issue: Is paying for the police and prisons worth the risk? No. The issue of whether an MDTF should finance expenditure for police and prisons invariably comes up at the start of the discussions on reconstruction funding. On the one hand, restoration and maintenance of internal law and order are basic civilian governmental functions, and there is no conceptual justification for excluding the requisite expenditure from an MDTF that otherwise funds recurrent costs. On the other hand, the fog of post-conflict and the uncertainty concerning the prospective behavior of the new government, make funding of police and prisons far riskier from a reputational point of view than financing other civilian government activities.67 Although a recent Bank legal opinion makes police- related expenditures eligible for inclusion in a Bank-administered trust fund, it appears highly advisable on balance not to do so. The counterargument is that through such administration the Bank could exert a positive influence on the human rights climate in the recipient country. But there are better alternatives: assistance for police and prisons can be mobilized from a bilateral donor with both interest and expertise in law and order, or can be administered through UNDP, and the Bank can insist on human rights considerations as part of its overall dialogue with the government. (However, expenditure on the legal system and the judiciary should certainly be covered under an MDTF, as the judiciary is essential for good governance and assistance to it does not carry the same type of risks as aid to police and prisons.) Organizational and Procedural Issues Is there a correlation between trust fund effectiveness and governance simplicity? No. In East Timor, the relatively simple arrangement of semi-annual TFET Donor Council meetings, complemented by the customary Bank-chaired donor coordination committee in the capital, went hand-in- hand with unwieldy aid arrangements fragmented among other vehicles. In Afghanistan, by contrast, despite their apparent complexity, governance arrangements have so far worked well. None of the strategic criteria for an MDTF listed earlier is directly related to the trust fund governance arrangements. When these criteria are met, it matters little whether there are one, two, or three concentric circles of governance and oversight; conversely, absent one or another of the essential criteria, even the simplest MDTF governance arrangement will degrade into second-guessing and attempts at micromanagement. 67With World Bank activities under close scrutiny from various quarters --some hostile--it could be disastrous if some assistance from a Bank-administered trust fund were somehow diverted by the recipient government to financing torture, inhuman punishments, or other human rights violations. 36 Speed vs. safety: The issue of advance authorization Time is of the essence for aid interventions in post-conflict situations. The cases examined have shown the importance of early planning, and it is equally desirable that the funding mechanism becomes operational quickly. However, there is a reputational and fiduciary risk to the Bank as trustee, especially if a recurrent cost component is included in the MDTF. Two issues are at stake. First, there can be no compromise with the need to put in place measures to minimize corruption and leakages before the MDTF begins operations. Second, concerning the link between the timing of donor contributions and that of commitments from the MDTF, at one extreme no commitments should be made until the actual deposits from donors are in the special account. At the other extreme, the MDTF could be allowed to enter into firm commitments as soon as indicative donor pledges are received. The former position would be safest but delay MDTF operations; the latter would entail a risk for the Bank if the pledges did not materialize. An intermediate position can be adopted: permit the MDTF Administrator to enter into commitments up to a certain percentage threshold of the total of binding pledges--where binding may be defined on the basis of a simple letter from the donor agency in question. (The percentage would of course be on a rolling basis, as new binding pledges come in.)68 Now that substantial experience has been gained with MDTFs, it would be appropriate to conduct a quick fact-finding exercise to ascertain the delay between first pledge and actual deposit in previous MDTF operations--for different donors and in the aggregate. This would permit defining an empirical frequency distribution to use as the basis in future cases for setting the percentage threshold at a safe level while avoiding undue delays in TF operations.69 Non-project technical assistance Large investment projects embody their own advisory and TA components and funding. Non-project TA, for essential institutional development and capacity building, needs to be programmed and financed separately. Previous post-conflict cases provide no clear evidence of whether such non-project TA is best handled as a component of an umbrella MDTF, or by a separate MDTF, or by a single-donor fund. In any case, the key requirement is the closest possible involvement of local counterparts. An airport can conceivably be built by foreigners, but for institution building there is no substitute for local ownership. East Timor exemplifies the extreme case of expatriates exercising directly all basic administrative functions without meaningful involvement of local Timorese, and Kosovo shows the even worse predicament of experts financed by different donors contesting for influence, capturing local clients, and hiding information from one another. However, the temptation of only doing, rather than also helping to build local capacity, is ever present. The first lesson of experience is therefore that non-project TA activities should be linked with a host-government capacity-building program, and that every set of terms of reference and every contract should include an explicit capacity-building element. 68 Such a position would have the added advantage of obviating donor reluctance to sign the formal grant agreement promptly, as there would be no immediate claim on their funds. 69Note that the issue here is not to prevent an MDTF from running out of funds. If liquidity problems begin to loom the expenditure path will be slowed down to prevent such an outcome. Ex post, no liquidity problem will be visible. The issue is instead to provide predictability of resources and effective expenditure programming over a meaningful period of time - thereby facilitating good budget preparation and execution. Note also that in several MDTF cases, as for example the disarmament/demobilization/reintegration (DDR) trust funds in Africa, the constraint has not been lack of money at the right time, but weak Government capacity to formulate sound programs and implement them in accordance with Bank procedures. 37 Other lessons of experience are: · capacity is inherently relative to the tasks to be performed, and TA should take special care not to aggravate local capacity problems by introducing overly complex systems or methods unsuited to local conditions, especially in public expenditure management; · because non-project TA expenditures are prone to expensive advisory activities of limited value, an especially close watch and tight quality control are advisable; · tying TA exclusively to services of donor nationals should be discouraged; · non-project TA should not be allowed to replace whatever local capacity exists (or short-circuit its creation); · a framework of priorities is essential, but funds should not be fully allocated in advance, and a substantial cushion should be maintained for urgent TA needs as they arise. How are recurrent costs to be financed and monitored? For project aid, physical and financial monitoring methods are fairly standard and well established, including some measurement of results. The same is true to some extent of special programs, such as resettlement of returnees. For recurrent costs, however, it is difficult to verify that all expenditures have in fact gone to finance the specific agreed beneficiaries and activities, especially when some of these activities are heterogeneous. It is intrusive and ineffective to try and micromanage specific inputs. Experience suggests instead the following criteria for monitoring recurrent cost financing: · monitor the (agreed) broad composition of expenditure; · finance large discrete categories of recurrent expenditure that entail regular and predictable payments, e.g. salaries, pensions; · engage in close dialogue among the donors and with the recipient government on policy affecting those categories of expenditure; · ensure an efficient payments mechanism, even if initially subcontracted; · complement the above with spot checks and random inspections of vouchers, payroll, etc; and, most important and politically most difficult, · be prepared to halt disbursements in the event of serious and uncorrected deviation from the agreed policies and expenditure composition, and/or substantial corruption. On balance, in post-conflict situations it is advisable to hire an Agent for the MDTF to verify the eligibility and correctness of withdrawal applications, and carry out spot checks of the validity of transactions. As and when local structures improve, it should be possible to gradually rely more and more on them for monitoring of expenditures and results of activities at local level. Internal Bank Processes Hit the ground running The importance of advance planning and an early start has been underlined earlier as crucial in all post- conflict cases for both donor coordination and the Bank's capacity to interact constructively with the client. Inclusiveness vs. confusion: Resist disaster tourists While there is a clear need for an open and inclusive posture by the Bank, post-conflict situations are newsworthy and high-profile, and tend to be magnets for underemployed organizations and individuals who look for bragging rights or psychic incomebut have little or no value added to contribute. By and large, the Bank has carefully assembled its post-conflict reconstruction teams to keep to a minimum the 38 internal hangers-on, whether prospective team members or managers. (Some other donors have been less successful.) More problematic has been the challenge of dealing with peripheral official agencies or NGOs looking for the limelight. Ignoring them altogether risks shifting the transactions burden onto the recipient government, by encouraging those organizations to search for and deal directly with internal clients. Ideally, the Bank can help find useful niches for them. In any case, there is a responsibility to help the recipient government to resist disaster tourism. Improve the recording of institutional memory In almost every instance, the first step taken by Bank staff in its preparatory activities was to examine the experience of earlier cases and tap the memory of those concerned, and every program shows evidence of learning from earlier experiences. However, the documentary record was stronger in some cases than in others and, naturally, some of the key players were no longer around. OED evaluations are generally the best source of relevant information and lessons learned--second only to the personal recollections of key staff--but by definition they are not available until a later stage. While the present paper will hopefully be of some use in improving the experientia l basis for future programs, the real solution lies in more systematic recording of experiences in real time. It would be counterproductive to suggest that a Bank team navigating the maelstrom of a fast-developing post-conflict situation should subordinate its essential work to record the developments. Other solutions might be found, however. Possibilities include a semi- permanent panel to provide advice on request, or ex post interviews with key consultants and staff (similar to the after-action debriefings typical of military operations), or the inclusion in the Bank team of a regular observer from the CPR Unit.70 The Aid Management Agency (AMA) Discussing an MDTF without referring to the government's own aid management agency would be the sound of one hand clapping, for it is the interface between the two that provides the bridge from donors to government and is the source of initial ownership. However disastrous the situation on the ground may be after a prolonged conflict, receivership is never an option. Thus, among the very first priorities will be to assist the government to set up its own aid management agency (AMA) to deal with reconstruction, in parallel with donor coordination and with the construction of funding arrangements. The fundamental strategic challenge of the post-conflict reconstruction process is the reconciliation between immediate urgencies and sustainability over the longer term. In some cases, government institutions are sufficiently functional that the AMA can be limited to regulating the flow of "traffic" and provide selected implementation services upon request of government ministries. In most cases, however, the government is extremely weak during the immediate post-conflict period. In these cases, the agency must also help formulate the reconstruction program, implement directly a number of activities, and be accountable for expenditures and, at least in some respects, for results. For this, adequate AMA capacity is a must. However, in the face of a virtual institutional vacuum and from predictable bureaucratic impulses, the agency has a natural tendency to monopolize decision-making, act prescriptively, and-- intentionally or not--hamper the rise of regular government organs. Paradoxically, this tendency will be stronger the more competent and effective is the AMA. This was well understood at the time of the establishment of PECDAR in WBG, and a sunset clause was written into the PECDAR Bylaws. As explained earlier, PECDAR got off the ground rapidly to become an effective counterpart for the donors, implementer of a complex program, and incubator of proto- government. The core economic management functions of PECDAR--economic policy, investment programming, finance--did in fact disappear as they were envisaged to, as soon as the regular ministries of planning and finance were established. However, the peculiar governance style in the Palestinian 70 It would be essential in the latter case to assure that the observer's role is merely to take note of key decisions and their context and in no way serve as critic or reviewer. 39 Authority produced extremely weak line ministries (with some exceptions), and most donors preferred to continue implementing projects with and through PECDAR. In general, the circularity of the problem-- strong AMA/weak ministries/reliance on AMA/continued weak ministries--precludes identifying its origin with certainty. In the case of WBG, the problem arose more from the way in which many line ministries were allowed to expand than from deliberate obstacles on the part of PECDAR.71 The "sunset dilemma" is common in post-conflict reconstruction. An AMA is created because the formal governance structures do not have the capacity to implement reconstruction, and because donors require transparent and reasonably corruption-free financial management. Over time, as the regular governmental institutions grow, competition emerges between the governmental structures and the parallel AMA.72 Instead of a smooth handover, the parallel tracks tend to persist, partly because the AMA has built up greater implementation capacity and partly because accountability and financial transparency remain a must for donors. And so, the AMA stays active longer than envisaged, competing with regular government ministries for resources and authority. Thus, government and donors should agree from the outset on an exit strategy for the AMA; link that strategy to appropriate conditionality vis-à-vis the emerging government structures; and at the same time assist proactively in building institutional capacity in the regular organs of government. Other lessons of experience with aid management agencies in post-conflict cases are: · putting needs assessment, program formulation, aid management and project implementation under the same roof can work well during the initial period; · separation between the policy function (by the Board) and the AMA management is essential, and should be closely monitored; · the Board should be representative of the key groups and local actors (in multi-ethnic countries, the AMA itself can serve as a laboratory of ethnic cooperation); · the AMA should operate by clear, public and transparent procedures, and assure the closest possible linkages with the emerging line ministries; · whether or not the AMA is given direct implementation responsibilities, building implementation capacity in the line ministries should be a priority, but only in the context of an agreed vision of government machinery and administrative expansion. Proliferation of weak ministries is not a sound alternative to a technocratic monopoly of decision-making. A Trilateral Division of Responsibilities Where there is an umbrella MDTF covering the bulk of civilian expenditures, and a Monitoring Agent is contracted (which would normally the case if recurrent costs are being financed), the responsibilities of the trust fund administrator (the Bank), the government (through the AMA), and the MDTF Agent can be delineated as follows. The Administrator: · enters into donor agreements (for contributions into the MDTF) and grant agreements (for expenditures out of the MDTF); · defines, in close consultation with the AMA and with the donors contributing to the trust fund, the major components of the fund, and the broad categories of recurrent expenditure eligible for funding; · identifies a suitably qualified contractor (the MDTF Agent) and contract for its services. The Administrator is not be liable for the activities of the Agent, but evaluates its performance and takes corrective action if necessary; 71The opposite is probably true in the case of Lebanon's Council for Development and Reconstruction (CDR), which actively worked to retain authority over activities normally exercised by regular organs of government. 72A version of this issue emerged in Kosovo, between the official agencies and the informal parallel structures. 40 · provides for the establishment of a Special Account for the MDTF (or for each of the major components), makes the initial deposits from the MDTF, and replenishes the Account(s) monthly, upon request by the AMA; · defines procedures and documentary requirements for withdrawals from the MDTF, and the format and coverage of statements of expenses, reports, and audits; · ensures independent external audit of the Special Account(s); · evaluates MDTF activities (usually quarterly) and reports accordingly to the participating donors; and · organizes meetings of contributing donors (usually semiannually), the AMA and other concerned parties, to review progress in disbursements and activities, exchange views, and resolve major issues and in other ways acts as "secretariat" to the group of contributing donors. The AMA: · is the principal government interlocutor with the MDTF and with the donor community on aid issues; · drafts policy statements, analyzes economic developments and prospects, and prepares estimates of the broad economic and financial aggregates, in consultation with the Administrator; · assists the government in formulating a program of activities, in consultation with all relevant stakeholders, including in particular the line ministries, and proposes the program to the Administrator for funding; · assists the government in formulating a budget consistent with the program; and · either implements or assists line ministries in the implementation of program activities, depending on the extent of administrative capacity in the line ministries. The Agent: · screens withdrawal applications for eligible expenditures submitted by the AMA, for transmittal to the Administrator (and, following his clearance, to LOA for processing); · provides to the Administrator a monthly statement of expenditures, by category of expenditure but not differentiated by donor; · assures the application of procurement procedures satisfactory to the Administrator (as mentioned earlier, procedures should be clear, simple, transparent, and flexible, especially in the case of smaller expenditures to meet urgent needs); · provides capacity-building assistance to the AMA (and the sectoral PIUs, if they are set up) on disbursements, payments, budgeting, accounting, and other technical tasks required for their eventual autonomous functioning. 41 Stages Toward Establishing an MDTF in Post-Conflict Situations The following stages of Bank involvement emerge from the earlier experience. Some are obvious and none are novel or sophisticated, nor are they necessarily prescriptive. The sequence is offered here only as a useful checklist of possible steps in future cases. In-conflict watching briefs Keeping a watching brief during the conflict itself is now Bank policy.73 This generally involves, among other things: · Listing the major agencies active on the ground (normally, international NGOs and/or UN agencies), and their primary activities; · establishing frequent and systematic contacts with these agencies, including occasional Bank staff field visits, if the security situation permits, or, if it does not, visits to their headquarters combined with "listening posts" in neighboring countries; · obtaining information from foreign ministries of concerned countries; and · tracking the conflict and, to the extent possible, its geographic articulation and the damage to the economy, physical and human capital. Pre-post-conflict early planning · Form an internal Bank working group/shadow country team; · Identify a "sponsor" for funding and facilitating the early planning by the Bank staff;74 · Carry out a preliminary assessment of economic prospects, problems and possibilities, and sketch out a menu of strategic options; · Interact affirmatively with the major potential donor(s), including the UN, EC, and the regional development bank(s) concerned,75 share all relevant information, and explore attitudes toward post- conflict cooperation, including through a joint trust fund; · Ascertain the position of the key political players, on the outline of a possible political settlement, their preferred coordination and funding mechanisms, and requirements for high-level governance of a possible multi-donor trust fund; and · Participate in the negotiations toward a political settlement, if invited.76 Immediate post-conflict period · Mount a full mission, as soon as possible after the settlement, jointly with the IMF and key donors, with maximum feasible interaction with the local authorities and liaison with NGOs and UN agencies active on the ground, to: (i) assess immediate and longer-term reconstruction needs (using the headstart provided by the preliminary assessment); (ii) advise and agree on the recipient entity's broad organizational arrangements for aid management and reconstruction implementation; (iii) agree on the parameters for basic economic policy positions; and (iv) agree on a rudimentary budget for the recipient government; · Simultaneously, formulate the outlines of the funding arrangements, and specifically of an MDTF, including particularly its governance, coverage, and understandings on lead donors for different activities, consult informally other key donors, and incorporate their views/preferences into the MDTF design; 73 See Operational Policy (OP) and Bank Procedure (BP) 2.30 on Development Cooperation and Conflict. Keeping a watching brief for in-conflict situations or failed states was suggested, among others, in Schiavo-Campo, 1996. 74The Dutch ED performed this role in the case of Bosnia before Dayton. The organizers of the Madrid talks funded the extensive Bank assessment of the Palestinian territories prior to the Oslo Agreement. 75EBRD in Bosnia, ADB in East Timor and Afghanistan, and IsDB in Afghanistan 76As in the Dayton negotiations on Bosnia. Note that Bank involvement in the activities and contacts mentioned above will make it more likely that the Bank will be invited. 42 · Select a prospective task manager for the MDTF; · Hold a donor conference immediately after the assessment mission, co-chaired by the Bank and at least one of the other key donor agencies, mainly to: (i) mobilize funding commensurate with the assessed requirements; (ii) endorse the economic policy positions and proposed budget for the recipient; and (iii) obtain official backing of the MDTF and its general outlines; if such backing is obtained, · Appoint the MDTF task manager; flesh out MDTF arrangements, including measures to address fiduciary risk along with flexible procurement procedures and--if recurrent costs are to be included-- TORs for the agent contracted to handle disbursements and reporting and audit provisions; hold a rapid informal meeting of the key contributing donors to finalize the structure; · Advise the recipient on all issues relevant to setting up their aid management and implementation entity, with a view to having it in place before the MDTF is formally approved and ready to begin disbursements; · Request Board approval of the reconstruction strategy, including the MDTF; and · Most important, be sure to share liberally the "credit" for decisions and successes, without fragmenting the responsibility for leading the effort or diluting the Bank's authority and accountability as MDTF Administrator. 43 IV. GENERAL ISSUES AND SELECTED RECOMMENDATIONS The sample of post-conflict cases examined in this paper is reasonably representative of the variety of country situations and of funding arrangements. While specific solutions must be tailored to the country and circumstances, certain general issues do emerge. Among these, establishing and maintaining order and security is a fundamental prerequisite. Although this is neither a funding nor an aid management issue, it is evident that the best needs assessment, reconstruction program, ownership, staffing, financing arrangements, resource availability, and implementation capacity are worth little if program implementation and general economic activity are constrained by lack of physical security. The General Context Make sure of a clear mandate Clarity of mandate given to the Bank, flowing from a political resolution and from a genuine consensus of donors, is the first requirement for effective post-conflict financing. In almost all post-conflict cases, a political resolution sanctioned internationally had to be achieved before organized economic and reconstruction assistance could begin to flow. Prior to such a resolution, even humanitarian assistance is inevitably fragmented and turf-dominated or, at worst, captured (literally or metaphorically) by combatants for their own ends. Work for genuine partnership Genuine partnership among the key international institutions concerned is a corollary of the need for a clear mandate. Enabling each donor agency to contribute in their own way improves the cohesion and quality of programs and financing, and offers a predictable and clear interface for the local authorities. In the special complexity of post-conflict situations, these well-known precepts about donor partnerships have shown themselves to be practical requirements for MDTF effectiveness. The need for partnership should not be misinterpreted as erecting firewalls between different partner activities. The frequently used "comparative advantage" terminology is especially misleading in a relationship that should be characterized by cooperation and not by trading. In some cases, problems have arisen from administration of one particular trust fund, because of lack of systematic opportunities for dialogue with or compensating intervention from other donors. Conversely, all donors should contribute in appropriate ways to the dialogue with the government on formulating the reconstruction and aid strategy and, once such strategy is established, should cooperate in its implementation--whether or not they contribute to the MDTF that is the main financing mechanism. Don't neglect economic policy issues Even though the priority in post-conflict situations is naturally on urgent short-term relief and reconstruction needs, economic policy issues must not be neglected. When the focus on a funding mechanism was allowed to lead to complete neglect of economic policy options, initial quick wins in reconstruction were obtained at the expense of sustainable results down the line. Fortunately, experience suggests that an intrusive and time-consuming policy dialogue is not necessary to prevent neglect of economic policy issues. What can be recommended is a simpler "negative list" approach--i.e., setting boundaries and tripwire mechanisms to prevent major policy mistakes in selected areas, as for example by discouraging extensive price controls or excessive expansion of government employment. Neglect of longer-term issues could also be prevented by appointing a "designated hitter" in the team with the responsibility of staying focused on the long run. 44 Communicate proactively Communication and dissemination by the Bank and its funding partners is important to deflate unrealistic public expectations and pre-empt the criticism that inevitably follows the initial post-conflict euphoria. In most of the cases examined, the donor community found itself constantly on the defensive-- particularly concerning delays in speed of disbursements, the allocations of funds to activities benefiting different interests, and the unavoidable occasional diversion of funds. Without a solid external relations and communications strategy, early erosion of public support is likely. The best communication strategy will be of little use, however, unless buttressed by the reality of early and visible results on the ground. The utility of a tight timetable The first phase of post-conflict reconstruction work is always carried out at a frenetic pace--to stay abreast of fast-moving circumstances. In addition, the East Timor and Afghanistan cases show the impetus that can be provided by a clear and tight timetable of political and economic progress, with appropriate benchmarks.77 Much like the prospect of being hanged in the morning, the potential cost of missing a key deadline in an integrated process helps every participant to concentrate. Furthermore, cooperation by other development partners is facilitated by a tight timetable, as there is less time to dedicate to turf battles and territorial positioning. This leads to the next topic. Insist on the Script, Not the Cast: Relating to Donors Relating to bilaterals and the regional development banks The experience shows the payoff from initiating an early dialogue with the major partners, and from building a consensus on relative roles and responsibilities in post-conflict organizational arrangements-- even if it takes extra time to achieve it. The Bank should welcome a substantive role for each of the major partners, but should not compromise for the sake of inter-donor amity on the need for a common reconstruction program; one budget reflecting that program; and streamlined and efficient funding arrangements to match, all with the active and informed participation of the recipient government. Relating to the UN In many post-conflict situations, relations with the UN system are an especially complex issue, with three aspects. First, in post-conflict situations the UN naturally has responsibility for political and security matters. While, as noted, the restoration and preservation of security is a prerequisite, the political timetable is necessarily different from a reconstruction timetable that looks to fiscal sustainability and long-term development. This difference in perspective and timing is surmountable (as the positive Afghanistan experience shows), but only if it is recognized and explicitly addressed. Second, in infant states (East Timor) or crypto-states (Kosovo), where the UN also exercises all administrative authority and serves as de facto government, there arises an ambiguity between that function and the UN role as a participant in the external assistance program. It is especially necessary in such unusual situation for the Bank and the UN to arrive at very clear early understandings of each other's roles and responsibilities. Third, the UN system itself must first succeed in persuading its various agencies to cooperate with a common policy framework and funding arrangements. The UN Resident Coordinator (or Special Representative, depending on circumstances) has therefore a special role to play. 77It is important, however, to resist the temptation to overdo this, and to keep the benchmarks to the few clear and monitorable ones that are essential milestone for economic and institutional progress. 45 The Importance and Limits of Ownership: Relating to the Client When should key issues with political implications be tackled? In most post-conflict cases, important development challenges with political implications were avoided in the early stages. International experience provides no clear message as to the relative advantage of tackling hard issues earlier in the reconstruction process rather than after a period of consolidation. On the one hand, lack of resolution of basic issues inevitably interferes with the effectiveness of reconstruction efforts. On the other hand, a certain amount of trust capital needs to be rebuilt before these issues can be addressed successfully. The key variable here is the severity of internal cleavages, and the external political constraints. Thus, in WBG, the very nature of the Oslo Agreement called for leaving resolution of the hardest issues until after Israeli-Palestinian trust had been built by working together and resolving less difficult problems. In hindsight, views differ on the wisdom of this approach. In East Timor, given the absence of serious internal cleavages and external political constraints, earlier resolution of such issues (e.g., land property rights) would have been desirable. In the Balkans, neither the political settlement nor the social climate would have permitted anything other than a focus on specific practical achievements. The only general counsel that can be offered here is neither to assume that hard issues must necessarily be left until later, nor to allow early pragmatism to lead to forgetting that at some stage the underlying basic issues must be addressed. Whose ownership? The outsider-insider problem A large diaspora can be a major human and financial asset for post-conflict reconstruction, contributing capital and needed skills. However, a balance between "outsider nationals' and "insider nationals" is very important for sustainability. Such a balance will differ in different circumstances, and in any case must be found by the country's people themselves. The issue for external assistance is that external agencies, both multilateral and bilateral, have a natural bias in favor of "outsiders"--who share many of the working habits of aid officials as well as ease of communication, and often were educated in the donor countries themselves. The Bank and other donors must be mindful of this bias and find ways to both correct it and assuage possible feelings of resentment among "insider" locals. One, simple, budget The importance of a comprehensive government budget--whatever the arrangements for financing it-- has been underlined repeatedly throughout this paper, owing to its profound implications for governance and sustainability of reconstruction. This should not be interpreted as suggesting sophisticated budgetary systems. On the contrary, post-conflict budgeting should stick to basics: a simple functional classification complemented by the basic economic classification, with few line items; a cash-based budget; a streamlined procedure for allocative decisions; single-entry bookkeeping; an agile payments mechanism (even if contracted-out during an initial period); and robust but selective audit focused on financial integrity. In post-conflict situations the imperative of quick wins and rapid disbursement leaves for a later stage standard budgetary issues such as development of a medium-term expenditure framework, detailed estimation of the eventual recurrent costs of investment projects, etc. It is important, however, to provide for some interaction between investment decisions and current expenditure;78 to present capital expenditure alongside and in the same document as the recurrent budget; to ensure adequate debate and public approval; and to sow the organizational seeds for an eventual move to normal budgeting processes. 78 This interaction can be facilitated by tracking employment and salaries ­ since investment projects invariably have a salary component, and some start-up expenditures procure fixed assets. This is an additional consideration in favor of a single umbrella TF, mentioned later. 46 The limits to conditionality in high-profile cases As evident from experience in WBG, Bosnia, Kosovo and thus far in Afghanistan, the importance of restoring peace, gaining the cooperation of the parties, and fueling the process of recovery poses a severe constraint to the Bank's capacity to insist on certain policy conditions. Bank staff have been well aware of the importance for sustainable fiscal and institutional development of setting early standards-- especially in public administration and economic management--and of their enforcement. Yet, attempts to arrive at clear policy agreements and to stand by them can be frustrated by external pressures to prevent any interruption in the flow of financial support. Conditionality in "stroke-of-the-pen" reforms, e.g., exchange and trade liberalization, price decontrols, etc. can be comparatively effective, but conditionality to contain the tendencies toward proliferation of governmental structures and expansion of government personnel is weaker. Among other problems, absence of reliable data, loose supervision mechanisms and, sometimes, lack of security on the ground, hamper monitoring of the implementation of policy understandings. Nevertheless, while direct limits on recruitment, wage rates, and governmental machinery are difficult to enforce, they can be compressed "from the top" through the budget--for the financing of which reconstruction assistance is essential.79 External pressure to keep the money flowing is to be expected, but reasoned Bank firmness may have a chance to prevail--if it is grounded on a prior donor consensus on policy direction, detailed homework, and open relations with the recipient government. Watch for centralizing tendencies Centripetal tendencies by the recipient government should be carefully watched. While aid must encourage capacity building in government, post-conflict experience suggests viewing the local population as the ultimate client. The good experience in all MDTFs with community-driven projects (and, to a lesser extent, with NGOs) suggests encouraging the government to act as much as possible as a conduit, channel, intermediary, and facilitator of local initiative. Such an approach should be explicitly reflected in the structure and bylaws of the aid management agency as well. This consideration leads to the next and final point. Go Soft, Go Local: Relating to NGOs and Local Communities Post-conflict reconstruction is inherently a top-down affair. Donor decisions come from their capitals; strategy meetings take place in Washington, Paris, Tokyo or Brussels; the recipient government itself, as noted earlier, has a naturalcentralizing tendency; even the collection and dissemination of information is top down, as security and logistical problems make it difficult for the media to travel to and report from the periphery of the country. Yet, some organizations have managed to remain active even in the worst conflict cases and the most dangerous years. At the very minimum, extreme caution must be exercised lest these organizations and local structures (which are usually unknown or neglected) are inadvertently suffocated by the post-conflict reconstruction programs. For this, knowledgeable local sociologists, economists and anthropologists should be brought into the dialogue at a very early stage. Beyond protecting what exists, the Bank should aggressively encourage the recipient government (and the donors) to incorporate into the reconstruction program the contribution of local communities, and of the local and international NGOs that have been laboring in conflict areas for years. It is true that using NGOs can cause overhead costs to add up pretty quickly, and that in many cases NGO accountability (especially in financial matters) is loose. While mindful of these limits, specific provisions on NGO roles should be incorporated in the trust fund agreements. 79 It is important to complement top-down enforcement through the budget with measures to protect governance and administrative efficiency. Thus, specific provisions are needed on wage rates (to avoid the syndrome of large numbers of employees at below subsistence wages); on financial management and accountability mechanisms; and on the opening of some channels for user voice and local empowerment in service delivery. 47 NGOs and local communities possess valuable information and experience, and are often the only ones capable of implementing activities on the ground. But their potential contribution goes much beyond assisting in implementation, or even acting directly as implementing agencies. Practical ways must be found to give voice to local communities in project selection and design. Indeed, the post-conflict experience demonstrates that some of the strongest components of the reconstruction programs have been local--resting on empowerment of communities and their partnership with NGOs. Moreover, capacity building at local level is a necessary condition for the evolution of the rule of law, accountability, and transparency mechanisms that make for good governance and sustainable development. Finally, and returning to the core issue mentioned at the start of this paper, local structures can make an essential contribution in rebuilding the social capital destroyed by the conflict. __________ 48 Selected References80 Bodewig, C., Emerging from Ethnic Conflict: Challenges for Social Protection Design in Transition Countries, Social Protection Discussion Paper Series, World Bank, Washington, December 2002. (socialprotection@worldbank.org) Danish Government (Danida), The International Response to Conflict and Genocide: Lessons from the Rwanda Experience, www.um.dk/danida/evalueringsrapporter/1997_rwanda/b4 Eriksson, J., The Drive to Partnership: Aid Coordination and the World Bank, OED, World Bank. Washington, 2001. Fox, W. and C. Wallich, Fiscal Federalism in Bosnia-Herzegovina: Subsidiarity and Solidarity in a Three-Nation State, unpublished discussion draft of a chapter in a forthcoming book. McKechnie, A., Humanitarian Assistance, Reconstruction and Development in Afghanistan: A Practitioner's View, Paper presented at the International Conference on Post-Conflict Reconstruction, Hiroshima, November 2002. Norwegian Government (Ministry of Foreign Affairs), Evaluation of Norwegian Assistance to Peace, Reconciliation and Rehabilitation in Mozambique, Evaluation Report 4.97, Oslo, 1997. ______, Reconciliation Among Young People in the Balkans, Evaluation Report 7/2001, Oslo, 2001. Rohland, K. and S. Cliffe, The East Timor Reconstruction Program: Successes, Problems and Tradeoffs, CPR Working Paper No.2, World Bank, November 2002 Schiavo-Campo, S., Institutional Change and the Public Sector in Transition Economies, World Bank Discussion Paper No.241, Washington, 1994. ______, Government Employment and Pay, Policy Research Working Paper No.1771, World Bank, Washington, May 1997; senior author, with A. Mukherjee and G. de Tommaso. ______, Lessons for Afghanistan of the Experience in Post-Conflict Institutional Arrangements, Paper presented at the Conference on Afghanistan Reconstruction, Islamabad, November 2001. ______, and D. Tommasi, Managing Government Expenditure, Asian Development Bank, 1999. Sierra Leone Government, National Disarmament, Demobilization and Reintegration Programme, Ministry of Finance, Doc. no. 225612 of 11 October 1999. Shihata, I, H. Abushakra and H-J Gruss, Legal Aspects of the World Bank's Assistance to the West Bank and Gaza Strip, The Palestine Yearbook of International Law, vol II, 1992/94 Sorbo, G. et al, Norwegian Assistance to Countries in Conflict Lessons of Experience from Guatemala Mali, Mozambique, Sudan, Rwanda and Burundi, Ministry of Foreign Affairs, Evaluation Report 11.98, Oslo, 1998. 80Additional references to specific Bank documents, Board reports or websites relevant to the individual case studies are found in the footnotes to those studies. 49 Woodward, Susan, Bosnia and Herzegovina: Priorities for Recovery and Growth, Brookings Institution, Washington, 1995 World Bank, Bosnia and Herzegovina: Towards Economic Recovery, Report No. 15668, 1996 ______, Developing the Occupied Territories: An Investment in Peace, Oct 1993. ______, Development Impact: Japan Policy and Human Resources Development Fund, PHRD Monitoring and Evaluation Unit, 2000. ______, Greater Great Lakes Regional Strategy for Demobilization and Reintegration, Report No. 23869-AFR, March 2002. ______, Holst Peace Fund ­ Supporting Development in the West Bank and Gaza, 2002 ______, Proposal to Fund Technical Assistance to the USSR, Doc R91-184, July 30, 1991 ______, Trust Fund Handbook, www.wbln0023/Institutional/tfhbk.nsf/Wdocs?openview&count=99 ______, Operations Evaluation Dept, World Bank's Experience with Post-Conflict Reconstruction, 1998. ______, West Bank and Gaza: An Evaluation of Bank Assistance, 2002. ______, Country Case Studies: · Bosnia and Herzegovina: Post-Conflict Reconstruction, (A. Kreimer, R. Muscat, A. Elwan, M. Arnold), 2000. · El Salvador: Post-Conflict Reconstruction, (J Eriksson, A. Kreimer, M. Arnold), 2000 · Uganda: Post-Conflict Reconstruction, A. Kreimer, P. Collier, C. Scott, M. Arnold), 2000 ______, Country Assistance Evaluations: · Cambodia (Report No. SecM2000-708, Dec 2000); · Guatemala (Report No. 25212-GT, Dec 2002) · Haiti (Report No.23637, Feb 2002) · Mozambique (Report No.17209, Dec 1997) · Sri Lanka (Report No. 21771, Jan 2001) 50 ANNEX: Partial List of Persons Contacted 81 A. On Post-Conflict Cases: The following persons are or were directly associated at the relevant times with the major cases examined in this paper and other post-conflict situations. Many have since been working on other assignments and countries, and some have retired from the Bank: Abushakra, Hadi (Afghanistan, WBG) Balaj, Barbara (Lebanon, WBG) Barlas, Dilek (BiH) Bjorgung, Anders (Afghanistan, WBG) Bradley, Sean (Africa) Byrd, William (Afghanistan) Chopra, Ram (WBG) Cisse, Hassane (Africa) Cliffe, Sarah (East Timor) Ebel, Robert (Sudan) Garg, Prem (WBG) Gilbo, Tim (BiH) Herfkens, Eveline (BiH) Horvai, Andras (BiH) Kostner, Markus (Africa DDR) McKechnie, Alastair (Afghanistan) Mertz, Robert (WBG) Michalopoulos, Constantine (Balkans) Nordlander, Karin, LEG (East Timor) Peters, Kyle R. (Kosovo, BiH) Pilch, Gennady, LEG (FR Yugoslavia) Rohland, Klaus (East Timor) Sheehan, Mary (BiH) Van Gelder, Linda (Afghanistan) Walker, Mark (LEG, Kosovo) Wallich, Christine (BiH) B. Others: Balasundaram, Rema (OED) Duvall, Thomas (CoFinancing and Project Finance) Farid, Mazhar, LEGEA Farooq, Zafar, ACTTF Gopal, Rajesh, ACTTN Gruss, Hans-Jurgen (LEGECA) 81 A few other knowledgeable individuals from other institutions were also contacted, but requested not to be identified.