82126 WORLD BANK INTERNATIONAL MONETARY FUND OPERATIONS EVALUATION INDEPENDENT EVALUATION DEPARTMENT OFFICE SUMMARIES OF TEN COUNTRY CASE STUDIES UNDERTAKEN AS PART OF THE IEO EVALUATION OF THE PRSP/PRGF AND OED REVIEW OF THE POVERTY REDUCTION STRATEGY (PRS) PROCESS (report CODE2004-0052) July 6, 2004 -2- Contents Page Abbreviations and Acronyms ..............................................................................................3 Albania.................................................................................................................................7 Cambodia ...........................................................................................................................10 Ethiopia ..............................................................................................................................14 Guinea ................................................................................................................................22 Mauritania ..........................................................................................................................26 Mozambique ......................................................................................................................28 Nicaragua ...........................................................................................................................35 Tajikistan............................................................................................................................41 Tanzania.............................................................................................................................49 Vietnam..............................................................................................................................54 Annex Guidelines for Country Case Studies.................................................................................58 -3- Abbreviations and Acronyms ADB Asian Development Bank APR Annual Progress Report BWI Bretton Woods Institutions (World Bank and IMF) CAS Country Assistance Strategy CDF Comprehensive Development Framework CFAA Country Financial Accountability Assessments CPAR Country Procurement Assessment Review DAC Development Assistance Committee (OECD) DFID Department for International Development (UK) ESAF Extended Structural Assistance Facility ESW Economic and Sector Work (World Bank) EU European Union HIPC Heavily Indebted Poor Countries IDA International Development Association IEO Independent Evaluation Office (IMF) IGR Institutional Governance Review IMF International Monetary Fund JSA Joint Staff Assessment M&E Monitoring and evaluation MDG Millennium Development Goals MTEF Medium-Term Expenditure Framework NGO Non-governmental organization ODA Official Development Assistance OED Operations Evaluation Department PA Poverty Assessment PEM Public Expenditure Management PER Public Expenditure Review PIU Project Implementation Unit PRGF Poverty Reduction and Growth Facility (IMF) PRS Poverty Reduction Strategy PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper PSIA Poverty and Social Impact Assessment SEDP Socio-Economic Development Plan SPA Strategic Partnership with Africa WB World Bank -4- Executive Summaries for Ten Country Case Studies Undertaken as Part of the IEO Evaluation of the PRSP/PRGF and the OED Review of the Poverty Reduction Strategy (PRS) Process This volume contains summaries of ten country case studies conducted for the IMF’s Independent Evaluation Office (IEO) evaluation of the PRSP/PRGF and the World Bank’s Operations Evaluation Department (OED) evaluation of the PRS process. Evaluators interviewed government officials, civil society groups or representatives, international NGOs and other external partners during field visits in all ten countries, and also reviewed internal and publicly available documents prepared by the IMF and World Bank staff. Stakeholders commented on the preliminary findings of the evaluation teams, and in most cases workshops were held at the end of the mission. The governments of the countries were asked to comment on the full case study reports, which reflect their remarks. Background The World Bank and the International Monetary Fund launched the Poverty Reduction Strategy (PRS) Initiative in 1999 to improve the planning, implementation, and monitoring of public actions geared toward reducing poverty. The Initiative centers on promoting a PRS process in each country that is expected to be country-driven, results focused, long term, comprehensive, and partnership-oriented. The introduction of the PRSP was accompanied by the transformation of the Enhanced Structural Adjustment Facility (ESAF)—the IMF’s concessional lending window—into the Poverty Reduction and Growth Facility (PRGF), with a view to giving a more central role to poverty reduction and pro-poor growth considerations in the design of IMF-supported programs in low-income countries. Both the PRSP and PRGF were intended to transform the way the Bank and the Fund operated in low- income countries. OED and IEO evaluations In 2003, the independent evaluation units of the World Bank and the IMF initiated collaborative evaluations of the PRS process and the PRGF to assess whether they had, indeed, resulted in transformations in the approaches of the BWIs to development aid for the poorest. OED, building on earlier reviews of the Comprehensive Development Framework (CDF) and the Heavily Indebted Poor Countries (HIPC) Initiative, undertook an assessment of the PRS Initiative’s progress against its objectives, and the effectiveness of the Bank’s support.1 The IEO reviewed the PRS approach as a framework for the delivery of IMF 1 OED Review of the Poverty Reduction Strategy (PRS) Process (OED 2004, report CODE2004-0052). -5- concessional assistance, the IMF’s own contribution to the approach, and the PRGF.2 The respective Boards of the two institutions discussed the evaluation reports in July 2004. Purpose of this volume The evaluations drew evidence from the experiences of ten countries in which OED and IEO carried out case studies—Albania, Cambodia, Ethiopia, Guinea, Mauritania, Mozambique, Nicaragua, Tajikistan, Tanzania, and Vietnam. Drawing on field visits in each country and extensive interviews with national stakeholders, the case studies provide an in-depth assessment of the impact of the initiative on development policy-making. In order to make this material accessible, OED and IEO have summarized each study and collected the summaries in this volume, designed to complement the main evaluation reports. The full country case studies have been posted on the respective evaluation websites of the IEO and the OED.3 How the case studies were carried out In order to assess changes in the process by which national stakeholders interact to work toward a country’s development goals, the evaluation teams, consisting of international and local evaluators, conducted interviews with government officials, civil society, international NGOs and other external partners in the case study countries and with IMF and World Bank staff in Washington DC. An in-country stakeholder workshop was convened at the end of each field visit. In-country stakeholder surveys were conducted in each of the 10 case study countries. Finally, OED and IEO held a multi-country workshop in Addis Ababa on January 13–14, 2004, hosted by the UN Economic Commission for Africa, to discuss findings across all the case studies.4 Four of the case studies were undertaken jointly by OED and IEO (Mozambique, Nicaragua, Tajikistan, and Tanzania), two by the IEO alone (Guinea and Vietnam), and four by OED alone (Albania, Cambodia, Ethiopia, and Mauritania). Each institution is responsible only for the judgments in the case studies in which it was directly involved. The case study methodology, developed jointly by OED and the IEO, is presented in an Annex. All 10 case studies covered three main areas: relevance of the PRSP approach, the extent to which the 2 Report on the Evaluation of Poverty Reduction Strategy Papers (PRSPs) and the Poverty Reduction and Growth Facility (PRGF) by the IEO is available as EB/EVC/04/3, June 1, 2004. 3 The IEO evaluation website is http://www.imf.org/external/np/ieo/2004/prspprgf/eng/index.htm. and OED’s is http://oed.worldbank.org/prsp. 4 Financial support of the workshop by the Commonwealth Secretariat and the Swiss Agency for Development and Cooperation is gratefully acknowledged. -6- underlying PRS principles were applied and preliminary evidence on the results. In addition, the effectiveness of World Bank support was assessed in the case studies in which OED participated (8 of 10), and similarly, the effectiveness of IMF support was addressed in the case studies in which IEO participated (6 of 10). How the 10 countries were selected The case study countries were chosen to cover a variety of country situations and stages of PRS and PRGF implementation. Selection criteria were: maturity of the PRSP process, geographic balance, coverage of non-HIPC countries, and country initial conditions. Coverage provided by case studies in recent or on-going OED work was also a factor in the selection of OED case studies.5 5 Toward Country-Led Development: A Multi-Partner Evaluation of the CDF (OED 2003) covered the PRSP process in Bolivia, Burkino Faso, Ghana, Uganda, and Vietnam. Debt Relief for the Poorest—An OED Review of the HIPC Initiative (OED 2003) conducted case studies in Cameroon, Guyana, Malawi, Uganda, and Zambia. -7- ALBANIA 1. In the mission’s view the PRSP has added value to Albania’s development planning and management. Over time this is likely to translate into better outcomes for growth and poverty reduction. 2. The PRSP was a highly relevant intervention for Albania. After the setback of the 1997 crisis precipitated by the collapse of the pyramid schemes there was a vacuum of strategic process and direction in the Albanian Government. The PRSP helped to provide a new focus for the Government’s planning system. Eight months were lost between the drafting of the document in 2001 and its eventual submission to the World Bank and IMF Boards in June 2002, due to the unsettled political situation in the country. But this provided time to shore up the weak analytic base. The process was reinvigorated in late 2002 by the decision of the Government to formally adopt the PRSP as a national strategy and to put in place monitoring and evaluation systems within most ministries. 3. With respect to process three aspects of the PRSP have been satisfactory. The Government wisely turned to a group of local consultants to handle the preparation of the PRSP document, given the limited capacity within the Government to focus on medium-term issues. The consultants worked closely with Government officials to ensure that this was a document which would be owned by the respective ministries. The consultations were well planned and reached a wide spectrum of civil society representatives. Discussions were open and reported in the media. The Government has put a great deal of emphasis on monitoring and evaluation of the PRSP; it required individual ministries to put this capacity in place, and funded a central coordinating unit for this purpose. While these achievements are considerable, there is still much to be done in ‘mainstreaming’ the PRSP, by putting in place the structures, processes, regulations and trained staff which can ensure the continuation of the PRSP as the Government’s core strategic instrument. 4. The area of partnership and donor alignment has been a weak aspect of the PRS Process. This is a difficult area in Albania. The country has received very large amounts of aid relative to its size and is of geo-political importance to a number of key donors. It is also an aspirant for EU membership and remains the poorest country in Central Europe. Donors have not been willing to align their agendas with the PRSP and the Government, mirroring these tensions, has been unable to come to a clear approach on the structure and form of its aid coordination activities. 5. The quality of the strategy needs substantial strengthening. The PRSP is particularly weak on the cross-cutting themes which need to be at the center of Albania’s planning in the years ahead: • Private Sector Development: Albania’s growth will depend on the development of small and medium enterprises. This in turn will require improvements in the legal and tax systems and more effective control of crime and corruption. -8- • Improved basic service delivery through decentralization: The Government has a decentralization program, but it is being implemented at a different pace and in a different way by each ministry and there is no clear roadmap which would help to put in place the local capacity needed. • Urban Infrastructure: There can be few European countries with as poor a stock of urban infrastructure as Albania. In most places roads are potholed and electricity and water available for only hours a day. Simple amenities like traffic lights, street lighting, public parks, etc. are few and far between and streets and sidewalks are often strewn with garbage. The problems with infrastructure are a legacy of the pre- transition under investment in maintenance by government and have been exacerbated by the high levels of migration into the cities of the coastal plain which are now ringed by peri-urban squatters and have makeshift infrastructure. 6. The PRSP is making significant contributions to better alignment of the strategy with the budget and to capacity enhancement. The PRSP has taken root as a key part of the overall budget process, linked to the annual budget through the MTEF. This said, the process is at an early stage and it would be hard to argue that there is effective alignment between the PRSP and MTEF at this point. The PRSP has made a useful contribution to capacity enhancement in areas such as statistical development, monitoring and evaluation, and through supporting the public debate on development issues. 7. The PRSP has resulted in a greater focus on poverty reduction. The Government has consistently increased the share of pro-poor expenditures in the budget. Most of these increases went to pay higher salaries for workers in health and education, but this is significant in a context where very low salaries made it difficult to retain qualified individuals and to motivate them to undertake the training and other efforts needed to improve their performance. The expenditure on maintenance of schools and hospitals has also risen, but it has been unevenly distributed. 8. The World Bank’s contribution to the PRS Process has been appropriate and effective in the three areas of preparation, consultation and monitoring and evaluation. Although the Bank played an unduly dominant role in formulating the I-PRSP, it learned from the experience and confined itself to comments on the PRSP itself. The Bank applied what one official called ‘steady pressure’ to ensure that preparation remained on track and that the PRSP remained in view during the year of political instability. The Bank supported the consultation process inter alia through a dedicated international staff member in the Tirana office, and through its dialogue was able to develop a better understanding of the benefits of an effective consultation process. The Bank also devoted substantial resources to training and support for monitoring and evaluation. 9. The Bank’s overall lending program did not support the development of a more strategic focus for the PRSP. The Bank’s lending program in Albania is distributed over a large number of development sectors. The Bank resisted focusing too narrowly on the specific pro-poor sectors, given the huge needs for infrastructure and the shortage of -9- alternative sources of donor support. But at the same time, the Bank program lacked the strategic underpinning which could have made the benefit of the whole program exceed the sum of projects’ individual benefits. Bank staff felt that as long as the PRSC was tackling core strategy issues, the content of the rest of the program was less relevant. As a consequence potential synergies may have been neglected. For example, even where individual projects tackle decentralization issues they do not add up to a coherent and consistent approach which could have supported similar developments in the Government’s program. 10. The Bank was not able to promote a more coherent approach to donor coordination in Albania. The Bank was very active and effective in persuading the EU that the PRSP was well aligned with the EU accession program. But except for a few bilateral donors (most notably DFID, the Netherlands and Switzerland), the donor community viewed the PRSP as a Bank driven initiative. The Bank might have done more to use its own program to support donor alignment through promoting sector wide approaches and increased budget support. It could also have moved away from the approach—widely copied by other donors—of numerous small projects administered through Project Implementation Units. It is difficult to argue that these efforts would have succeeded, however, given divergent donor interests. A more significant step would have been Government leadership in resolving the conflict over responsibilities for aid coordination of the Ministries of Finance and Economy. 11. Improved coordination between the Bank and IMF is an important benefit of the PRSP in Albania. Most observers agree that the recent past has seen a more consistent approach between the two institutions with regard to the balance between expenditures on poverty reduction and the overall budget deficit, and also on the balance between the need for increased tax revenues and the need to retain adequate incentives for private sector activities. - 10 - CAMBODIA 1. Cambodia’s Poverty Reduction Strategy Paper (PRSP) was approved by the cabinet in December 2002 and discussed by the Boards of the World Bank and IMF in February 2003. This case study presents a snapshot of the Cambodian experience with the PRS Process as of mid-2003, covering strategy formulation and the very early stages of implementation. OED concludes that progress toward improved development planning and management is limited. However, it is important to bear in mind that the constraints faced in Cambodia were, and continue to be, daunting (e.g., severe public sector capacity weaknesses, heavy aid dependence, problems with donor coordination). Whether over time the PRS Process will translate into better outcomes will depend on the Government’s willingness and capacity to take steps to institutionalize the PRS Process and the use of the framework by donors to provide coordinated assistance focused on sustained poverty reduction. 2. The objectives of the PRSP initiative are highly relevant to the Cambodian case, though demanding in light of the country’s current development constraints. The focus on poverty reduction is indispensable in a country with major problems of both income and non- income poverty. The PRSP focus on Government ownership and participation also strikes an important chord in the need for institutional development of both the administration and civil society, and the need to move from a donor driven program to one which is both designed and perceived to be home-grown. But the timing of the PRSP preparation pushed by the Bank and Fund was awkward in that the pre-existing planning process supported by the Asian Development Bank (the Socio-Economic Development Plan, SEDP) had just begun a new cycle with a legally mandated time-frame which was too short to allow for the integration of the PRSP principles. Rather than moving quickly to prepare a second planning document, the time could have been better spent in getting a consistent poverty baseline and building the capacity and processes, within the Government, needed for implementing the PRSP principles. 3. Ownership of the PRSP—both the document and the process—resides in a small group of Government officials. The Ministry of Planning (MOP) took leadership of the PRS process and liaised effectively with planning officials in the four key line ministries. But senior officials in the line ministries were not sufficiently engaged, and the Ministry of Economy and Finance (MEF), played little part until the final stages when it was actively involved in drafting the sections on prioritizing and costing the program. As a consequence the PRSP is not perceived as having broad ownership within the Government, let alone outside the Government. The donor community in Phnom Penh is also skeptical of the degree of Government ownership of either the process or the document, perceiving the PRSP as Bank driven with limited Government ownership. This is partly a consequence of the tension with SEDPII and the concern that the Bank was imposing an unnecessary burden on the Government by insisting on a separate strategic process. 4. There was a genuine effort to consult with NGOs in the preparation of the PRSP and to incorporate their views which represented a change on behalf of Government relative to past planning processes, but participation was limited with regard to broader civil society - 11 - involvement, and consultation at the regional and local levels. The NGOs were consulted through three national workshops held in Phnom Penh. While they coordinated among themselves to prepare comments on the various drafts, the time available for this was viewed by them as insufficient. In addition, drafts were only made available in English and this in turn limited the potential for consultation. Parliament and the private sector were not significantly involved in the formulation process; regional workshops were held in only 3 of the 24 provinces; and there was little attempt to involve the media prior to the launch of the final PRSP. 5. The PRSP is helping to make poverty reduction more central to the thinking of the Government, donor and NGO communities. The MOP has made clear to line ministries the need to think through the poverty impact of the programs they support. Donors are also more sensitive to this requirement. However, the PRSP leaves considerable work to be done in identifying clear priorities. Many stakeholders interviewed were disappointed that the PRSP document did not deliver as an operational plan—as advertised to be a complement to the general framework of the second Socio-Economic Development Plan. Ministry submissions during PRSP preparation were generally wish lists of projects for donor support rather than reflecting clear prioritization on the basis of poverty impact as the MOP had requested. There are exceptions that can serve as positive examples, such as the health and education sectors where clearly defined strategies and coverage by the medium-term expenditure framework (MTEF) reflect the substantial past support by donors, in human and financial terms. In other sectors, most critically agriculture, the strategy remains unclear. An effort is being made to improve the alignment of project selection with poverty impact for the authorities’ first PRSP Progress Report. The PRSP does not provide a longer term vision, which can guide short and medium term programs, in some of the key cross-cutting strategic issues facing Cambodia, such as private sector development, the management of natural resources, and civil service reform. 6. There is not as yet sufficient linkage of the PRSP to the MTEF and the annual budget, with little evidence currently available at this early stage that this is impacting resource allocation—the build up of expenditures in the social sectors which pre-dated the PRSP is continuing. As in many other countries, MTEF implementation is at an early stage in Cambodia and lacks ownership beyond the MEF. The limited involvement of the MEF in the PRSP between the I-PRSP and the final stages of document preparation is in large part responsible for the lack of more effective budget alignment. 7. Two areas that warrant attention in the PRS Process going forward are agricultural growth and capacity enhancement. Over 70 percent of Cambodia’s population depend on agriculture as their primary source of income, and most of the poor fall into this category. Productivity in the sector is well below that of Cambodia’s neighbors. The PRSP could be a very effective instrument for bringing together the four concerned ministries, the donor community, local authorities and the many NGOs in the sector, to develop a strategy and coordinate effectively in implementing it. Similarly the PRSP could provide a framework for analyzing capacity enhancement needs and providing guidance on how to address them. A consistent approach needs to be taken to the problem of inadequate training and motivation - 12 - of Government officials which moves away from the arbitrary top-ups for staff in donor supported Project Implementation Units (PIUs) and Technical Units. At the national level the focus needs to be on enhancing the capacity of the judiciary, the police and customs and tax officials so as to reduce the constraints, imposed by corruption, on the growth of small and medium enterprises. 8. The instruments and processes for implementation and monitoring of the PRSP are still at an early stage of development. A central monitoring unit has been established in the MOP, but it is only weakly linked to the data collection units at the central and sectoral levels. In general the areas where implementation is likely to proceed well are the education and health sectors where the capacity of the ministries was strengthened through well- coordinated donor support prior to the PRSP. Implementation in Cambodia is largely handled through individual donor-supported PIUs and there is no consensus on changing this model, which is not consistent with longer term capacity building in this area. 9. Cambodia is highly aid dependent and improved partnership among development partners and with government can yield significant gains in aid effectiveness. The PRSP has increased the potential for improved donor coordination in Cambodia. The PRS Process promoted a framework for discussion of the government’s priorities with donors that did not exist in previous national planning processes. A major objective of the PRSP initiative is to turn aid management into a government-driven process but shifting to government control over the allocation and management of aid will be a long term process. 10. Following the launch of the PRSP, attention has shifted to the role of donors in supporting and aligning to the government’s strategy, as well as synergies among partners. As new country assistance strategies will be developed by several key partners in the coming months, including DFID, WB, and ADB, it is too early to assess the nature and degree of donor alignment. The education and health sectors in Cambodia represent good examples of effective donor coordination. In other areas, however, donors appear to be focusing on their own projects with specific implementation mechanisms which are often not well linked to the ministry in which they are housed. 11. The World Bank’s support to Cambodia in formulating the PRSP was substantial, but its efforts seem to be have been dictated by concerns that the Government should prepare a credible document. Accountability seems to have been less in terms of assisting Cambodia to put in place the PRSP principles. The Bank financed through an IDF grant a support unit in the MOP which was key to keeping the document preparation process moving forward. At periodic intervals, staff from Washington provided comments and advice on the document preparation, and a staff member from the Bangkok Office supported the MOP in getting the consultation process off the ground. A major challenge for the Bank in supporting preparation of the PRSP was balancing quality and ownership, given the government’s past treatment of planning processes. The Bank provided close consultation in the final stages of document preparation that led to positive quality improvement but at the expense of perceptions among some in the donor and NGO communities that the draft document “had gone to Washington”. - 13 - 12. The Bank needs to refocus its support towards the ‘mainstreaming’ of the PRSP principles in Cambodia’s development planning and management. During the formulation process, the Bank was not well placed to provide support for the PRS Process. The country office was relatively new and not staffed at the level needed for such support until quite late in the cycle. The country team did not see itself as having a joint responsibility for assisting the Government and did not put the activities of the Bank in Cambodia into a PRSP context. It was not seen as a change in the business model. Further progress in strengthening the country office and a more cohesive approach on the part of the Country Team will be needed. One important advantage going forward is the recently completed Public Expenditure Review which fills an important analytic gap in underpinning the PRS Process. 13. Though it is premature to assess the effectiveness of the Bank’s support for PRSP implementation, alignment in terms of lending and non-lending activities with the PRS Process is an area of concern that should drive the Bank’s forthcoming Country Assistance Strategy (CAS). The Bank’s analytic work seems to have been a matter of business as usual during the preparation of the PRSP and there was poor sequencing of key PRSP-related products. The Bank could have moved much more quickly to gear up donor activities in support of agricultural development in Cambodia. There continue to be major weaknesses in this sector, which is the dominant sector in the Cambodian economy and a priority area for addressing poverty reduction. The new CAS being prepared provides an important opportunity for the Bank to fully align its lending and analytic work with the PRS Process in Cambodia. 14. Moving ahead, the effectiveness of the PRSP in Cambodia as an instrument for supporting growth and poverty reduction, will depend on the Government and the international community taking a number of steps over the coming months and years. First, the same coordinated approach, which has been effective in health and education, needs to be applied in other key areas: a) agriculture/rural development; b) small and medium private sector development; c) capacity enhancement and civil service reform. Second, an effective cross-ministerial steering group needs to be established with a secretariat responsible for providing regular monitoring reports. These need to be available to the media and the NGOs to ensure proper accountability. Many other supporting steps need to be taken, but these dominate in their urgency and potential pay-off. Given its starting point, Cambodia has moved some way in developing ownership of the program, and carrying out consultations on it. There is a clear upward trend in the understanding of the ministries of the potential inherent in the PRSP approach and the dialogue between government and donors, and among donors, has improved. If this can be maintained then Cambodia should begin to realize the benefits of the PRS Process in the near future. - 14 - ETHIOPIA 1. Poverty in Ethiopia is widespread with almost half of its population of 66 million estimated to be in absolute poverty and an average of 4 million people per year, and between 11 and 13 million during the most recent drought, facing the prospect of serious food insecurity. Conflict, most recently, the 1998-2000 war with Eritrea, and commodity price shocks have only served to exacerbate these problems. In 2002/3, due to a combination of drought and a collapse in coffee prices, real GDP growth plunged to - 3.8 percent. Ethiopia relies on development assistance for 40 percent of public sector spending and is likely to remain heavily dependent for many years to come. Relevance of the PRSP Approach 2. Ethiopia became a pilot country under the Comprehensive Development Framework (CDF) in 1998. In 2000, after gaining eligibility for HIPC II assistance, Ethiopia began preparing its IPRSP. Preparation of the full PRSP began in 2001. While closely linked to HIPC II, the timing of the PRSP also proved opportune in domestic political terms. The period since the war with Eritrea was a particularly active one in policymaking terms in Ethiopia, with several high-level initiatives6 setting out the core development policy and implementation strategies of the country as well as some far reaching changes in the formal structures of government. With significant development challenges ahead of them, the PRSP provided the GoE with an opportunity to articulate their policy agenda to internal and external constituencies while bringing together a range of policy commitments under one comprehensive and long-term framework. The PRSP approach also provided added incentives to focus on results and the resource allocation framework, areas that had been given insufficient attention in the past.7 3. The timing of the PRSP was also opportune from the standpoint of government-donor relations in Ethiopia. Aid relations were in a poor state leading up to and during the war with Eritrea, due to political differences about the war and also because of mistrust that had developed during the 1990s with government concerned about the motives of some donors, and donors concerned about key areas of government policy and a climate of limited transparency. After the war, both government and donors were keen to see an improvement in aid relations. Building on the CDF experiment, the PRSP offered a tangible organizing framework to focus this dialogue and bring to the fore key issues around donor behavior and the role of the international community in supporting poverty reduction in Ethiopia. Application of Principles 6 Starting with the EPRDF congress in 2001. 7 Sectoral strategy documents, for example, in health, education and food security include M&E frameworks but are often quite detailed (the monitoring framework for HSDP I had 110 indicators) and not systematically applied as a measure of progress against specific policies and programmes focused on poverty reduction. - 15 - 4. Overall the PRSP has been well received by government in Ethiopia. Ownership at central government level has been high because of strong continuities between the existing policy agenda of the EPRDF (the ruling party) and the PRSP (SDPRP). The strategy adopts a multi-dimensional approach to poverty and identifies a general M&E framework. There are some signs of second-generation thinking in the PRSP document, particularly around the role of the private sector and how to address key constraints to growth, but in general terms, as government has indicated, the SDPRP is a continuation of existing government policy under a more comprehensive, multi-sectoral umbrella. 5. The consultative process for preparing the SDPRP was broad-based, closely mirroring the federal and administrative hierarchy of the country - a major undertaking given the size of the country and its population. Most respondents to an OED survey in government, the private sector and in the donor community regarded the consultative process as an important first step towards increased voice and local accountability. Woreda (district) level officials noted the benefits in terms of awareness-raising about poverty issues and an increased emphasis on self-help at the community level. Amongst domestic and international NGOs working at federal and at state level, views on the consultative process were less sanguine. The majority recognized the importance of the process and its intrinsic value for those who participated, but a significant number felt that the consultations promised more than they delivered. In particular, there was little space to consider policy alternatives, a finding also supported by survey results. Several were unsure of how their own participation had contributed to the preparation of the document and were concerned that government had sought to get endorsement for existing policy and program commitments without discussing whether or not they were working. The expectation from these groups seemed to be that the consultative process would herald a more fundamental shift in political accountability in the country. Arguably this goes well beyond the PRSP alone. What the consultative process does appear to have done, however, is to expose the importance of the government’s agenda for institutional and political transformation which will prove to be a critical factor affecting the success of PRSP implementation over the medium to long run. 6. The PRSP has also provided a context for a range of new partnerships between government and donors and between government and private sector partners. The 40-strong donor Development Assistance Group (DAG) played a vital role during SDPRP formulation and subsequently in supporting greater alignment and harmonization around SDPRP implementation, while the public-private sector consultative forum was established to address major problem areas identified by private sector partners during the consultative process. Preliminary Evidence on Results 7. Although the SDPRP was only completed in July 2002 and the year following it has been marked by one of the worst food crises for many years, there is some preliminary evidence of progress as a result of the PRSP process. Bank staff in particular point to important shifts in government’s attitude towards policy discussions in areas previously ‘out of bounds’, such as business regulation, land tenure, telecoms and financial sector - 16 - development. The hope is that this will lead to some rebalancing of old policy agendas in favor of more market-based approaches. However, no significant policy adjustment has yet taken place and donors are aware that the SDPRP remains only one of a number of government agendas, including outstanding issues with Eritrea and increasing national security concerns. 8. Progress is also evident in the recent push towards more devolved and responsive local government. While the system is still in flux the decision to push ahead quickly with un-earmarked transfers (block grants) to the woreda level in the four largest regions (Amhara, Oromia, SNNP and Tigray), has provided a clear indication of the government’s intention to address critical service delivery blockages and promote downward accountability—both central to SDPRP implementation. A major challenge will be providing the right incentive environment and overcoming capacity weaknesses at lower levels of government. 9. Progress with monitoring and evaluation around the SDPRP was initially slow. Interviewees from within and outside government spoke of a loss of momentum around M&E linked to the emergency situation, the demands of the new block grant system and the time taken up by donor-related processes, including work on harmonization.8 At the time of the evaluation mission progress had picked-up, mainly because of the need for government to prepare a common performance assessment tool for PRSC-budget support from donors. The Annual Progress Report submitted to the IFI Boards in December 2003 sets out a comprehensive plan for an M&E system, however, progress in developing key indicators, clarifying how they will be monitored and how the M&E system will work in the context of decentralization is still required. 10. Although Ethiopia stands out amongst low-income countries as having a fairly sound PEM system, progress with internal budget alignment with the SDPRP has also been slower than hoped. The medium term macroeconomic and fiscal framework (MEFF) is still something of a shadow exercise and the full procedural integration of recurrent and capital budgeting is yet to be effected. A comprehensive program for public expenditure reform is in place but the pace of change has been uneven, with the demands of the government’s decentralization program placing the system under added strain, including ongoing efforts at budget consolidation. Budget out-turns for 2002/3 were above those in 2001/2 but well below those forecast in the SDPRP. Out-turns for poverty targeted expenditures amounted to 77.6 percent of budgeted levels. Nevertheless budgeted levels for 2003/4 continue to show strong optimism on the part of government. The APR details expected increases ranging from 42.5 percent in agriculture and food security to 10.6 percent in health and 2.3 percent in education. No clear indication is given of the basis for such projections in the view of the estimated out- 8 Somewhat ironic given that donor efforts at greater harmonization, especially in relation to budget support, are reliant on government having a credible performance assessment, and hence M&E framework in place. - 17 - turns for 2002/3 nor is there any further prioritization of core poverty programs, particularly given expected shortfalls in required funds over the medium term. 11. Of all the principles underlying the PRSP approach, it is the context for aid partnerships that has probably changed most significantly. Since completion of the PRSP, government and donors have formed a joint task force on harmonization and agreed on architecture to facilitate a coordinated dialogue on high-level policy issues. The establishment of a High Level Forum for taking forward government-donor dialogue was one of the first of its kind in the region. Ethiopia was also the first country to host an SPA mission to discuss ways of aligning multilateral and bilateral budget support behind the PRSP. Since the SPA mission, budget support donors have formed a DBS group and have agreed with the World Bank to align any DBS arrangement with the PRSC. The PRSC appraisal mission in August 2003 included bilateral donor representatives committed to providing DBS. Further evidence of progress with harmonization was difficult to find, however. While progress had been made in pulling together a harmonization framework covering donor project procedures, procurement and fiduciary management standards, there was still limited progress9 by donors in moving the framework forward. At the time of the evaluation mission there was some frustration emerging over the harmonization agenda as donors confronted the practical (and political) difficulties of signing up to procedural changes at country level and in donor capitals. World Bank Effectiveness 12. The Bank has played a consistently supportive role in the PRSP process in Ethiopia. Its role has been in line with the principle of country ownership while supporting the government with advice and technical inputs as and when required. The PRSP Task Manager and the country team played an active and constructive role in guiding and supporting the process. The Bank also provided various forms of technical and analytical input to enhance local capacity, for example, in support of an M&E framework for the SDPRP. Workshops and seminars were provided through a combination of Bank and WBI support, including a workshop for NGOs on the PRSP approach towards the start of the process. 13. Bank engagement was also shaped by the CAS preparation process, which ran virtually in parallel with SDPRP formulation. The consultations around the CAS—mainly a series of high level workshops during 2002/3 on themes such as decentralization, private sector development, capacity building and rural development—were intended to pick up on issues and themes that the government was grappling with in the SDPRP. The country team viewed the two processes as more or less synergistic, despite some Bank staff being concerned about possible challenges to government ownership of the PRSP process. There is some evidence to suggest that the CAS consultations did have some influence on government thinking, especially around the PSD agenda, but equally there is limited evidence to suggest 9 Confirmed by the donor group in their Joint Partner Review of the APR. - 18 - that this in anyway challenged government ownership.10 What is harder to determine is whether by running the CAS preparation process in parallel with the PRSP process there were any added costs to government. In terms of time costs on government, the only real indication was a certain frustration from some government officials about the number and frequency of workshops hosted by the Bank and others during 2002/3. In the circumstances, should the CAS process have been delayed? It is hard to come up with a simple answer. The consultations did overlap with SDPRP preparation but the final CAS document was delayed to ensure that presentation to the Board did not precede SDPRP endorsement. In addition, the Interim Support Strategy had come to an end requiring that a CAS be prepared as soon as possible. The strong synergies between the CAS and the SDPRP is in accordance with Bank guidance, although avoiding too much additional burden on government during a parallel process such as this seems equally important. 14. Responses from government officials about Bank support for the PRSP were almost universally positive. Government officials also noted that relations with the Bank had improved with the increased status of the country office and the Country Director based in Addis. This was also echoed by members of the DAG who had witnessed a significant improvement in donor-Bank relations since 2000/1. Particular high points in collaboration were the common position on revisions to the SDPRP policy matrix, joint participation in the SPA mission, combined support for work on harmonization and the joint preparation of key analytical products such as the CFAA, CPAR and PER. In relation to the PRSP principles themselves, several donors acknowledged that the Bank had acted appropriately in ensuring that the government was able take control of the process. 15. Some different views were expressed, however. Representatives of NGOs interviewed during a roundtable meeting were unsure of whether the Bank was playing a role as a referee (who also sets the rules) or a team player. They felt that expectations had been heightened around the room for policy adjustment as part of the PRSP process only to find that such room disappeared during the consultations themselves. The Bank’s support for country ownership was also perceived to be a double-edged sword by some non- governmental stakeholders. On the one hand, they agreed that a country-led, country owned process was crucial. On the other hand they were concerned that government paid insufficient attention to their views and demands for inclusiveness, and wanted the Bank to advocate more strongly on their behalf. This kind of double bind is familiar to Bank staff, but once again it suggests a mismatch in expectations. Perhaps a more comprehensive communication strategy at the beginning of the PRSP process could have helped set out clearer expectations about the process, the role of government, donors and NGOs, and what could reasonably be achieved? 10 Some NGO representatives were concerned that there was more ‘open’ debate during these CAS workshops than there was during the PRSP consultations. - 19 - Bank Support and Alignment 16. The CAS clearly defines its objectives in terms of support for SDPRP implementation. It also goes further in detailing weaknesses in the current strategy for promoting growth and private sector development and reducing vulnerability and managing risks. The assistance program focuses on three clusters of objectives – enhancing pro-poor growth, enhancing human development outcomes by improving governance and reducing vulnerability – that track closely with the four pillars of the SDPRP. At the heart of the assistance program is a series of PRSCs intended to support each of the strategic objectives of the CAS. Initial preparations for the PRSC coincided with the SPA mission in September 2002, so from the very beginning, the PRSC appraisal process was connected to wider discussions about alignment and harmonization with the SDPRP. Other new aid modalities are also considered, including capital grants to local authorities for infrastructure development and a programmatic approach to public sector capacity building. Perhaps one of the most innovative elements of the CAS is the plan to use the non-lending program not only as a tool for furthering national dialogue but as a way to strengthen monitoring and evaluation of key policies included in the SDPRP. The Institutional Governance Review (IGR) was already being prepared in this vein. A key challenge is to ensure adequate buy-in from government so that the results of such exercises can utilized as part of the SDPRP review process. 17. While innovative, the CAS program still remains very wide-ranging. It covers a broad spectrum of projects and programs across PSD, roads, energy, rural capacity building, health, education, water, empowering civil society, pastoral development, drought recovery, HIV/AIDS and public sector capacity building. The breadth of the program reflects the scaling up of IDA assistance following the end of the war and the fact that the SDPRP itself is a wide-ranging and relatively weakly prioritized document. The CAS explains the breadth of the program in terms of IDA's dominant role in Ethiopia (it is larger than all other donors combined) and its commitment to supporting the transition from SDPRP vision to tangible improvements in policies and capacities. The expectation is that in the medium to long term, as other donors scale up their programs, IDA will become more selective, focusing in areas attracting less support from other donors, such as infrastructure. Compared to recent CAS's, the current CAS is ambitious with a much larger IDA allocation. The challenge for the Bank will be in managing such a broad (and high volume) strategy and ensuring sufficient complimentary with what other actors are doing or planning to do in support of the SDPRP. There are also some risks to moving upstream acknowledged by the Bank team. As the CAS states "...the Bank will closely monitor results during CAS implementation, and if necessary, move back to more traditional instruments” (2003:43). 18. In terms of ways of working, the PRSP approach, along with the PRSC, have resulted in some important changes. There is more emphasis on multi-sectoral working in the country team, there is greater collaboration and information sharing with other donors through the DAG, and the on-the-ground presence of the Country Director has definitely improved relations with both government and other donors. However, the Bank’s size and dominance in policy dialogue in Ethiopia remains a concern to some donors. Although the PRSP has - 20 - increased opportunities for joint working within the donor community there is still a concern from some of the smaller donors that the Bank can go it alone when it wants to. The breadth of its CAS only serves to fuel this worry. Remaining fully engaged in donor coordination processes will be vital to ensure that this doesn’t happen. Bank-IMF Cooperation 19. Bank-Fund collaboration was not a topic many in Addis felt able to comment on. Nevertheless stakeholders with some contact with both institutions, and particularly amongst donors, acknowledged that relations between the Bretton Woods institutions had improved and had ‘warmed up’ significantly during the PRSP period. 20. From the perspective of both Bank and Fund staff based in Addis and in Washington, the PRSP process had been a period of constructive engagement. The IMF resident representative had participated in several consultations and was an active member in DAG discussions. Washington-based Fund staff were less involved on a regular basis but did take the lead in encouraging government to undertake an assessment of the sources of growth and to consider alternative macroeconomic scenarios in the SDPRP. The IMF country team also played a central role in preparing the JSA in collaboration with the Bank country team. Donor representatives noted several areas of increased cooperation between the Bank and Fund, most recently around financial sector development and the inclusion of a Bank country economist on the latest PRGF review.11 The Fund resident representative also participated in various PRSC missions, with the office economist participating in almost all of the August 2003 PRSC mission meetings. 21. Areas where collaboration could have been better included the early stages of the CFAA, where there was the perception that the IMF could have supported (or been asked to support) the process more directly and, as a result, taken greater ownership of the findings.12 IMF participation in PERs was also an area that some—particularly donors—felt needed strengthening. The IMF did commit resources to the 2001 PER, including drafting a chapter on fiscal policy, nevertheless the perception was that Fund ‘ownership’ of the PER’s findings was limited with implications for the take-up of priority issues in the PRGF. In the most recent PER 2003, the Fund was not invited to participate, this time due to the dominance of bilateral donors. However a future role for the Fund in a strengthened PER process seems vital, especially in enhancing future PRGF-SDPRP alignment. 22. At the time of the evaluation mission, some senior staff in MOFED were still expressing reservations about the IMF’s stance in Ethiopia. Some were concerned about a possible disconnect between the macroeconomic framework set out in the PRGF and the financing needs of the PRSP. Although the PRGF was considered to be consistent with the 11 According to Fund staff, Bank staff have been invited to participate in all PRGF missions, though by their own choosing they did not attend all meetings. 12 The resident representative did participate in CFAA discussions alongside other donors. - 21 - policy thrust of the SDPRP, the perception was that the Fund’s programming rules were not altogether in alignment with the financing demands of the SDPRP. The Fund’s view, however, was that the authorities assumptions in the SDPRP were too optimistic and a more realistic macroeconomic framework based on achievable growth and revenue targets and identified financing was necessary to ensure that programs were prioritized and targets could be met. The PRGF review in 2002 did allow for some relaxation of the deficit rules following requests for greater flexibility from MOFED, but the perception remained that more should be done to address the financing needs of the PRSP and ultimately, Ethiopia’s realization of the MDGs. Conclusion 23. The Bank has played a constructive role in the PRSP process in Ethiopia. Ownership of the strategy by central government is clear. The Bank program is broadly aligned with the SDPRP. Progress has been made in taking forward some key areas of the government’s program, most notably in telecoms and private sector development. Tangible progress has also been made in donor alignment with the SDPRP. While the SDPRP is still in its infancy, the evaluation uncovered some potentially divergent expectations about the process. The government views the SDPRP as an important opportunity to articulate its longstanding policy program to both internal and external constituents. Many donors view it as the first step on the road to more responsive and transparent government and a new context for aid relations. CSOs and NGOs view the PRSP approach in a more radical light, hoping for a more fundamental shift in the political economy of policy making and the consideration of policy alternatives. Although these expectations are not necessarily in conflict, at least over the medium to long term, they are the source of some tensions already. Being aware of these differing expectations and finding the right balance between respect for country ownership and providing intellectual and financial leadership is likely to be a critical challenge for the Bank going forward. - 22 - GUINEA 1. Guinea is a small country endowed with rich natural resources but very low human development indicators and a GDP per capita of only US$550. A centrally planned economy until 1984, it has since embarked on a gradual process of economic opening and liberalization. Public finance imbalances and weak governance are among the major challenges it faces today. From 1995 onwards, Guinea launched several participatory initiatives to draft development strategies that, although incomplete and not directly operational, provided a good starting point for the PRSP process. Guinea was declared eligible for the HIPC initiative in December 1999 and reached its decision point a year later, having prepared an interim PRSP in October 2000. The full PRSP was completed in January 2002. Guinea has been engaged in IMF-supported programs (first under the ESAF, then under the PRGF) almost continuously since 1987, although with a poor track record of implementation. At the time of the evaluation, the current PRGF-supported arrangement— approved in May 2001—was off-track. 2. While the key incentive for Guinea’s PRSP was that it conditioned access to debt relief and concessional IFI lending, the formulation process was country-driven in the sense that the government organized the process and drafted the document without pressures from the donor community. The participatory element rested on the work of seven thematic groups composed of representatives of government, civil society and donors, and on a series of workshops with broad public attendance held both in the national capital and at the decentralized level. The IMF kept a low profile throughout the process, largely limiting its involvement to observing the debates. 3. Stakeholders generally viewed this process as having significant value added over past practices and particularly appreciated the government’s outreach efforts to present its policies to the public. At the same time, concerns were expressed about the representativeness of participants to thematic groups, and the actual impact of the participatory process on the final PRSP appears to have been limited, particularly in the macroeconomic area. It was more significant with respect to the poverty diagnostic and the prioritization of the policy agenda—bringing to the fore, in particular, governance issues. Factors mentioned to account for the limited impact include weak civil society capacity to participate, restrictions to the freedom of expression of actual and potential stakeholders and poor communications. All these factors were accentuated by the modalities of the consultation process itself (such as limited access to the documents under discussion, absence of choice between alternative policy options and of discussion of the tradeoffs involved, insufficient time, strong government involvement in the process etc.). As a result, it does not seem that the process led to broad-based ownership of the PRSP, even within government broadly defined, although it is strongly owned by the relatively narrow circle of civil servants most closely associated with its drafting. 4. Guinea’s PRSP is generally viewed as a good “document” in the sense that it contains all of the expected building blocks and that it adopts a comprehensive definition of poverty, identifies expected pro-poor outcomes in some specificity and contains elements of - 23 - a long term perspective. At the same time, it suffers from critical weaknesses which have limited its usefulness as an operational roadmap for policy making to date. In particular, its analytical contents are poor, providing few insights on the links between policy actions and poverty/growth outcomes. The macroeconomic framework is based on overoptimistic assumptions, and not well integrated with the rest of the strategy, and sources/factors of growth receive only limited attention. Furthermore, there is a general sense that the policy agenda contemplated in the PRSP is at variance with the government’s implementation capacity, and that it is insufficiently prioritized, including in the sense that it does not offer guidance to solve tradeoffs between competing objectives, or contingency plans to adapt to changes in the external environment. 5. The joint staff assessment of Guinea’s PRSP fell short of meeting several of its purposes. It offered a balanced assessment of the PRSP’s contents and was candid in pointing out many of the areas that required continued attention from the authorities going forward, which was valuable to the authorities. However, it was less candid in its description of the participatory process which was assessed very favorably. Likewise, the assessment tended to downplay implementation risks. In both areas, the JSA was at odds with the perceptions of most nongovernmental stakeholders on the ground. The JSA did not seek to incorporate inputs from other parties than the authorities and received very little dissemination in Guinea. Thus, most stakeholders were either unaware of it or viewed it as irrelevant. IMF Board members, when reviewing Guinea’s PRSP, made few references to the JSA, but when they did, accepted its appraisal at face value. 6. Limited changes have been observed in the IMF’s way of doing business under the PRGF. Some efforts were made in the early days of the initiative to undertake limited qualitative social impact analysis of key elements of the program, but they have faltered since. There is also evidence that more policy space is given to the authorities in the negotiation of structural conditionality, and that meeting macroeconomic targets by squeezing social spending is now viewed by the IMF as an unsustainable strategy, that should not be supported. However, the IMF failed to use the opportunity of the PRSP process to ensure that key policy issues and tradeoffs in its areas of competence were duly aired and debated. At the same time, IMF staff’s outreach efforts to explain the rationale for the PRGF- supported program to a broader audience than the authorities have been very limited. Key explanations for these shortcomings include lack of time and staff resources, reluctance of the authorities, and lack of clarity about what was expected of the staff in what was meant to be a government-led process—especially if some key policy issues were not emerging in the public debate. 7. The design of the PRGF-supported program reflects some progress towards most of the seven key features supposed to distinguish the PRGF from the ESAF, but it is still a long way from being fully embedded in the PRSP: • Alignment: The PRGF was approved before the finalization of the PRSP, and its original framework adopted as macroeconomic framework of the PRSP. That framework, however, was soon overtaken by adverse shocks, so that by the time of - 24 - the first review of the program, there was a clear mismatch between the macro framework of the PRGF-supported program, which had been updated, and that of the PRSP, which had not. Moreover, the latter did not provide guidance as to how to handle tradeoffs required by such shocks, leaving the PRSP underfinanced and with unattainable macroeconomic objectives. Regarding structural policies, by contrast, the PRGF-supported program and the PRSP are broadly aligned. • Program design did protect priority sector spending and allowed for substantial flexibility in fiscal targets, both to accommodate higher pro-poor spending and in response to shocks. However, that flexibility was not always present in the initial program design and instead often resulted from ex post adaptations which, not being determined on the basis of fully transparent criteria, gave rise to misunderstandings and criticism of arbitrariness, both with the authorities and with the donor community. Also, problems remain with the efficiency and targeting of “pro-poor” spending. • Structural conditionality was streamlined in the sense that the number of formal conditions in the program declined and that they became more focused on IMF core competencies. Conditionality was streamlined in areas where either the authorities had demonstrated strong ownership of the reform agenda or an understanding was reached with the World Bank to make it the “lead agency” in that area. Within that framework, significant emphasis was put on governance and public accountability. However, this trend coincided with the proliferation of informal conditionality, frequently in the context of informal interim programs negotiated to put performance back “on-track.” These interim programs had little linkage to the PRSP. • Areas where progress has been lacking include the use of poverty and social impact analysis in program design, and the realism of key underlying macroeconomic assumptions. 8. Collaboration between the IMF and the World Bank was significantly enhanced at the working level from late 1999 onwards, thanks to systematic joint missions to the country and a clear articulation of respective “lead agency” responsibilities, leaving few areas of overlap. However, collaboration in connection with the streamlining of conditionality encountered problems, most critically in the area of SOE reform, where the indicated reform outcomes were not achieved. This resulted in large part from lack of country ownership of the reforms concerned, but also from differences in approach to conditionality and lending instruments between the Bank and the Fund that implied limited leverage in the timeframe relevant for the Fund-supported program. The PRSP process itself did not include a significant debate on SOE reform and alternative policy options, which were consequently left to more traditional negotiating vehicles and conditionality, which proved ineffective. 9. The impact of the PRSP has been limited to date, in part because its implementation has been hindered by substantial shortfalls in external financing, but also for - 25 - lack of government support. Impact on the policy making process is visible only, to some degree, in the budgetary process, where priority sectors have been protected, and through some progress in decentralizing revenues and spending authority. Impact on program ownership and on the policies adopted is, likewise, hard to detect, including in areas, such as governance, that the PRSP process identified as critical. Impact on donor coordination has been elusive, as most donors have taken a wait and see attitude towards the PRSP process, while reducing their aid owing to governance-related concerns (including concerns related to the political aspects of governance that go well beyond the PRSP). More critically, the sustainability of the process is far from ensured, as two years after the adoption of the full PRSP, there were no clear institutional arrangements for its implementation, monitoring, and adaptation over time. 10. The main lessons arising from this case study are as follows: (i) initial expectations of what the PRSP process could deliver in the short term in the Guinean context were probably too high given in particular the political context and the fragility of macroeconomic stability; (ii) for PRSPs to be effective as an operational roadmap for policy making, they need to provide strategic guidance for setting priorities among competing demands and for solving unanticipated tradeoffs. They also need to be backed by working institutional arrangements for their implementation, monitoring and regular updating; (iii) in their assessment of PRSPs, the staffs of the IMF and the World Bank should focus on whether or not these elements are present at least as much as on the specific contents of PRSPs as documents or on whether various process requirements have been met. Taking into consideration the views of local stakeholders in the JSA process could help reach a balanced and candid assessment; (iv) IMF staff needs to be proactive in the discussion of the macroeconomic framework of the PRSP to ensure that it offers a suitable basis for policy formulation. Without restraining the policy space, it should in particular ensure that major policy issues are discussed, alternative options considered, and likely tradeoffs considered; (v) IMF staff need to be more explicit about the criteria taken into account to reconcile macroeconomic stability and pro-poor/pro-growth policies in the formulation and monitoring of PRGF-supported programs; (vi) streamlining conditionality has two distinct dimensions: enhancing ownership and improving the division of labor between the IMF and the World Bank. With regard to the first, problems can arise if the PRSP is overly vague about the strategy envisaged, thereby leaving the key elements to be addressed as part of negotiations on the PRGF-supported program. On the second, the integration of Bank and Fund conditionality can be highly complicated in practice and the assumption that conditionality will be more effective simply by letting the Bank take over responsibility for conditionality in areas where the Fund is reducing its direct involvement may not be warranted, as the example of conditionality on state enterprise reform illustrates. - 26 - MAURITANIA 1. The PRSP process is relevant for Mauritania and represents a more comprehensive approach to poverty reduction than in the past. The PRSP built on a pre-existing Government focus on poverty reduction which had already resulted in upward trends in poverty reduction and social indicators. Mauritania was one of the first countries to join the PRSP initiative and was motivated to complete a PRSP by the prospect of debt relief under HIPC. An Interim PRSP document was not prepared and a full PRSP was completed in one year based on the existent outdated poverty diagnoses. These deficiencies were addressed in the two Progress Reports when more recent data was used to update the strategies. 2. Ownership of the PRSP by the Government was strong at the outset of the process and the final draft of the PRSP document was felt by the Government to be country owned and an improvement on the past frameworks for donor assistance such as the PFP. The PRSP was supported by high-level leadership in Government and was approved by Parliament. The participatory process undertaken during PRSP preparation was more than had previously been attempted in Mauritania, and a number of external commentators have perceived the process to have been successful. Stakeholders in-country, however, report excessive Government control over the choice of participants and issues discussed during formulation of the PRSP and note that the depth of participation has not evolved during implementation. It is also difficult to assess to what extent participation has affected or added value to the PRSP program. Sustained Government commitment to the PRSP is an area of concern in Mauritania. The implications of recent political instability in the country are unclear; after achieving irrevocable HIPC debt relief, stakeholders perceived there to have been policy fatigue on the part of Government. 3. Implementation has lagged behind original plans due to overoptimistic targets and capacity constraints. Expected costs of the PRSP program have varied from year to year and remain overoptimistic. The first Progress Report included a vast increase in the projected allocations for education and health, although the specific programs remained constant. These cost estimates were subsequently reduced, as expenditures on PRSP priorities reached only 43% of the target. Also, while allocations for growth promotion are projected to grow to almost 30% of total PRSP expenditures by 2004, actual spending only reached 0.4% of total PRSP expenditures in 2002. Despite analytic work undertaken by donors including the Bank, the Government’s limited absorptive capacity is a serious risk to implementation of the PRSP program. The increased role of the Commissariat in PRSP implementation has also raised concerns over fiduciary controls and safeguards. 4. The Bank’s program has been well aligned with the PRSP priorities. The ESW program provided inputs into the PRSP preparation process. The CAS of May 2002 states that the Bank’s main objective for FY03-05 is to assist the authorities in implementing the PRSP and the proposed operations are in fact designed to support implementation of the four strategic axes of the PRSP. The level of Bank lending has not, however, kept pace with that foreseen in the CAS and the PRSC planned for FY03 has not materialized. The near-term - 27 - introduction of PRSCs will challenge the implementation capacity of Government and there are fiduciary risks associated with the implementing bodies. 5. The PRSP provided a relevant mechanism for donor coordination beyond the CG framework and donors have sought to align their programs with country priorities as set out in the PRSP. OED interviews with donors suggest initial enthusiasm diminished as donors came to see the PRSP process as being directed from Washington. Informants specifically referred to BWI control over the deadlines and structure of the PRSP, and to instances where the Bank met separately in Washington or Paris with the Government in order to prepare draft documents before other donors had the opportunity to make any input13. 6. Donors report that they have an excellent relationship with the Bank’s office in Nouakchott. However, in some instances poor timing on the Bank’s part limited the participation of donors in the PRSP process, for example, the presentation of the PRSP to the BWI Executive Directors significantly in advance to its enactment into law and prior to its presentation at the Consultative Group Meeting to donor partners. 7. The quality of the Joint Staff Assessments is uneven and the mission identified many gaps in the JSAs for both the original PRSP and the two Progress Reports. The JSAs did not discuss key issues such as public sector efficiency, expenditure levels, implementation capacity. Revenues from hydrocarbon are expected to constitute a major source of income for the county, but the use of the PRSP to channel hydrocarbon resources into poverty alleviation has not yet been discussed. The JSA of the PRSP assesses the quality of participation despite the JSA guidelines and without identifying objective, measurable criteria for assessing participation. JSA certification of a PRSP process as “participative” should not be confused with frequent, transparent, accountable and democratic political processes. 13 According to Regional Bank staff, the practice of meeting in Paris/Washington to prepare draft PRSP-related documents is now actively discouraged by the Bank. - 28 - MOZAMBIQUE Overview 1. In Mozambique, poverty reduction as a key objective of government policy predates the introduction of the PRSP approach by the IMF and the World Bank. PARPA— Mozambique’s PRSP—drew heavily on existing sectoral policies and programs, but it also put more firmly on the public agenda issues related to participation in domestic policy processes (including the role of parliament and political parties), and has brought greater attention to implementation constraints through a focus on monitoring results. It has become the common point of reference on poverty reduction policies within government—across line ministries and between central and provincial levels of government—and for dialogue between the government, other national stakeholders, and international development partners. 2. The principal objective in the PARPA is reduction in the incidence of poverty from 70 percent in 1997 to less than 60 percent in 2005, and to less than 50 percent by 2010. Six priority, and 11 complementary, areas are identified to promote human development and create a favorable environment for rapid, inclusive and broad-based growth. The priority areas are: (i) education; (ii) health; (iii) agriculture and rural development; (iv) basic infrastructure; (v) governance; and (vi) macroeconomic and financial policies. Application of underlying principles 3. The PRSP approach has proved to be relevant to Mozambique, and the underlying principles have been applied to varying degrees in the PARPA process. Initially, the authorities resented what they saw as an additional externally-driven procedural requirement. Overall, however, the process has been country-driven with strong government ownership. A draft of the PARPA was subjected to a consultation process that involved the participation of business associations, labor unions, religious bodies, nongovernmental organizations, media, central and provincial government institutions and donors. Even though the consultation process had its shortcomings—for example, representatives of civil society organizations have pointed to insufficient time for preparation (e.g., materials were distributed late, and there was insufficient time for CSOs to consult members at grassroot levels)—the consultations did lead to significant changes to the draft, most notably the inclusion of good governance as one of the “fundamental priority” areas. However, after approval by the government, little has been done to disseminate PARPA (including preparing popular versions in main local languages); it is largely unknown in any detail outside Maputo and provincial capitals. 4. The PARPA process is partnership-oriented and builds on existing mechanisms for government-donor relations. The PARPA process was launched at a time when there was broad consensus between government and donors on key priorities for poverty reduction. Most of the sector policies around which PARPA was built had been developed by the government in close cooperation with donors. This relatively harmonious process reflected the fact that Mozambique already had a comparatively well developed donor coordination - 29 - mechanism. A number of Sector Wide Approaches (SWAps)—including in health and agriculture—have existed since the mid-1990s. Also, several donors in Mozambique were coordinating their balance of payments support prior to the PARPA process. 5. The PARPA recognizes the multi-dimensional nature of poverty and adopts a comprehensive approach to poverty reduction. It employs a wide range of indicators for measuring poverty; e.g., in addition to income- and consumption-based indicators, it also used available data on illiteracy, mortality and water access rates, and drew on the results of a qualitative participatory poverty assessment. Reflecting its broad definition of poverty, the PARPA adopts a comprehensive approach to poverty reduction that includes broad-based growth in a framework of macro-economic stability, actions to enable market-based rural development, and social service delivery. Although it provides a good description of regional differences of poverty incidence, it does not provide provincial- and district-level targets. There is a need for greater clarity on how national average targets relate to provincial and district priorities. 6. The PARPA is based on a long-term perspective, but its operational framework is appropriately medium-term oriented. The medium-term nature of PARPA is underlined by two additional factors: government intention to link future versions of PARPA to the election cycle (specifically, the economic program presented to parliament by a newly elected government), and ongoing work towards defining a “Vision 2025.” 7. The PARPA is results-oriented, but institutional weaknesses pose a challenge for translating plans into action. The operational targets are largely taken from sectoral plans, and vary considerably in scope and precision: some are set at the output level (e.g., enrolment rates in primary education), others are input-oriented (e.g., curriculum reform, teacher training), while others are set on intermediate outcomes (e.g., mortality rates). In a few areas, targets were not yet specified (e.g., anti-corruption). Implementation, monitoring, and preliminary results 8. Integration of PARPA into the government’s planning, budgeting and reporting processes is underway but at a slow pace. The annually updated five-year medium-term budget framework (MTBF) is the principal instrument for translating the public expenditure priorities in PARPA into budgetary allocations. Weaknesses in Mozambique’s public expenditure management system need to be addressed to ensure that budgeted funds are available to spending units and that the funds are spent as budgeted. Quality control (i.e., ensuring quality of spending) and monitoring and reporting functions in government also need to be enhanced. Senior officials emphasized to the evaluation team that procedures for monitoring and reporting on progress in implementation should be oriented around the strengthening of domestic processes, rather than fulfilling BWI procedural requirements. In this regard, the recent merging of the annual PRSP process report and the review of the 2003 Economic and Social Plan (submitted to Parliament) is an important step forward. - 30 - 9. A high level forum—the Poverty Observatory—has been established to bring together the government, a broad range of national stakeholders and Mozambique’s international development partners to review PARPA implementation on an annual basis. The focus is intended to be on drawing lessons from experience and coming up with recommendations for improving implementation and monitoring. 10. Evidence on progress towards meeting PARPA targets and objectives is mixed, with significant progress in some areas (e.g., education and health), and less progress in others (e.g., agriculture). But in most areas, there is uncertainty about the extent to which progress is being made in improving quality of service delivery. These uncertainties reflect system weaknesses in budget execution, monitoring and reporting, and are being addressed as part of the reforms of the public expenditure management system. Enhancing capacity 11. There is a wide range of ongoing reforms in the public sector aimed at addressing capacity weaknesses related to implementation and monitoring of the PARPA. Although these reforms are meant to address capacity constraints, they appear to be taxing current capacity to the limit. At the same time, different levels of government (i.e., central, provincial and district) are putting great stock in these reforms—especially aspects dealing with decentralization of the planning and monitoring system—to improve the implementation of government policies and programs. 12. The PARPA process seems to have enhanced policy discussions on poverty issues within the government (especially between central and provincial government officials), and to a lesser extent, between the government and non-government stakeholders. However, a broad range of those met by the evaluation team stressed the need to strengthen policy analysis capacity in and outside government for a more meaningful and sustainable participatory process. IMF effectiveness 13. Reflecting the wishes of the authorities, IMF staff did not participate directly in the preparation of the PARPA. In particular, there was no IMF staff participation in the government-led consultations with stakeholders. 14. When the ESAF was transformed to the PRGF in 1999, this was supposed to signal a new way of doing business for the IMF. In particular a number of “key features” were supposed to distinguish PRGF-supported programs from those supported under the ESAF: (i) broad participation and greater country ownership; (ii) embedding the program in an overall strategy for growth and poverty reduction; (iii) government budgets that are more pro-poor and pro-growth; (iv) appropriate flexibility in fiscal targets; (v) more selective structural conditionality; (vi) emphasis on measures to improve public resource management and accountability; and (vii) social impact analysis of major macroeconomic adjustment and structural reforms. - 31 - 15. The evaluation’s findings on the application of the “key features” in Mozambique’s PRGF-supported programs are as follows: • Participation in the formulation and monitoring of IMF-supported programs has remained narrow, but there is some indication of greater country ownership of the program. • The macroeconomic framework of the pre-existing PRGF-supported program influenced the PARPA’s macroeconomic framework, but the PRGF objectives have become broadly aligned to PARPA goals. • The PRGF emphasizes poverty reduction interventions (e.g., pro-poor, and pro- growth government budgets) and measures to improve public resource management and accountability (in line with PARPA objectives on good governance). • The envisaged fiscal stance in programs continues to reflect an “objective” of significant reduction in aid dependence over a relatively short time horizon, but in reality, programs have become more flexible in dealing with aid inflows. • There has been significant streamlining in structural conditionality under the PRGF- supported program. In practice this has meant transferring responsibility for areas vacated by the IMF to the World Bank. Thus, “streamlining” has been interpreted by the staff in terms of the division of labor with the World Bank rather than reducing the burden of aggregate conditionality on the country. • Not much has happened with respect to undertaking social impact analysis of major macroeconomic adjustment and structural reforms. One exception was PSIA of a proposed fuel tax increase, where the policy decision was postponed until the assessment was completed. 16. There was little public discussion of the macroeconomic policy content of the PARPA, and more generally, there continues to be little public discussion of macroeconomic policy issues. Broader participation would be facilitated by the establishment of a government-led macroeconomic working group open to representatives from all stakeholder groups. The IMF can contribute towards broadening participation in the discussion of macroeconomic policies in the country by facilitating wider dissemination and discussion of the analytical work that forms the basis for its policy recommendations, including the work of TA missions. 17. Feedback from the authorities and donors on the role of IMF resident representatives was very positive. However, CSOs indicated that the IMF was invisible to them reflecting its low profile outside official circles. The office of the IMF resident representative needs strengthening to be able to effectively play all the roles expected of it. World Bank effectiveness - 32 - 18. The Bank supported PARPA formulation the way the government wanted, with appropriate support without being intrusive. Preparation of the PARPA was strongly country- driven, with the Bank providing comprehensive informal and formal comments as well as offering relevant analytical inputs. There was also considerable work by the Bank preceding the PARPA process that authorities drew from, including development of sector-wide approaches, a Country Economic Memorandum, the participatory activities conducted in preparation of the 2001 CAS, and preparation of a Public Expenditure Management Review starting in September 2000 that was recognized by government as aiding the costing of sector strategies. The end product of the formulation phase was clearly country-owned, with several areas of continued differing viewpoints between the Bank and government. These included the scope of the strategy, which the Bank feared was too broad, too sector focused, and included public policies which were not core to poverty reduction. On specific issues, Bank staff wanted more focus on the quality of primary education, a stronger separate focus on HIV/AIDS, and changes to the land rights regime. 19. The government expressed satisfaction with its relations with the Bank, with discussions now more linked to the realities of Mozambique. The Bank has broadened its contacts with civil society actors, including the business community, with relations with these stakeholders also notably improved. It is important to note, however, these assessments did not relate the perceived change to the advent of the PARPA process per se, but to a broader improvement over the past five years. Some of this improvement was noted to be a result of the significant decentralization of Bank staff including the country director to Maputo. Donors expressed a more ambivalent opinion of their relations with the Bank, painting a mixed picture of the quality of sectoral dialogue. On the one hand, the Bank is perceived to be more participative and sensitive to other donors, by seeking dialogue and participation in joint donor groups (e.g., agriculture). On the other hand, the Bank is still seen as too driven by Washington-based task managers who occasionally overrule country office staff who participate in regular donor working group meetings (e.g., the social sectors and infrastructure). 20. The evaluation’s findings on the alignment of the Bank’s work with the PARPA initiative include: • Bank lending and non-lending assistance since the PARPA have largely been in line with its priorities. Significant adjustments have been made to the analytical work program, in line with government priorities, including work on decentralization as well as shifts to procurement and financial accountability analyses in line with furthering the governance objectives of the PARPA. However, activities dropped or delayed have resulted in gaps in the Bank’s support for building on important pieces of the PARPA strategy, most notably the critical rural development area. • A new CAS was discussed by the Bank’s Board in November 2003, and the planned assistance is broadly aligned with the PARPA. In addition, the results framework of the CAS is linked to PARPA and PAF goals. - 33 - • The Bank intends to move with other donors towards providing assistance through budget support linked directly to the PRSP, via a Poverty Reduction Support Credit (PRSC), in spite of cautionary signals from its own analytical work regarding the use of government procedures and systems. Fiduciary risks may be partially offset by linking the PRSC directly to performance indicators related to progress in enhancing public expenditure management capacity (and targets in PARPA). While risky, the move to budget support the signals a strong vote of confidence in the ability of Mozambique to implement the PARPA agenda. Preliminary planning for the PRSC has involved discussions with the G-11 group of donors. This new direction, in conjunction with decentralization by the Bank of key staff to the field, has provided the framework for improved coordination with other external partners. • The Bank’s experience with the PARPA process has implications for the way the Bank operates. First, a qualified local presence of the Bank in key strategic areas is highly appreciated by government, national stakeholders and other donors. It is also a sine qua non for continuous participation in policy dialogue processes, which are becoming less dependent on timing priorities of donors and more linked to domestic needs and events. Second, Bank presence can be of special importance not only as an additional voice among many, but also as a voice that may have a balancing, unifying or if needed mediating function in the donor community. Third, the ability of the Bank to play this role is closely linked to the ability to deliver timely and high quality analytical and advisory inputs. It is important that the Bank ensure adequate resources to provide knowledge services. IMF-World Bank collaboration 21. The JSAs for Mozambique have been candid in their assessments; they have pointed out strengths and areas of significant improvement without glossing over weaknesses. However, they have tended to understate the challenges to implementation posed by capacity weaknesses. In terms of their role in providing a road map for future improvements, the JSAs have contained suggestions of areas in need of improvement, and the full PRSP (PARPA) did respond to many of the issues highlighted in the JSA of the I-PRSP. The JSAs do not seem to have contributed much to enhancing partnership; donors saw it as geared almost exclusively to the Executive Boards of the IMF and the World Bank. A number of donors would like to participate in the assessment rather than merely provide input to Bank and Fund staff. 22. Problems associated with the production of the first PARPA progress report reflected weaknesses in monitoring and reporting arrangements, but also raised important issues of temporal alignment between government processes and requirements of external development partners (including the IMF, the World Bank). Assisting the authorities to strengthen the analytical content of the national reporting instruments that are subject to parliament scrutiny would enhance the prospect for closer alignment. 23. The transfer of responsibility from the IMF to the World Bank for structural conditionality in areas that the Bank is expected to take the lead in has been facilitated by the - 34 - existence of Bank adjustment lending. During periods when there has been no such Bank instrument in place, the PRGF has provided the main vehicle for exercising conditionality by the BWIs (e.g., banking sector conditionality during most of 2000-2002). - 35 - NICARAGUA Country background 1. Nicaragua is among the poorest countries in Latin America with a GNI per capita of US$710 in 2000. Poverty remains largely a rural phenomenon, heavily concentrated in the Eastern and Northern parts of the country. Between 1993 and 2001, the share of the population living below the poverty line declined from 50 to 46 percent and extreme poverty fell from 19 to 15 percent, but poverty reduction has been uneven across regions. 2. Nicaragua’s political context is one of a polarized society and political system, and weak institutional capacity and governance. The nature of Nicaragua’s polarized society, a century old phenomenon, and political instability have been a strong obstacle to consensus building, constraining the process of participation and opening the policy debate, and on reaching common grounds for broader ownership. Nicaragua’s political characteristics have contributed to a weak institutional framework, blurring the separation of state powers, weakening the rule of law and the protection of property rights, and leading to governance problems. Moreover, the state has limited capacity to implement and control activities, with a highly fragmented structure. 3. Throughout the 1990s, after the dismal economic situation left by the Sandinista regime, Nicaragua has struggled with major macroeconomic problems, principally unsustainably large fiscal deficits and debt levels. Repeated efforts to consolidate the fiscal situation during successive government proved unsuccessful owing largely to recurrent expenditure overruns. These recurrent failures put Nicaragua several times in extremely vulnerable positions, on the verge of fiscal collapse and balance of payment crises. Hurricane Mitch at the end of 1998 inflicted serious damage to Nicaragua’s infrastructure and disrupted the macroeconomic situation and fiscal consolidation efforts. The Bolaños government has made significant efforts to improve macroeconomic stability. 4. Nicaragua’s macroeconomic vulnerability is one of the main challenges to poverty reduction, including external factors (such as terms of trade shocks and foreign aid dependence). Other major obstacles are the challenges in enhancing non-farm rural activities and in sustaining rural income growth, the inefficient health services, high rate of population growth, and chronic malnutrition. The PRSP process 5. The timetable for implementing the PRSP process was rapid; in a year the government moved from launching the Interim PRSP to reaching a HIPC decision point and completing the PRSP. The Nicaraguan government embarked on the PRSP process because it was a precondition for entering the enhanced HIPC initiative. This underlying motivation appears to have driven many subsequent processes and decisions, including the depth and scope of policy debate, which affected ownership perception across stakeholders. - 36 - 6. The PRSP approach has on balance been relevant for Nicaragua’s national strategy process and its poverty reduction efforts. The approach has addressed important gaps found in past experiences, including through a widening participation efforts, appropriately focusing on a comprehensive vision of poverty, and addressing a sorely lacking results- oriented framework. In this connection, it has represented a more binding commitment than past national strategies. 7. However, the PRSP process has initially showed modest results, which needs to be viewed realistically in the context of political and institutional realities and very difficult starting conditions. The PRSP did not initially strengthen policy processes, while poor public expenditure management has been a hindrance to implementation. The fragmentation of policy making has generated problems for coordination and overview of public investment, and weak implementation capacity at the municipal level has remained an obstacle. 8. The consultation process for the PRSP was a step forward, but differences in views about the purpose and legitimacy of such participation have caused assessments of the participatory process to vary widely. This highlights the inherent difficulties of participation in a country with a highly polarized political system. As a result, broad country ownership of the PRSP has remained limited. Nevertheless, most observers agree that it was a helpful attempt to foster a dialogue among the government, donors, and civil society. Survey results highlight, however, that the area where participation was specially lacking was that of the formulation of the macroeconomic framework. 9. The new government’s desire to modify the PRSP, with the introduction of a new development strategy that sought to change the expenditure composition guided by a greater emphasis on directly productive investment, was marked initially by tensions in the process. The government felt constrained in its ability to pursue a revised strategy, while donors were initially surprised by the initiative, expecting the PRSP to bring a more lasting and resilient policy framework. This illustrates some of the potential tensions between different elements of the PRSP process, notably between the provision of greater policy space to the government and other objectives such as partnership. These tensions were increased in Nicaragua by the poor alignment of the PRSP process with the domestic political cycle and the limited policy debate early in the process on the tradeoffs involved in the growth strategy. The potential tensions between normal political processes and the PRSP approach are illustrated by the fact that, despite the emphasis on greater participation, the issues concerning the growth strategy were not a topic of debate during the general elections. However, the approach did eventually contribute to bringing critical issues in the open, including through the Plan Nacional de Desarrollo (PND), thereby facilitating subsequent work to create a more effective approach to poverty reduction. Even with the considerable imperfections in the process, the fact that the debate has started to take place should be welcomed. - 37 - World Bank effectiveness 10. A number of key findings and lessons emerge from the Bank’s support to the PRSP process thus far. The Bank has provided important support to the PRSP process. • The Bank was closely involved in the PRSP process, providing relevant analytic inputs and advice. These non-financial services have been well aligned to the needs of Nicaragua through the PRSP process, including two completed Poverty Assessments, a PER, and an agriculture growth study. Key on-going analytical work includes a CFAA, a CPAR, and a PSIA. The quality of the technical assistance and analytical advice provided by the Bank is held in high regard by most informants. • The Bank’s CAS is well-aligned with the PRSP and closely follows the main strategic pillars. The assistance strategy details the specific support to each pillar, in recognition of the main constraints to poverty reduction outlined in the PRSP. The move to programmatic lending through the PSAC and planned PRSCs is the most explicit expression of the alignment of Bank support to the PRSP on the ground. The Bank is improving its behavior in terms of promoting the PRSP principles. • According to stakeholders the Bank has become more open to dialogue with the country and is taking steps to increase its responsiveness to civil society, both reported as areas of weakness in terms of country ownership. The Bank is increasing the cohesion of its large country program and is addressing the decentralization of task managers. The agriculture growth study is an example of unprogrammed analytic work being undertaken at the request of the government and demonstrates greater responsiveness on the Bank’s part. The study was also produced in very close cooperation with government officials, and represents a more participative approach to the Bank’s work. • While the Bank may contribute to a technically sound strategy, the durability of a strategy requires both the substance and process dimensions. The Bank could do more in focusing on improving processes in Nicaragua and not just working towards preparation and refinement of a strategy. The process management skills and collaborative attitudes of the Bank need strengthening to allow country ownership to grow, to allow negotiation processes to get to better results, and to achieve the vital balance between technical and political quality in advisory services. The Bank could do more to advance the debate on broad based growth and the trade-offs between growth and poverty-related spending in Nicaragua. • There is widespread perception that the emphasis on social sector spending has been imposed by the Bank. In reality, there is an inherent incentive and bias towards social - 38 - sector spending by government given the link with use of HIPC resources and, rather than an imposition of a “social agenda” by the Bank, the key issue is the absence of a frank discussion of growth issues in Nicaragua. • The growth-social balance is a technically very complex discussion. It is connected to highly contended political issues and to conflictive political processes—but it needs precisely for that reason to become a more public debate in an appropriate forum, based on arguments rather than mutual mistrust. The Bank should promote and participate in this discussion, integrating the process under the PRSP framework. The ongoing Development Policy Review, with its focus on sources of economic growth, may play a positive role in this regard. Involvement of stakeholders during preparation, for example through feedback on drafts or presentations of preliminary findings, and appropriate dissemination are likely to be important factors in this works value-added to the PRSP process. The Bank could play a more aggressive role in supporting government led donor coordination. • The Bank has in the past not played an effective role in coordinating its activities with those of other donors in Nicaragua. There have recently been promising steps taken towards improving the situation on which the Bank can build. These include improved coordination and participation in its analytical work and the emphasis on SWAps in the CAS. • Improved aid coordination will require continuing efforts to clarify own policies, positions and procedures, willingness to cede donor-side leadership to other donors, and, as a particularly important signal in this area, assistance to the government enabling and motivating it to coordinate the support from the World Bank and the IDB. The PRGF and IMF support 11. The Fund’s internal policy formulation process for Nicaragua under the PRGF/PRSP approach was initially not very different from previous practices, although there is some evidence of improvement over time. Macroeconomic stabilization efforts took up most of the attention, and while emphasis on growth and poverty issues has occupied a more prominent place in the internal process, they have not received the consideration that would have been expected under the new approach. Fund technical assistance has generally supported the priorities highlighted in the PRSP, including tax reform and banking crisis resolution. 12. Program negotiations, rather than the PRSP, remained the driving force behind the formulation of the macroeconomic framework because of several factors including: pressing initial conditions and lack of government credibility; a PRSP macroeconomic framework that was quickly superseded by events; and lack of operational guidance on some major policy tradeoffs in the PRSP. Some sensitive but critical issues (i.e., the resolution of the banking - 39 - crisis and its fiscal consequences) were not part of the broader policy debate, reflecting government choice. The focus of the PRGF-supported program on macroeconomic stability issues diverted efforts from its linkage to the PRSP’s growth pillar. Moreover, limited growth analysis prevented the considerations of tradeoffs, including between alternative fiscal consolidation paths and infrastructure investment. 13. The initial design of the PRGF-supported program did not differ significantly from earlier ESAF-supported programs, with a key element of the macroeconomic framework being the path of fiscal consolidation. This was in part because both the government and the IMF staff considered stabilization a priority and made it the main axis of the program. However, more recently there appears to be greater fiscal flexibility, including with regard to the accommodation of potentially higher aid flows. Structural conditionality in the PRGF- supported program was streamlined, focusing on financial system and public sector issues that have major macroeconomic relevance in Nicaragua and have been highlighted in the PRSP as key priorities. But the burden of quantitative performance criteria did not diminish. 14. The policy space for government-led initiatives and flexibility to political considerations in the context of the PRGF-supported program did expand over time—mainly reflecting increased government credibility and a more stabilized macroeconomic environment. This was reflected, for example, in the context of policy choices associated with the approval of the 2003 budget, the passage of tax reform legislation, and the process of asset recovery from liquidated banks. 15. Key components of the PRGF-supported program—the need of fiscal consolidation and macroeconomic stability—have been government owned, but the program has lacked broader ownership. Despite recognition of increased outreach efforts by the Fund, the formulation of macroeconomic policy is still regarded as not being open to a broader policy debate outside of traditional negotiations between the IMF staff and the authorities. To a large extent, this reflected the choice of the government. Opening program negotiations to non-government actors is not realistic, but efforts by both the government and the IMF to broaden the discussions of macroeconomic policy could contribute to an explanation of various policy choices and a better understanding of trade-offs. So far, the PRSP has had little impact in this area, although the use of technical commissions, akin to the one employed to formulate tax reform proposals, may be one approach worth exploring further. 16. The case of Nicaragua illustrates that, when macroeconomic stability issues still require significant attention, PRGF arrangements may only be expected to move gradually to cover in-depth growth and poverty aspects, taking into consideration initial country circumstances. But the issues of how to transit from stabilization to growth and enhanced growth analysis—including consideration of key tradeoffs—would eventually require heightened attention to move towards the key features of the initiative. - 40 - Bank-Fund cooperation and the JSA 17. Joint Staff Assessments were insufficiently candid in some key areas, notably the constraints to participation, but nevertheless played a useful role, identifying country-specific risks to the implementation of the strategy as well as areas where further development of specific policies in the PRSP or policy prioritization was needed. However, the Progress Report did not play the central role that would have been expected in the ongoing debate on the growth strategy and modification of the PRSP, and the associated JSA could have signaled the limited use of the Progress Report in this respect. 18. The World Bank and IMF have collaborated closely since the PRSP, with country offices playing a key role, providing inputs to the respective programs, and participating regularly in meetings with civil society groups in various fora. The area of governance and legal reform, of prominent importance in Nicaragua, is an example where collaboration appears to have been effective, leading to a clearer delineation of responsibilities and conditionality. - 41 - TAJIKISTAN Overview 1. Tajikistan experienced a tumultuous decade following the collapse of the Soviet Union, marked by a protracted civil war and sharp reductions in the population’s incomes and living standards. The civil war and a serious brain drain left their mark on the country’s economic infrastructure and capacity in the public service. Economic reforms gathered momentum in 1997 following a peace accord that marked the end of the war. Real GDP growth has been strong since 1998—albeit from a low base—and a measure of macroeconomic stabilization has been achieved. 2. The PRSP sets a target of reducing the proportion of the population living below the poverty line from 83 percent in 2001 to 75 percent in 2006 and to 60 percent by 2015. The strategy for reducing poverty involves: (i) export-led labor-intensive growth, with increasing reliance on the private sector; (ii) efficient and fair provision of basic social services; (iii) targeted support to the poorest groups; and (iv) efficient governance and improvement in security. 3. The authorities see the PRSP mainly as an instrument for attracting external development assistance. Application of underlying principles 4. The PRSP approach was relevant to Tajikistan, given: weak country-ownership of ongoing economic reforms, the lack of a forum for the participation of a wide range of stakeholders in policy discussions, and weak coordination of external development assistance. 5. The PRSP was prepared through a well-organized government-led process that involved broad participation—including line ministries, scholars, and representatives of civil society organizations (CSOs). Regional participation was a strong feature of the process; scores of workshops and conferences were held across the country to discuss sector notes prepared by Sector Working Groups. A number of other initiatives were undertaken to supplement the official regional consultations: a donor-funded NGO umbrella group “Counterpart Consortium” sought the views of local NGOs and communities on poverty and compiled these in a key document, the “NGO Poverty Assessment Report;” and the World Bank and the Asian Development Bank supported a “Voices of the Poor” study which provided qualitative feedback from the poorest segments of the population. 6. The PRSP process opened up the policy making process to some extent, but there is little familiarity with the final version of the PRSP outside government. Furthermore, the role of the participatory process in implementation and monitoring, and in future iterations of the strategy, has not been defined by the government and is a source of concern among stakeholders. NGOs described their interactions with government as having “petered out” after the formulation of the PRSP. - 42 - 7. The PRSP recognized the multi-dimensional nature of poverty by defining poverty broadly, to encompass income and non-income aspects of poverty (including indicators that measure well-being and access to social services). The PRSP also highlighted differences in poverty incidence across geographical regions. 8. To tackle poverty issues comprehensively, the PRSP proposed a cross-cutting, inter- sectoral approach, whereby, for example, the improvement of infrastructure would need to be linked to the improvement of social services, and vice-versa. The organization of the PRSP work around nine “sector” working groups under the direction of an overall Steering Committee helped to provide a comprehensive view. However, in a number of areas, greater coverage and more detailed discussion could have given the PRSP more cohesion (for example, constraints to private sector development and links between the formal and informal sectors). 9. The PRSP is based on a long-term perspective for poverty reduction, with key targets linked to the MDGs. However, as a guide to operations, it is insufficiently result-oriented, as the vast majority of indicators specified in it relate to processes and policy measures rather than to outcomes. Furthermore, public actions are not well prioritized and there is a disconnect between the PRSP and the government’s public investment program, with neither being well integrated into the government budget process. The institutional arrangements for monitoring the implementation of the PRSP revolve around a PRSP Implementation Unit that became operational in the first half of 2003. The Unit, in conjunction with other agencies, is in the process of developing a more focused set of indicators, including for measuring intermediate outcomes. 10. A weakness of the PRSP which adversely affects its results-orientation is inadequate prioritization. All major issues are covered in some form, but public actions needed to achieve goals are weakly prioritized. By way of better orienting the government budget towards poverty reduction, the PRSP highlights reform of the public expenditure management system as critical. Three key elements are marked as being in need of strengthening: the medium-term budget framework (MTBF), the public investment program (PIP), and the preparation of the annual budget. 11. The PRSP is partnership-oriented in that it has provided a point of reference for donor interactions with the government. However, it has not provided a basis for reconciling what appears to be a substantial difference between government preferences (top priority to infrastructure development) and what donors are willing to finance (spending on social sectors and on safety nets). 12. The size and composition of external aid to Tajikistan has evolved significantly over the last decade, moving from being predominantly humanitarian assistance to becoming mainly development assistance provided by an increasing number of donors. Donors were involved to varying degrees in the preparation of the PRSP, mainly through commenting on sector notes circulated by the sector working groups. On the whole, donors felt that their inputs were taken into account in the preparation of the PRSP. - 43 - 13. There is some evidence of donors aligning their programs to PRSP objectives, but with no clear links to the PRSP policy matrix. In discussions with the evaluation mission, a number of donors noted that the unprioritized nature of the PRSP meant that it was easy for donor programs to be aligned to some PRSP objective or other. Preliminary results 14. The authorities indicated to the evaluation mission that they have been implementing measures specified in the PRSP “policy matrix” to the extent that they can, and cited lack of external financing as the major factor hindering implementation. All the ministries contacted said that their budgets for 2003 and the proposals under formulation for 2004 were based on the PRSP. However, the evaluation mission formed the view that not much has happened by way of integrating the PRSP, the PIP, the MTBF and the government budget. This is an important task since currently, the PRSP and the PIP appears to have different priorities. 15. The government’s attempt to coordinate more effectively donor assistance has led to a proliferation of units with responsibilities that include liaison with donors. This has resulted in overlapping functions and accountability. World Bank support 16. PRSP formulation. There is awareness and appreciation of the Bank’s role in support of PRSP formulation both within government and outside. Bank staff (including local office and sectoral staff) provided useful comments on the draft PRSP document, they assisted the working groups in mounting the workshops and consultations with civil society, and they helped to familiarize key officials about best practices in tackling poverty, through the holding of a series of workshops in Moscow and Budapest. 17. A key part of the Bank assistance came as a result of the Poverty Assessment Report of June 2000, which was a well-argued, analytic document. While this report preceded the PRSP, it formed a cornerstone on which the authorities could base their PRSP proposals and measures. The report established the poverty line benchmarks and it highlighted strategic issues to be tackled in the labor market, agriculture, education, health, social protection, and governance sectors/areas. The report did not cover energy, infrastructure, and private sector development. 18. In assessing sectoral differences, the evaluation mission’s discussions with representatives of the PRSP Working Groups revealed that the Bank’s work on education, poverty analysis, social protection, health, energy, and agriculture/land reform have been the most appreciated as having direct relevance to the PRSP, while private sector development, the financial sector, water, and the transportation sectors are seen as perhaps being relevant, but have had the least input into shaping the PRSP. 19. Alignment of Bank assistance to PRSP priorities. The Tajikistan CAS of February 2003, which was formulated after the PRSP, takes a medium-term perspective of 5-6 years, and as such, represents a good fit with the long-term perspective of the PRSP. The guiding - 44 - principles of the CAS are in line with the four PRSP pillars or elements mentioned above, and this alignment is further reflected in the areas of focus in the Bank’s lending program: a mix of adjustment lending and investment lending in the social sectors (education and health), agriculture, watershed management, banking, private sector support, and municipal infrastructure. The economic and sector work represents a key element of the Bank support to the PRSP. To assist in improving governance, the Bank will focus on public sector management reform, including civil service reform, and the strengthening of the customs and tax departments, and improving the budget process. 20. The ongoing Public Expenditure Review (PER) is a key exercise in providing the prioritization that was missing in the PRSP, especially as regards highlighting the main expenditure items or programs to be supported in the budget. All potential Bank actions/activities are aimed at the same basic problem that was flagged in the PRSP formulation stage: strengthening the weak institutional capacity that exists within Tajikistan. The Bank is increasing its program of analytical work, which has important potential to assist the PRSP implementation phase. All major non-lending products are associated with key focus areas of the PRSP, including the Business Environment and Investment Climate Review, PER, Living Standards and Measurement Survey and Poverty Update, Social Sector Expenditure Review (completed), and Education Sector Review (completed). 21. A key element in the CAS revolves around the community-driven development (CDD) approach. Board discussion of the CAS highlighted the riskiness of this approach, while agreeing that it had a role to play in the country, and stressed the importance of considering the broader growth agenda and sources of future growth. Risks outlined by Board members included the substantial concentration of authority at the central government level, capacity constraints with regard to coordination between the central government and the communities, and limited prior Bank activity in this area. The evaluation mission’s discussions with NGOs and field visits outside Dushanbe confirmed that there are significant challenges and risks to the CDD approach: (i) factoring in capacity constraints at the local level, as well as at the center (national capital and Oblast capitals); (ii) constraints in scaling- up quickly to assist in achieving the PRSP’s time-bound targets; and (iii) meshing with the CDD projects or programs of other donors in the country. 22. Capacity building in PRSP implementation. In 2003 the Bank approved a grant for Provision of Implementation Support and Development of Monitoring Indicators for the PRSP. The grant is channeling technical assistance to the PRSP unit in the Presidential Administration, which has been set up to coordinate the implementation of the PRSP. The World Bank Institute’s selection of Tajikistan as a focus country for its capacity-building programs should allow training and global distance learning opportunities to be delivered to key officials, and to have these activities geared around supporting the PRSP objectives. Regional learning opportunities and seminars at the Bank’s Kazakhstan hub office, and at other ongoing PRSP exercises such as in the Kyrgyz Republic, could also play a valuable role. - 45 - 23. In order to ensure effective support of the PRSP process, the evaluation mission sees five other areas that merit the Bank’s close attention: i) Creating stronger partnerships with other donors on PRSP issues. At the CG meeting in Dushanbe in May 2003, the Bank exercised a direct role in keeping the PRSP on the radar screen. The Bank should continue to use the CG process to help identify shared priorities and policies and continue this process throughout the year with donor/government discussions on policy and implementation issues. ii) Monitoring the effectiveness of Bank support in the PRSP implementation phase. This requires a holistic view: how to integrate the work of the new implementation grant into drawing out the poverty-related linkages of regular project work, and how best to tackle the poverty issues that are embedded in the major ESW studies -- all of this represents a vital part of the Bank’s support role to PRSP implementation. iii) Sequencing the Bank’s work on PSIAs. Government officials noted to the evaluation mission that the PSIAs on energy and land reform had not been carried out, yet conditionalities on tariffs (in natural gas) had been agreed without the benefit of having the PSIA on Energy completed. iv) Designating a focal point on PRSP work in the Bank’s program. The Dushanbe office is short of a specially-appointed or designated person on PRSP issues. v) Interactions with the hub setting. Bank leadership on Tajikistan is exercised through a hub setting in Almaty, Kazakhstan. This places a particular challenge in keeping the triangular communication patterns close and frequent. IMF support 24. IMF staff, especially the resident representative, provided substantial support to the authorities in the preparation of the I-PRSP, but were more removed from the PRSP preparation process. In particular, they did not participate in the government-led consultative process; staff saw this as essentially a domestic stakeholders’ affair. However, the PRSP acknowledged the influence of the IMF on its macroeconomic framework; for example, it indicated that monetary policy would be guided by the PRGF-supported program (then under negotiation). 25. The inclusion of “macroeconomic management” as one of the nine “sector” working groups ensured that there was some discussion of macroeconomic issues in the PRSP consultation process, including in the NGO Poverty Assessment Report. Several of the policy recommendations in the NGO Report indicate a state-interventionist rather than market-based approach to economic management, but others emphasized the importance of strengthened institutions and greater transparency as essential elements of a sound macroeconomic framework. The issues raised demonstrate the potentially useful role IMF staff can play in raising broader awareness of macroeconomic issues, and in promoting broader participation in policy discussions. - 46 - 26. We found the IMF’s technical assistance to be broadly consistent with priorities and objectives set out in the PRSP. In recent years, the areas covered by this assistance have included: bank restructuring, tax policy and administration, external debt management, monetary operations, macroeconomic statistics, and economic legislation. Much of the technical assistance has been closely coordinated with program activities; for example, providing key elements of structural conditionality and/or aimed at meeting capacity gaps revealed by poor implementation of program commitments. 27. Overall, we found evidence that the internal IMF policy formulation and review process has adapted to the PRSP/PRGF approach in some areas—such as allowing room for alternative options for meeting some program objectives and in paying attention to guidelines on streamlining structural conditionality. However in other areas, notably related to fiscal flexibility and to PSIAs, there was little evidence of movement towards a markedly “different way of doing business” suggested by the PRSP/PRGF approach. 28. Participation and country ownership. Participation in the formulation of Tajikistan’s IMF-supported programs continues to be limited to a relatively small circle of government and central bank officials. Typically, the discussions revolve around a draft “memorandum of economic and financial policies” (MEFP) prepared by IMF staff. In our view, this approach is not conducive to encouraging government ownership and runs the risk of perpetuating dependence on IMF staff. We recommend that the staff and the authorities establish a time table for turning over to the authorities increasing responsibility for the preparation of the first draft of the MEFP. We formed the view that the PRGF-supported program is playing a crucial role in maintaining a measure of discipline in Tajikistan’s public finances, but we also formed the strong impression that the authorities have gone along with important aspects of the PRGF-supported program that they do not fully agree with, mainly because they perceive that to do otherwise would reduce development assistance to the country. 29. Fiscal policies. There has been a modest increase in government spending on social services in recent years and this trend is projected to continue under the current PRGF- supported program. The program also provided for some compensation to low-income households to mitigate the impact of tariff increases in the energy sector, and has recently introduced a “floor” on government spending as part of conditionality under the program. With respect to pro-growth orientation, the PRGF emphasizes governance measures— especially to stop public officials’ interference in the affairs of private enterprises—more than fiscal measures. 30. The key feature “appropriate fiscal flexibility” is intended to allow room for programs to be more accommodating of increased public spending (linked to poverty reduction goals) that can be financed in a non-inflationary way (usually through concessional external financing). Although the program’s formal conditionality does not cover concessional borrowing, there is an understanding that external borrowing to finance the PIP will be limited to no more than 3 percent of GDP. IMF staff have justified this limit on grounds of debt and fiscal sustainability, and absorptive capacity constraints. However, a number of senior government officials question the rationale for the limit, viewing it as arbitrary. In our - 47 - view, justification of the 3 percent of GDP limit needs to be more firmly grounded in a quantitative sustainability framework and also in growth analysis. 31. Structural conditionality. The total number of structural conditions has fallen, and its composition has changed slightly. In particular, conditionality on public enterprise restructuring/privatization has been dropped, but conditionality remains on sectoral issues deemed to have macro-relevance, especially when other IFIs—especially the World Bank— have no lending operation in the sector. Although staff have cited capacity constraints as a reason for streamlining structural conditionality, in practice “streamlining” has been interpreted to mean clearer division of labor vis-à-vis the World Bank. 32. Social impact analysis. There has not yet been a PSIA on any aspect of Tajikistan’s IMF-supported program, although the programs recognized that measures aimed at eliminating quasi-fiscal deficits in the energy sector would have severe adverse effects on the well-being of a large part of the population. IMF-World Bank collaboration 33. Tajikistan was a pilot case for an “enhanced Bank-Fund collaboration” project that was initiated in late 1998. With the advent of the PRSP approach, it was expected that staff would build on the progress achieved under the pilot project. 34. In general, the staffs have tried—not always successfully—to coordinate their work programs. Areas in which collaboration has been very effective include support to the government in ensuring that the priority status of the education sector in the PRSP is translated into appropriate budget allocations (based on World Bank analytical work). Collaboration has also been good on issues related to the agriculture sector and land reform, though the IMF has now pulled back from these areas. Collaboration on energy sector issues appears to have improved over time, but could be closer. 35. Areas in which collaboration has either not been effective or has been slow in producing results include PSIAs and expenditure policy issues. Regarding current issues that have an important bearing on poverty, IMF staff indicated that they looked forward to receiving the report on the Public Expenditure Review (PER) being undertaken by the World Bank staff. Also, the PSIA for energy was delayed, which meant that it will be undertaken after a number of important policy measures have been implemented, rather than before. 36. The staffs of the IMF and the World Bank maintain contacts with staff of other international financial institutions, especially the AsDB which is a major source of concessional loans to Tajikistan. An economist from the AsDB routinely participates in IMF missions as an observer; both sides indicated to the evaluation team that this arrangement has proved to be mutually beneficial. In our view, closer collaboration among the three institutions would enhance the effectiveness of assistance being provided to Tajikistan by each in at least three areas: (i) support for the development of a well functioning medium- term expenditure framework that integrates the PIP into the budget process (requiring closer coordination of the various strands of the assistance that each institution is already providing - 48 - to the budget process and/or the PIP); (ii) energy sector reforms (AsDB has taken the lead in assisting the authorities but its operations and conditionalities have not addressed the issue of quasi-fiscal activities, which IMF and World Bank staffs have highlighted as significant problems); (iii) PSIAs—the three institutions can collaborate in a manner that complements each other’s operations. - 49 - TANZANIA The PRSP process 1. The initial formulation of the PRSP took place within an extremely compressed timetable, under pressure to reach the HIPC completion point. Civil society inputs in the preparation of the PRSP were limited and hence had little influence on the initial policy content of the PRSP. The Ministry of Finance officials indicated that at the beginning they were frustrated over the lack of clarity of what was expected from a PRSP. A country-led and comprehensive poverty reduction strategy, set in the context of the long-term Vision 2025, already existed in the form of the National Poverty Eradication Strategy (NPES). With respect to partnerships, the Helleiner report had initiated a fruitful dialogue with the country’s external partners and the Tanzania Assistance Strategy (TAS)—aimed at coordinating donor assistance—was under formulation. So what potential did the PRSP have when it was introduced? A review of the existing strategy documents and the evolution of the process, as evidenced from the preparatory phase and the subsequent three years of implementation, suggests that the PRSP process has indeed added value in some important areas, although challenges remain. 2. A key contribution of the process has been the opening up of a dialogue between the government and civil society. Civil society is now more actively engaged in national policy- making through channels such as participation in the public expenditure review (PER). The PER has broadened to become the main avenue for public policy debate and strategy formulation in Tanzania. The institutionalization of the policy dialogue, through the PER process, where the World Bank played a critical supporting role, has been crucial—although it is not possible to attribute this development to the PRSP approach per se. The PER process provided a forum for follow up of the policy debate which was very important given initial gaps in the poverty reduction strategy. It is broadly participatory, government-led and encourages domestic and external partnership. It also facilitates a multi-sectoral results- orientation of the PRSP via links to the budget and assessments of the poverty focus of expenditure. 3. One area where stakeholder involvement has been limited is on macro policy issues— involving mainly the government, donors, and the World Bank and the IMF. While this is partly due to the weak technical capacity among civil society groups, it also reflects limited success by the IMF in framing key macro policy issues in a manner digestible to a wider audience and in ensuring that all major macro issues were included in that debate. IMF staff were not clear as to what their role should be if some key policy issues were not emerging and were reluctant to intervene too actively in what was meant to be a government-led domestic debate. 4. The PRSP has enhanced the mobilization of resources and focused them on priority sectors—with public expenditure reviews as major inputs. During the PRSP process the shares of priority sectors in expenditures have risen steadily, although PER analyses have shown that spending within priority sector needs to be better targeted to the poor. The PRS - 50 - process has substantively enhanced national processes for poverty monitoring and a master plan now guides all monitoring activities. But the feedback between monitoring of results and policy actions remains in need of improvements. 5. In retrospect, the issue of poor domestic capacity has not been adequately considered given the broad agenda of reform—both within the process of implementing the PRSP and more broadly as part of ongoing government reforms—particularly management and coordination at the sub-national level. Lack of capacities has led to serious delays, and may ultimately lead to inadequate outcomes. 6. The PRSP process has been partnership-oriented. External partners view the PRSP as the guiding framework for development assistance and have attempted to streamline their processes accordingly. The BWIs have also aligned their processes and their behavior in the field, along the lines of the PRSP approach, although in the view of domestic stakeholders more remains to be done. Two issues of concern on partnerships are the still low levels of involvement in the process of the business community and the continuing high transactions costs for the government of numerous donor missions, some of which are still not well coordinated. 7. In the initial phase of its implementation, the PRSP has been mainly focused on improving service delivery, particularly in education and health, by increasing social sector expenditures. Several stakeholders also noted that the poverty diagnosis of the PRSP was not comprehensive enough, with a bias towards non-income aspects. The income-oriented elements of a broader growth and poverty reduction strategy were delayed by the late completion of sector specific strategies, notably for agriculture and rural development. This also reflects the legacy of the initial focus on social expenditures of the HIPC process. The completion of sector strategies and action plans in later years and the more elaborate PRSP progress reports have helped to expand on and balance the implementation of the strategy. 8. The ownership of the PRSP is strongest at the level of top leadership and the Ministry of Finance, followed by priority sector ministries, but is found to be much lower in non- priority central ministries. Among civil society and local government, ownership has been mixed, being a function of their involvement in the process, with Dar-based groups more favorable to the process than those outside. Parliament has been largely outside the process, and its ownership is low. 9. The main conclusions from Tanzania’s PRS experience are fourfold. First, it is the process, not the document itself that matters. Despite the compromised quality of the initial participatory process, the insufficient poverty diagnostics and lack of a broader growth strategy in the initial PRSP, three years after the launch of the PRSP policy-makers have been able to flesh out the strategy and improve poverty monitoring mechanisms. Second, while there has been commitment to the strategy at the top leadership level, there is by no means a full consensus on many key issues among all stakeholders involved in the exercise. The sustainability of the PRS process depends upon a broader level of participation by Parliament, non-priority sector ministries, and civil society, including in the macroeconomic - 51 - policy framework. Third, capacity constraints and lack of clarity about the role of the BWIs if major policy issues are not emerging in the debate are currently hindering such participation and need to be addressed. It is also clear that wide differences of view about the scope and goals of participation remain. Fourth, the PRSP process requires closer World Bank and IMF collaboration, including a clearer delineation, as part of the PRSP framework, of what analytical inputs should be delivered. IMF role and the PRGF 10. The consultation processes under the PRSP did, over time, have a significant impact on the content of macroeconomic policies, and the PRGF-supported framework did show considerable flexibility. In this sense, there was eventually an “alignment” with the key objectives of the PRSP although the process was not driven by preparation of the original PRSP. In assessing how the PRSP process affected the scope of policy choices available to the government and how the PRGF adapted to these policy choices, we examined in depth four policies in the macroeconomic area. They included two cases, the PRGF programming framework and cash budgeting, where a broader debate in the PER working groups led to a more flexible fiscal stance and increased priority expenditure better insulated from temporary budget shocks; one case, trade reform, where the political nature of the problem and the lack of a domestic debate on the lines of the PER evaded easy resolution; and a last case, fiscal federalism, which, though considered a priority by the government and with clear relevance for the PRSP, was approached very cautiously by IMF staff, in light of the fairly limited local capacities and the IMF’s own experience from other countries, and thus the alignment of the PRGF remained incomplete. 11. Lack of capacities within government led to delays in policy formulation and implementation, although the process of interaction and debate finally led to the convergence of views on key macroeconomic issues. However, effective networking by key government officials, reinforced by donor concerns that the initial macroeconomic framework did not sufficiently accommodate increased aid flows, helped bring onboard technical expertise from local consultants, academics engaged in local research-oriented NGOs, and others from abroad. Their participation in the PER macro working group and workshops organized by the government, as well as technical contributions to policy discussions helped move the process forward. In this sense, the substantive revisions that were made in the original PRGF program design reflected an effective contribution of the institutional framework set up for consultations between stakeholders and the use of outside expertise to strengthen capacity. However, this participatory process was incomplete since it largely involved government, the BWIs, and bilateral donors, with civil society stakeholders largely on the sidelines. 12. With regard to internal IMF policy formulation and review processes, the overall conclusion is that the IMF does seem to have evolved towards better discussion of the PRGF- PRSP linkages, as well as poverty issues more broadly, and did show flexibility in modifying the program design. However, while the program was able to accommodate higher aid inflows and did modify the targeted fiscal path, in light of the slower-than-expected-response of private sector investment and demand for credit, the assumptions on growth remained - 52 - largely exogenous to program design. Moreover, PSIA still did not play a significant role as a guide to policy formulation. World Bank effectiveness 13. The Bank played a facilitative role, providing technical advice to the government during the short formulation period of the PRS as well as in identifying key process issues, which influenced the sectoral contents of the strategy and helped promote a more participatory process. Stakeholders were critical of the Bank’s “intrusive” influence in the initial stages of the formulation of the strategy, but acknowledged that this keenness was tempered by the country office. Overall, the principal drafters of the PRSP described the formulation as a government-led process with the BWIs not influencing the substance of the final strategy. Government officials, however, did complain about having to adhere to “formatting requirements” for the BWIs. 14. Civil society members, on the other hand, were critical that during formulation the Bank did not exert more influence on the government to increase their participation. But they were cognizant of the dilemma faced by the Bank in urging the government to expand the consultative process while respecting the “country-led” principle. Further, the Bank as well as the civil society was reluctant to hold up Completion Point under HIPC to demand a more extensive consultative process. During implementation, the Bank has promoted the participation of civil society within the national PRS processes with encouraging results. Civil society participation has been a feature of the PER process and is being further strengthened year by year. 15. An area where the Bank could have been better prepared for the PRSP was in assisting Tanzania to conduct a Household Budget Survey prior to 2000. During implementation of the PRS, however, Bank staff have provided substantial technical advice in analyzing the results of the Household Budget Survey and in establishing the Poverty Monitoring System. The Bank’s technical expertise and sectoral inputs in education and health and via the PER process have supported the 1st and 2nd Annual Progress Reports and have been greatly appreciated. The PER process in particular has been well recognized for its support of the PRSP and its principles, and the Bank has played a leading role in promoting and expanding the PER from an external technical assessment to a country-led participatory process involving a wide range of stakeholders. Besides providing secretariat services to the bi-weekly PER meetings until the end of 2003, Bank staff have fostered the momentum of the process and have continued to provide high-quality technical expertise. 16. In other areas of the Bank’s analytical support, some key shortcomings came to light in the formulation of the PRSP, and some of these remain. There are continuing knowledge gaps between the micro-macro linkages, specifically the micro-economic social and poverty impacts of macro-economic and structural policies, analysis on the sources of growth and the poverty impact of alternative policy and strategy options. At the sectoral level, the depth and breadth of the Bank’s analytical work was uneven prior to the formulation of the PRSP. - 53 - Work in agriculture, for example, should have started much sooner and the Bank could have been more assertive in promoting the growth agenda in the PER/MTEF processes. 17. The Bank’s lending portfolio in the last three years has been well aligned with the PRSP, and the recently introduced PRSC signals a further shift towards alignment with country processes. The PRSC1 is programmatic, aligned with the PRS policy matrix and has much less conditionality than previous policy-based lending instruments. 18. The Bank’s overall behavior vis a vis respecting the “country-led” principle is widely recognized as having improved significantly. The Bank was already moving in this direction but the PRSP has helped moved it much further and faster. Interviewees described a “sea change” in the Bank’s behavior, and have cited the strong role of the country office in promoting this change. But there is still room for improvement, with some national stakeholders feeling that the “Washington consensus” has become a “Washington-Dar” consensus. Government officials and NGOs also feel that some sector teams are more resistant to the principle of country ownership and often fail to adequately consider alternative viewpoints. 19. The Bank has continued to promote the “partnership” principle. The Bank’s proactive role in the participatory PER process, the alignment of the PRSC within the framework of the PRBS, and its engagement with the DAC, are important steps that have enhanced the collaboration of donors in Tanzania. There are still complaints about the Bank’s coordination at the sector level which relate to the large numbers of missions that regularly arrive from the Bank. At the same time stakeholders have acknowledged that the complaints at the sectoral level are strongly correlated with capacity problems within ministries, an area where the Bank’s and donors' technical assistance has not been successful in addressing. World Bank-IMF collaboration 20. The PRSP process has fostered greater coordination amongst the World Bank and IMF teams working on Tanzania, although there remain important gaps—including the setting of priorities for timely delivery of PSIA. On conditionality, the BWIs have narrowed their focus to areas of their specialty, but this has not resulted in substantial reductions in aggregate conditionality. The Joint Staff Assessments of the PRSP and the Progress Reports have been important in providing a forward-looking perspective on the process but have been fairly uncritical, especially with respect to implementation capacity at the local level. As commentaries on a multi-stakeholder process, however, questions have been raised on their limited inclusiveness. - 54 - VIETNAM 1. The Government of Vietnam (GoV) initiated the preparation of an I-PRSP in mid- 2000 which coincided with an intense planning phase in the country’s policy cycle leading to the production of the current 10-year Socio-Economic Development Strategy and 5-year Socio-Economic Development Plan (SEDPs). The I-PRSP was presented to the Executive Boards of the Bank and the IMF in March 2001 along with a request for a 3-year PRGF- supported program and the first World Bank-supported Poverty Reduction Support Credit (PRSC). The full CPRGS (as the PRSP is known in Vietnam) was finalized just over a year later, around the time that the second review under the PRGF-supported program was completed. 2. The CPRGS was produced by the Ministry of Planning and Investment (MOPIC) in association with relevant ministries. Actual drafting was done by a 52-member committee from 16 ministries and agencies. Organization of the supporting participatory process was centered around the Poverty Task Force (PTF) composed of government ministries, multilateral and bilateral donors, and international and local NGOs. The PTF had emerged from the poverty working group, a forum set up by the government to facilitate collaboration with the donor community and NGOs on poverty issues. 3. The content of the CPRGS is widely considered to be domestically owned. The authorities maintained a firm hand throughout the drafting process, much of which was done in Vietnamese with only selected drafts translated into English. The strategy’s emphasis on human capital and poverty reduction relative to other growth considerations reflects a perception that the donor community was the target audience for the document. A notable exception is the pattern of resource allocation for the public investment plan which does not appear linked to the poverty reduction orientation of the CPRGS, and which contains a stronger emphasis on infrastructure and industrialization. With the encouragement of Japanese donors, the authorities have since amended the CPRGS to include a chapter on the contribution of large-scale infrastructure to poverty reduction. 4. The participatory process in the formulation of the CPRGS was less “country driven” with international partners (the World Bank and a group of INGOs, in particular) initially assuming a leading role in the organization and financing of public consultations. To some degree, this reflected a lack of experience with, and initial discomfort on the part of, the authorities in conducting participatory exercises of the type called for under the PRSP Initiative. Government engagement with the participatory process intensified during the formulation process, but the true extent of the authorities’ ownership of, and commitment to, a participatory approach to policy making will only be tested once they take the driver’s seat in future participatory consultations. 5. Particularly absent from the participatory process were discussions of the macroeconomic framework and many of the associated structural reforms underpinning the strategy. There were numerous possible reasons for this, including a lack of interest and/or expertise in discussing the macroeconomic framework on the part of organizers of the - 55 - participatory exercises, the absence of a systematic effort on the part of the IMF to encourage public discussion of these issues, and the perceived lack of pressing macroeconomic problems that would warranted the allocation of scarce time at participatory workshops. 6. The relationship of the CPRGS to the existing indigenous planning processes and documents (SEDPs) remains unclear. Donors and IFIs look to the CPRGS as the key policy document while the authorities consider this to be the role of the SEDPs. The authorities have characterized the CPRGS as an “action plan” for implementing their SEDPs. However, the lack of clear priorities among competing demands and the absence of a transparent medium-term expenditure framework (or indeed a transparent budget planning process) in which to situate the SEDPs detracts from its operational value. 7. More generally, and despite proactive outreach on the part of the IMF’s Senior Resident Representative, there was a lack of clarity with respect to the role IMF staff were expected to play in the formulation of the CPRGS. At the same time, there was broad-based interest in the IMF playing a greater role in the policy debate (i.e., extending beyond traditional official interlocutors). While not a problem in Vietnam, it was felt that this would require IMF staff to more routinely develop their capacity to discuss complex macroeconomic issues with non-economists. Furthermore, it is unlikely that the current limited resources of the resident representative office would permit meaningful and sustained participation in a broader policy dialogue. 8. There has been some increase in donor coordination, most notably with the formation of a Like-Minded Donors Group, some of which are involved in the co-financing of PRSCs. However, there exists scope to improve coordination and some major donors are (for a range of reasons) reluctant to adapt their programs to align with the strategy. That said, the breadth and generality of the CPRGS has made it relatively easy to justify most pre- existing donor activities and therefore has not constrained donors in pursuing pre-existing priorities. 9. On balance, the PRSP (CPRGS) has made a positive contribution to the policy- making process in Vietnam. It is generally seen as having helped improve inter-ministerial coordination and the increased participation promoted transparency. The CPRGS has led to improvements in the mandate and capacity of information-producing entities such as the General Statistics Office. However, the monitoring and evaluation framework contained in the CPRGS appears to have been excessively ambitious and would have benefited from streamlining and priority setting to better reflect administrative capacity. With respect to the adaptation of IMF-supported programs under the PRGF to reflect PRSP principles and the “key features” of PRGF-supported programs, performance was mixed. 10. Alignment of polices in the PRGF-supported program with those of the CPRGS was only partial, although it has improved over time. This was particularly the case with respect to SOE reform where vagueness about the authorities’ priorities initially masked differences - 56 - between PRGF program commitments and the authorities’ desire for a substantial and continuing role for the public sector in the economy. While there existed a range of opinion among the authorities, the decision to maintain a substantial public sector role was based on a broadly-owned government policy derived from lengthy internal party discussions. While the IMF eventually went along with this approach, it remained concerned about the potential long-term fiscal consequences of large SOEs and tensions with the authorities in this area were not fully resolved. 11. At the level of the macroeconomic framework, there were persistent differences of view between staff and the authorities on GDP measurement and macroeconomic prospects. This resulted in two separate medium-term forecasts, with the authorities’ forecast underpinning the PRSP and that of staff underpinning the PRGF-supported program. However, the difference of view was transparently reflected in IMF staff reports and the CPRGS. 12. While program conditionality has become more focussed on the IMF’s traditional areas of expertise, there was little evidence of an actual reduction in the number of structural conditions. This is surprising in light of the widely-acknowledged strong government ownership of the announced strategy, but appears to have reflected IMF staff’s concerns with the strength of the political commitment in areas such as state-owned commercial bank reform as well as the adequacy of the agreed-to reform agenda. 13. The more defined division of labour between the Bank and the IMF in areas of overlapping interest complicated the design of program conditionality. The example of Bank- Fund conditionality on SOE reform illustrates two messages. First, the debate on the overall policy direction needs to take place in the context of the PRSP or equivalent strategy document; Fund conditionality that is not reasonably well aligned with this domestically owned strategy is unlikely to be effective. Second, in areas where the Bank has lead responsibility, it may well have its own priorities for analysis and conditionality design which may create “principal-agent” difficulties in setting conditions. This was the case in Vietnam where the focus of Bank lending operations in the SOE sector was on improving efficiency and privatizing smaller state enterprises, in line with the Government’s gradualist reform strategy. In this case, the effectiveness of IMF conditionality may have been further undermined by attempts by IMF staff to use quantitative conditionality on credit growth, which—while perhaps justified on its own merits in pursuit of macroeconomic objectives— was not well suited for pursing structural reforms in the SOE sector. 14. Perhaps reflecting the authorities’ good track record in conducting fiscal policy, significant fiscal flexibility was integrated into the PRGF-supported program. Higher levels of spending in priority social sectors were accommodated and these were not constrained by overly pessimistic projections of ODA flows. The authorities had significant leeway on the manner in which they could use higher-than-expected revenues. Perhaps most notable was the absence from the PRGF-supported program of quantitative conditionality (i.e., performance criteria or benchmarks) on major fiscal variables. - 57 - 15. Under the PRGF-supported program, budgets became somewhat more “pro poor,” as total poverty-reducing spending increased both as a share of GDP and of total expenditure. However, it is not clear the extent to which this trend can be attributed to the PRGF- supported program since it began well before the adoption of the program. The progressivity of the tax system was not addressed as part of the PRGF-supported program although there are plans to undertake PSIA to assess the impact of future tax reforms. 16. Some effort was made to assess—a priori—the impact of proposed program measures on the poor and most vulnerable. Of particular note are efforts to establish and finance a social safety for workers displaced by SOE reform. Preliminary work was undertaken to assess the social and poverty impact of trade liberalization but the analysis was of only limited value in informing policy design. Beyond these examples, PSIA did not figure prominently as an input to PRGF program design, partly reflecting IMF staff’s sense that proposed reforms would not have a major negative impact on the poor. Looking forward, an agenda for PSIA work has been endorsed, including with respect to trade and tax reforms. 17. There have been some improvements in Bank-Fund collaboration, most notably at the level of information sharing and as a result of the authorities’ insistence on joint missions. However, further improvement may be constrained by tensions between the degree of decentralization in the respective institutional structures of the two institutions. 18. At a more fundamental level, the case of Vietnam raises important questions about the value-added of an IMF-supported program in a low-income country experiencing macroeconomic stability and no pressing financing gap. While program financing was clearly appreciated, it is not obvious that a PRGF-supported program with higher-then- average access was the most appropriate mechanism to support sound macroeconomic performance. Most donors appear to have valued primarily the “signalling” role that the program provided, but discussions with bilateral donor representatives suggested that the absence of such a program would not have significantly affected their aid decisions. - 58 - ANNEX GUIDELINES FOR COUNTRY CASE STUDIES 1. The Operations Evaluation Department (OED) of the World Bank is undertaking an evaluation of the Poverty Reduction Strategy (PRS) process. The objective of the review is to assess progress of the PRS process toward meeting the challenge of poverty reduction and to assess the World Bank’s role in support of the process, with a view to informing, and where necessary, making recommendations to strengthen the implementation of the Initiative and to increase the effectiveness of the World Bank’s support. The evaluation will be conducted in collaboration with the IMF’s Independent Evaluation Office (IEO), who are carrying out a parallel evaluation of the PRSP/PRGF process. Together these evaluations will provide an independent assessment of BWI effectiveness in support of the PRS process to the Boards of the two institutions. The issues that the evaluations will cover are discussed in detail in the Issues Paper from IEO (http://www.imf.org/external/np/ieo/2002/prsp/index.htm) and Approach Paper from OED (http://www.worldbank.org/oed/prsp). 2. This note proposes guidelines for conducting country case studies answering the evaluative questions laid out in the Two Approach/Issues Papers. It is based heavily on related work between OED and IEO as primary input for both evaluations. While the individual evaluations will each focus on institutional performance, the OED and IEO teams will conduct joint desk and field work in overlapping country case studies. It is expected that collaboration between the units will result in a single country case study report for each of the joint case studies.14 Approach 3. Given the country-specific nature of the PRS process, case study analysis will be used as a key source of evaluative material. The methodology will combine two types of case studies identified in the evaluation literature, “program implementation” (primarily explanatory) and “illustrative” (mainly descriptive).15 The case studies will be “program implementation” case studies as they intend to investigate the implementation of the PRS process in each country relative to its underpinnings. They will also be illustrative in nature as they will add in-depth examples to other sources of evaluative material including external evaluations, thematic studies, cross-country analysis and survey work. 4. To capture the diversity of country experiences with the PRSP Initiative, OED plans to conduct eight case studies for the evaluation, four of which will be conducted jointly with 14 These guidelines refer to the expected coverage in the case studies undertaken jointly by the IEO and OED. Coverage in case studies conducted by either the OED or the IEO only will not address specific issues concerning, respectively, the effectiveness of the IMF or the World Bank. 15 See Datta, Lois-Ellin, United States General Accounting Office, “Case Study Evaluations”, GAO/PEMD-91-10.1.9, November 1990. Also Morra, Linda G. & Friedlander, Amy C., OED, “Case Study Evaluations” (available on OED’s website.) - 59 - ANNEX IEO. IEO is planning to conduct another two case studies independently from OED. In total, therefore, OED and IEO will perform an in-depth investigation of 10 of the 23 countries with completed PRSPs through December 2002. 5. A range of criteria were considered in selecting the PRSP countries for case studies: geographic balance; HIPC and non-HIPC; range of country conditions and PRSP experience; and both mature and recent programs. Country selection also took into account the country coverage of previous or on-going work.16 The following countries were selected based on the criteria set out above: Mozambique, Nicaragua, Tajikistan, and Tanzania (joint OED-IEO); Albania, Cambodia, Ethiopia, and Mauritania (OED); and Guinea and Vietnam (IEO). Details on Methodology 6. The following guidelines define the basic parameters for the case studies, and can be further tailored to reflect country characteristics. The guidelines have been informed by experience from lead missions in Albania (OED) and Tanzania (joint OED-IEO). 7. A local consultant will be sought to assist in each of the case studies.17 The local consultant will be expected to contribute to the team in line with their level of expertise and knowledge. This is expected to yield a range of local consultant assistance across the case studies, from assisting in scheduling meetings and providing local knowledge, to preparing background papers and briefs. Local consultants will also be used to coordinate implementation of the country stakeholder survey (see below). 8. Key informant interviews will be held in both Headquarters and in the field. Interviews will be conducted with relevant World Bank and IMF staff who have been associated with the country team and/or the PRS process. This will be followed by in-country consultations to be conducted for a period of at least two weeks. These consultations are the primary input for the fieldwork and will be targeted at a broad range of stakeholders including; key ministries (central and line), key government units involved in the PRS process, local governments, bilateral and multilateral donors, local and international NGOs, private sector representatives, media, and, wherever possible, Parliamentarians. Interview modalities will vary and may include group consultations. Finally, an exit workshop (see below) will be held which will target stakeholders consulted during the mission. 9. In-country field visits outside of capital will be conducted in each case country, unless not warranted, or not possible given country conditions. Field visits were conducted in both Albania and Tanzania and provided considerable value-added. The modalities of field visits will vary from country to country. 16 For example, the extensive country studies by the Multi-Donor Evaluation of the CDF covering countries with completed PRSPs (Uganda, Bolivia, Vietnam, and Burkina Faso) contain a significant amount of information on the PRSP process in these countries. 17 No local consultant was used in Cambodia and Vietnam. - 60 - ANNEX 10. An informal exit workshop will be held in each case country at the conclusion of the mission. The purpose of the workshop is to seek feedback on the emerging findings of the case study from a representative group of stakeholders, including stakeholders involved in the interviewing. 11. A stakeholder survey jointly designed by OED and IEO is being conducted in each country to provide input for both the OED and IEO studies. The survey will supplement information gained from key informant interviews and will target each category of stakeholder involved in the PRS process. Coverage will include the central government ministries (Planning and Economy/Finance) and Central Bank, line ministries, donors, NGOs, and representatives of the private sector and other civil society groups (e.g., trade unions, labor unions, academia, media, and faith organizations).The goal of the survey is to obtain institutional responses (not individual views) on the PRSP and Bank and IMF performance in supporting the process. As a result, the criteria for respondents includes good knowledge of the PRS process and familiarity with their respective institution’s positions on specific issues. A core set of stakeholders has been identified which each country survey must target, including key ministries involved in the PRS process, but outside of these parameters, the specific coverage and number of responses is likely to vary from country to country. Local consultants will be used to administer surveys in each case country further enhancing the independence of the exercise. 12. Case studies will use multiple methods for data collection, including reviews of relevant documents and databases, in-depth interviews of key informants both in Washington D.C. and in the country, and an in-country workshop. Relevant documents may include official government data and reports, Bank and Fund analytical and advisory work, related policy documents, staff reports to the Boards of the IMF and World Bank, internal documentation, prior OED and IEO evaluations, external evaluations and materials from research institutes and civil society. 13. An extensive analysis of the data will be carried out using the technique of triangulation in which the reliability of findings will be developed through multiple data sources and findings will be validated through agreement across multiple types of data. An example would be assessing process changes in public expenditure management in a country. The country’s Public Expenditure Review may yield an initial finding which could then be verified with multiple data sources e.g. interviews with Bank staff in the regional Public Expenditure thematic group, officials in the Ministry of Finance, and civil society. Economic data on budget execution could be cross-checked against budget laws and qualitative data on disbursements and procurement, i.e. data would be validated across multiple types of data. Likewise, to assess the realism of macroeconomic projections made by the IMF, one may assess the extent of deviations from recent trends, compare projections with outcomes, and cross-check that against views expressed at the time by the authorities, review departments and the World Bank. - 61 - ANNEX Phases and Key Deliverables 14. Phase 1—Desk Review: A desk review will be conducted in advance of field work in each of the case countries. The desk review will analyze and summarize main findings from available research materials and collate the necessary data. A desk brief will be prepared following a standard outline to provide an initial basis of information for the country study. 15. Phase 2—Washington-Based Interviews and Mission Planning: Phase 2 will build on the desk review through interviews with Bank and Fund staff. Key persons for interviewing in the field will be identified, evaluative questions will be tailored around the emerging evaluative hypotheses, and arrangements for the workshop and field interviews will be finalized. 16. Phase 3—Country Visit: The field work will investigate the key evaluative questions in-country and seek to address open issues. Structured interviews with key informants and an in-country workshop will form the basis for consultations. As noted above, an in-country stakeholder survey will be applied after field work has been completed. 17. Phase 4—Final Washington-Based Analysis: In Phase 4, the case study team will agree on the main findings and supportive evidence, submit them to a new round of triangulation with HQ-based staff from the Fund and the Bank, and draft a report. The case study teams may present their findings to key internal stakeholders at a World Bank or IMF brown bag lunch. The views of the authorities on the draft final report will be sought. Final Report 18. In joint case study countries, OED and IEO will coordinate to produce a single report as a key input to both evaluations. Some sections of the report will be under the exclusive responsibility of either IEO or OED, while others will convey a joint assessment. The suggested outline for case study reports is as follows, although the format may differ in those reports that are not joint: I) Country Background: • Poverty, political context, economic policy and performance, national strategies and their formulation. II) The PRSP Process: • Relevance and value added of the PRSP approach compared to past modalities for strategy formulation and donor assistance • Application of underlying principles and assessment of what has changed • Preliminary evidence on process results including alignment of budget to PRSP - 62 - ANNEX III) World Bank Effectiveness: • Bank support during PRSP formulation. • Bank support and alignment during implementation IV) The PRGF and IMF Support. V) Bank-Fund Cooperation and the JSA: • Role of the JSA • Bank-Fund cooperation VI) Main Conclusions. Staffing of Joint Case Study Teams 19. For joint case studies, teams will consist of at least one core evaluation team member from each evaluation unit, along with at least one senior in-country national consultant. Either OED or IEO will be assigned overall responsibility for coordinating the case study. The team will also be supported by junior consultants to be assigned by each unit as deemed necessary. The OED and IEO teams will nominate a case study coordinator. His/her responsibilities include overseeing coordinating interviews, arranging the in-country workshop, overseeing the production of inputs, and production of the final report.