Rwanda - Second Integrated Forestry Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 15329 IMPLEMENTATION COMPLETION REPORT REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (CREDIT 1811-RW) JANUARY 31, 1996 Agriculture and Environment Division Central Africa and Indian Ocean Department Africa Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS Monetary Unit: Rwandese Francs (FRw) Per US$ 1.00: Appraisal Date (May 6, 1986): FRw 83 Closing Date (June 30, 1994) : FRw 145 WEIGHTS AND MEASURES Metric System I stre = I st.cbm (m3) I m3 of stacked roundwood = 2 steres = 450 kg I mn3 firewood = 1.6 steres (st.cbm) = 580 kg 26 poles= I m3 (roundwood) I Toe = I ton oel equivalent 2.5 tons of air dried wood I Toe = I ton oel equivalent 1.45 tons of charcoal LIST OF ACRONYMS APEF Extension Unit (Actions a l'exterieur de la foret) CAT Technical Support Center (Centre d'appui technique) CFD French Development Fund (Caisse francaise de developpement) CTFT Technical Center for Tropical Forestry (Centre technique forestier tropical) DGF Forestry Department (Direction generale des forets) DCE Department for Control and Evaluation (Direction Contr6le et Evaluation) ERR Economic Rate of Return FAC French Aid and Cooperation Fund (Fonds d'aide et de cooperation) FAO Food and Agricultural Organization of the United Nations FAO/CP World Bank FAO/Cooperative Program FFN National Forestry Fund (Fonds forestier national) Frw Rwandese Francs (Francs rwandais) GEF Global Environment Facility GOR Government of the Republic of Rwanda GNP Gross National Product GBK Gishwati-Kigali-Butare Project (First Integrated Forestry and Livestock Project) IDA International Development Association ISAR National Research Institute of Rwanda (Institut des sciences agronomiques du Rwanda) MDC Ministry of Development and Technical Research (Ministere du developpement et de la cooperation) MINAGRI Ministry of Agriculture, Livestock and Forests (Ministere de l'agriculture, de l'elevage et des forets) ONF National Forestry Office in France (Office national forestier - France) ORINFOR National Information Service (Office Rwandais d'information) OVAPAM Office for the Development of Agriculture and Livestock in Mutara (Office pour la valorisation des produits agricoles du Mutara) PACZN Zaire-Nile Forest Conservation Plan (Plan d'action pour la conservation des forets de la crete Zaire-Nil) PASP Second Integrated Forestry Project (Projet agro-sylvo pastoral) PMC Project Management Committee PPF Project Preparation Facility PSA Agricultural Services Project (Projet services agricoles) RPF Rwandese Patriotic Front (Front Patriotique Rwandais) SAR Staff Appraisal Report TFAP Tropical Forestnr Action Program UGZ Forest Management Unit (Unite de gestion de zone) UNAMIR United Nation Assistance Mission for Rwanda (Mission des Nations Unies pour l'Assistance au Rwanda) UPEL Federation of Gishwati Livestock Owners (Union pour l'elevage laitier) FISCAL YEAR OF GOVERNMENT January I - December 31  FOR OFFICIAL USE ONLY REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (Credit 1811-RW) IMPLEMENTATION COMPLETION REPORT CONTENTS PREFACE ......................................................i EVALUATION SUMMARY ......................................................i PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE .....................................................1I Project Identification ....................................................... A. Project Objectives, Responsibilities,and Standard Steps ...................... ...............................1I B. Major Factors Affecting the Project ......................................................4 C. Achievement of Objectives ......................................................7 D. Project Sustainability ..................................................... 12 E.Performance of the Association and the Borrower ..................................................... 14 F. Assessment of Outcome ..................................................... 16 G. Lessons Learned and Recommendations for Future Operations ..................................................... 17 PART II: PROJECT REVIEW FROM BORROWER'S PERSPECTIVE ................................................... 20 PART III: STATISTICAL INFORMATION ..................................................... 21 Table 1: Summary of Assessments ..................................................... 21 Table 2: Related Bank Credits ..................................................... 22 Table 3: Project Timetable ..................................................... 23 Table 4: Credit Disbursements ..................................................... 24 Table 5: Disbursement Profiles per Component and Category ..................................................... 26 Table 6: Summary of Project Achievements ..................................................... 27 Table 7: Project Costs and Financing ..................................................... 30 Table 7: Status of Legal Covenants ..................................................... 31 Table 8: Bank Resources: Staff Inputs ..................................................... 37 Table 10: Bank Resources: Missions ..................................................... 38 Maps IBRD No. 20054, 20055, 20056, 20057 ..................................................... 39 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. I   IMPLEMENTATION COMPLETION REPORT REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (Credit 1811-RW) PREFACE This is the Implementation Completion Report (ICR) for the Second Integrated Forestry Project in Rwanda, for which Credit 1811-RW in the amount of SDR 11.0 million was approved by the Board on August 14, 1986. The credit was closed on June 30, 1994, one year behind schedule. At this date, the undisbursed balance of about SDR 6.0 million was canceled. At the Credit closing date, the country was in a state of civil war following the death of the President on April 6, 1994. The consequences of the war on project investments are difficult to evaluate under the present situation of political uncertainties, since all project information has been destroyed, with the exception of the offices, documents and equipment of the Nyungwe project component. Moreover, the majority of project staff, including the most influential people of the Project Management Committee, have either been killed or fled outside the country. Parts I and III of the ICR were prepared by the Agriculture and Environment Division of the Central Africa and Indian Ocean Department of the Africa Region. On August 21, 1995, the Bank sent the Borrower Part I and III with the request to prepare Part II. No comments have been received to date. Preparation of the ICR was done about nine months after the Credit was closed. Since most of the project documents were destroyed and offices looted during the 1994 civil war, the ICR is based mainly on documents available in Bank files at the time the report was written. They include, inter alia, the Staff Appraisal and Project Implementation Reports, the Development Credit Agreement, Preparation Reports, Supervision Reports, the 1993 Evaluation of the Zaire-Nile Ridge Forest Conservation Strategy, correspondence between the Bank and the Borrower, internal Bank memoranda, annual reports prepared by the three Project Units and reports on discussions with representatives of the Borrower and the Bank.   IMPLEMENTATION COMPLETION REPORT REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (Credit 1811-RW) EVALUATION SUMMARY 1. Project Objectives. The Second Integrated Forestry Project (PASP; Credit 1811- RW); under implementation from October 13, 1988 to June 30, 1993, with its original closing date extended to June 30, 1994) was the follow-up of the First Integrated Forestry and Livestock Development Project (GBK; Credit 1039-RW) that was financed by IDA from 1980 to 1987. Recognizing the environmental and economic threat posed by rapid destruction of natural forests, both projects were aimed at: (i) expanding the source of wood products through industrial plantation and rural afforestation; (ii) enhancing the value of protected natural forest ecosystems; (iii) developing a high productivity cattle industry from degraded natural forests in the Gishwati area; and (iv) strengthening the institutional capacity of the forest and livestock agencies involved. In the Gishwati area, it was anticipated that most traditional nomadic cattle owners in the forest reserve would adopt an intensive dairy farming technique. 2. Although effective in achieving physical targets for plantation and pasture establishment, the GBK project suffered from the following: (i) inadequate plantation sites and species selection; (ii) limited stimulus to agroforestry development; (iii) high costs of pasture establishment after forest-clearing in the Gishwati area; and (iv) deficient management of the dairy facility due to the high operating costs of the obsolete equipment. Moreover, the takeover of pastures by influential politicians and project staff induced an inadequate rate of cattle destocking from the Gishwati forest. Furthermore, unsatisfactory project management resulted in misprocurement and the misuse of funds. In addition, intensive forest clearing and pasture development in Gishwati hampered the development of surrounding communes and dislodged the Batwa (Impunyu) from their  ii Evaluation Summary natural habitat and traditional way of life. No resettlement program was foreseen under the project and the Batwa became internally displaced persons. 3. The Second Integrated Forestry Project was built upon the experiences of the first project and included five main components aimed at redressing first project deficiencies: (i) the protection of the Nyungwe forest (PASP II/UGZ3); (ii) the promotion of agroforestry and rural afforestation (PASP II/DGF); (iii) development of integrated forestry, agriculture and livestock in Gishwati (GBK Gishwati); (iv) the maintenance of fuelwood and industrial plantations established under the first project, particularly around Kigali and Butare; and (v) the rehabilitation of pastures in the Mutara region. These components were also aimed at providing a major impetus to the implementation of Government policies on environmental protection, the promotion of agroforestry and the development of integrated forestry, agriculture and livestock strategies in the Gishwati forest. 4. Project Financing. The project was financed by an IDA Credit in the amount of SDR 11.0 million (US$14.1 million), covering about 70 percent of project costs. For implementation purposes, three separate Special Accounts were opened and maintained in the name of the UGZ 3, DGF and Gishwati Project Units. At the closing date of the Credit, the undisbursed balance was about SDR 6 million. 5. Achievement of objectives. The project was plagued with problems throughout its implementation. Firstly, the Credit did not become effective until October 13, 1988, due to strong disagreements regarding the implementation of the Gishwati livestock component. These disagreements persisted during the first project years, thus compelling the Association to phase out its participation in the Gishwati livestock component. As a result, an amount of SDR 1.6 million was suspended from the Credit on February 28, 1989. Furthermore, the outbreak of civil war in 1990 seriously undermined the project since the Government delayed provision of local funds needed to cover its share of operating costs, counterpart investment capital and expropriation fees. As a consequence, project units suffered from increasingly acute financial problems and relations with the Bank deteriorated. The financial situation of the Gishwati component became such that suppliers not only stopped delivery but initiated legal actions against the Government in order to obtain payment for goods and services. Consequently, operations in Gishwati came to a virtual standstill after the resurgence of the 1993 civil war. 6. Results. The main consequence of implementation problems, policy flaws and poor management capacity was that agroforestry and plantation management objectives were not achieved and the institutional strengthening of forestry and livestock institutions involved in the project fell substantially short of the envisaged goals. First, a large proportion of the GBK industrial plantation, which was unused and unintegrated into local economies, was virtually transformed into an open-access resource following the 1990 civil war. As such this became an easy target for being illegally harvested for charcoal, cleared for agriculture or burned. Second, with the collapse and subsequent  Evaluation Summary iii discontinuation of the UMUGANDA system' in 1991, most of the forest seedlings in subsidized nurseries were produced for wealthy farmers and local officials. Third, the Nyungwe and Gishwati buffer plantation program was continuously disrupted by land- use conflicts with the local population, which had not been involved in management activities or in defining the boundary location. Fourth, the Gishwati industrial plantation management objectives were never met, in view of their low economic viability due to site distance, low standing quality and lack of market outlets caused by the limited purchasing power of urban end-users. Finally, because of reluctance to hire consultants from the credit, the training and technical assistance programs were inadequate, thereby curtailing institutional strengthening objectives. As a result, the Forest Department remained ill-prepared and incapable of gradually transferring the management responsibility to local communities and the private sector as had been anticipated and resolving the numerous implementation problems. In that respect, the motivations and attitudes of the staff involved in the project conflicted with the project's objectives. 7. With regard to the Gishwati livestock component, the Government decided to go ahead with the clearing of the remaining natural forest, pasture extension and purchase of dairy equipment, despite the Association's withdrawal. In early 1994, more than 12,000 ha had been cleared and 10,000 ha of pastures had been stocked with 12,600 heads of cattle. The most disturbing effect of this situation was the high claims on Government resources for the development and maintenance of pastures. In addition, pasture management faced numerous constraints: the livestock federation envisaged never materialized and cattle owners did not contribute to management costs through adequate pasture contracts. As no satisfactory land use solution was provided by the project, cattle continued to graze and degrade the last remains of the natural forest. Other project setbacks related to dairy activities included inadequate processing equipment, poor milking and sanitation, and an inadequate transport system. Equally, selling prices and output in excess of market demand led to considerable milk waste throughout the project. Moreover, as a result of the war in 1990, work on the Mutara pasture rehabilitation component came to a halt. 8. Until 1994, the successful implementation of the Nyungwe forest conservation component, was a contrast to the dismal results of the Gishwati livestock component. Despite the forest's difficult security situation, the project component had: (i) built up the national capacity to carry out a biodiversity inventory in other high-altitude forests in Rwanda and elsewhere; (ii) established a long-term database for iteratively accumulating knowledge on African high-altitude pristine forests and social systems, allowing for the development of predictive monitoring and trends analysis, and the evaluation of management scenarios; (iii) increased the knowledge of the biological status of the Nyungwe forest. Besides a simple taxonomic inventory, the methodology used has led to an in-depth analysis of the composition and structure of the communities, the diversity factors, the evolutionary dynamics of individual taxa, the hydrological and climatic Communal enforced labor.  iv Evaluation Summary regulation functions of the forest ecosystem; and (iv) established a software tool (the Ecobase) that will considerably increase research efficiency by making available a database on the various species that form the biological communities of the forest ecosystems of the Zaire-Nile ridge. 9. Sustainability. As a result of the 1994 civil war, a large proportion of project investments under the first and the second projects was lost. Offices, buildings, documents, and equipment were looted. All of the pre-war 12,500 pure-bred, cross-bred and Ankole cattle in Gishwati were slaughtered or taken out of the country. Deforestation and shifting cultivation are spreading rapidly in the Nyungwe and Gishwati natural forests and buffer and industrial plantations are gradually being cleared for agriculture. On the infrastructure side, bridges, segments of access roads and tracks built under the project, particularly along the periphery of the Nyungwe forest, were destroyed. However, ninety-five percent of the ecological information base, used to prepare the Nyungwe forest conservation and management plans, has been safeguarded on CD ROM and floppy disks. The Nyungwe Conservation Center remained protected by UJNAMIR. 10. Lessons Learned. With regard to forestry, five lessons can be learned from the project's failures. First, in countries like Rwanda, rural populations excluded from decision-making and management of their forest resources will be deprived from their potential benefits, and thus, will inevitably contribute to their destruction. Second, the 1985 forestry policy framework should have been reviewed at the project design stage to include significant reforms more conducive to participation. Among such reforms as decentralization of forest resource tenure, integration of resource utilization into the local economies, decentralization of management responsibility, trade reforms, pricing and tax incentives, etc. should have been included Third, buffer and industrial plantations cannot simultaneously resolve fuelwood supply problems and protect natural forests. Household energy problems could have been resolved by measures steering consumers toward alternative energy sources, promoting rural afforestation and agroforestry, and encouraging the sustained management of natural forests. In light of their productivity and location, investment and recurrent charges from industrial plantation are too high at the current urban-market prices. It might be possible to improve product value in the future, for example, by increasing the production of construction wood and timber. However, a number of issues would need to be resolved, such as improving low quality materials which would require very expensive silvicultural tending; competition from Zaire's natural forest production; and end-users' limited purchasing power. Fourth, in order to ensure the success of sustainable forest management, conservation and agroforestry related-strategies should have called for structural, institutional and organizational changes that would have ultimately led to repercussions on the role, responsibilities and functions of the Forest Department. In particular, the DGF's technical capacity should have been further strengthened and staff attitudes modified to support local forestry initiatives. Other points of concern might have included the realignment of mandate, institutional incentives, decentralization of the decision-making process, better capacity to deal with sectoral issues as well as better sectoral background information, particularly of mechanisms that govern the marketing of forest products.  Evaluation Summary v And fifth, disbursements should have targeted recurrent expenses related to forest management and conservation activities. 11. While good technical choices might have been made on the Gishwati livestock component, in a country like Rwanda, under extreme land constraints, such a space- consuming ranch system should have been discarded in favor of dairy projects located on the periphery of the main urban centers and based on intensive smallholder stable systems. Land development and forest conversion costs would not have been incurred. However, semi-extensive production was the only strategy left once the decision was taken by the Government to clear the Gishwati forest to establish improved pastures, with Bank endorsement during the first project. 12. The setbacks experienced by the project components undoubtedly contributed to a better understanding of the issues involved in forest --and natural resources-- management. This also helped the DGF to refocus its sectoral policy. Increasingly, reform-minded technocrats within the DGF were requesting policy and institutional changes to better integrate natural resource management into their mandate. In 1993, the forest policy was being revised as a result of a better perception of sectoral issues. This initiative, supported by Switzerland, was, however, interrupted by the 1994 civil war. To date, most of the core cadre of sectoral specialists capable of dealing with sectoral issues has disappeared. 13. Recommendations for Future Operation. Six major recommendations could be derived from the project experience. First, topography, soil and land use conditions in Rwanda do not allow for cost-effective wood production in industrial plantations. Under the current price system, fuelwood production is too expensive, as well as being insufficient to yield enough to effectively satisfy the rapidly increasing urban demand for fuelwood. 14. Second, sustainable management of forest resources in Rwanda (including fuelwood, buffer zone and industrial plantation resources) will depend in large measure on the willingness of Rwanda's Forest Service to formulate a new integrated forest policy, correct structural constraints and policy distortions and obtain the administrative means and legislative tools to implement it. Policy reforms should be developed in response to sector issues raised by the major forest users and institutions in advance of the design of future detailed investment operations. They should include economic policy reforms related to forest resource pricing and fiscal policies; and forest management reforms related to the nature of state intervention for resource allocation. Full recognition needs also to be given to the opportunities for rural population and private sector participation in forestry, conservation and forest industry activities. Appropriate policy reforms should require a strong focus on sector work to help identify intersectoral policy flaws, and integrate forestry objectives in a wider economic and environmental framework. The process of formulating forest policy should be open to all stakeholders.  vi Evaluation Summary 15. Third, because Government's role in forestry development and environmental conservation will remain essential, greater attention needs to be paid to institutional reform and capacity building, particularly human resources development, and the long term horizon which this implies. The successful implementation of forestry and natural resource management project will require not only considerable institutional reforms in the structure and operations of the DGF, but also drastic changes in the very structure of the present forest policy framework. Changes required from an institutional viewpoint include: a stronger reorientation of DGF's mission towards social forestry and integrated natural resource; organizational restructuring of the DGF; improved internal organizational systems and procedures; decentralization and devolution of forest management authority; and increased cross-sectoral coordination and collaboration. Future projects should help restructure the sector's institutions (including the DGF and other dependent agencies, training, research, etc.), define an appropriate policy framework and a long-term human resource development program. As this was not the case under this project, forestry and livestock institutions remain unprepared to achieve their technical, legislative, judicial, and regulatory functions. 16. Fourth, project preparation should fully take into account stakeholders' views on project objectives and how they are to be achieved The principle is that local people should be fully involved in issues concerning them and the environment in which they live. Sustainability of project objectives will depend in the most part on the commitment of all interested parties. 17. Fifth, full advantage should be taken of opportunities to make forestry (and conservation) activities self-sustainable on a financial basis, in order to avoid potential difficulties with counterpart funding. Because the country's financial situation will remain very weak in the future, alternative sources of funding should be identified (such as GEF, Debt for Nature Swap, endowments for nature conservation, etc.) so that forestry development and conservation recurrent costs need not depend entirely on Government counterpart funds. 18. And Sixth, improved Government/donor coordination in policy dialogue, project design and implementation is essential to maximizing potential development impacts.  IMPLEMENTATION COMPLETION REPORT REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (Credit 1811-RW) PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE Project Identification Project Name: Second Integrated Forestry Project Credit No. 181 1-RW Operational Responsibility: Africa Region, Central Africa and Indian Ocean Department, Agriculture and Environment Division Country: Rwanda Sector: Agriculture Subsector: Forestry A. Project Objectives, Responsibilities and Standard Steps 2.1 Sector Background and Issues. In the past, Rwanda has had a long standing commitment to forestry and natural resource management and biodiversity protection. According to the latest estimates, forest vegetation covered an area of 668,000 ha or 28 percent of the country land's area, including agroforestry and dispersed trees. Though densely populated, it was also one of only a few African countries with nearly 13 percent of its surface designated as protected areas containing a remarkable variety of unique habitats and species. They include the four remaining Afromontane natural forests that play an important ecological and hydrological role (Nyungwe, Mukura, Gishwati and Birunga protected forests) and the Akagera National Park with the adjacent Mukura game reserve. All are, however, under permanent threat through resettlement, illegal felling, poaching, mining and pit sawing. In addition, erosion control and reforestation programs were successfully conducted until 1990. Agroforestry, meanwhile, expanded significantly; on-farm production met some 70 percent of the country's wood requirements in the early 1990s. Since 1985, increasing attention has been given to the role of natural forest protection in Rwanda's socioeconomic development. In December 1992, recognizing the increasing environmental and economic threat posed by the rapid destruction of natural forest cover, the Government of Rwanda (GOR), assisted by the donor community, developed the Zaire-Nile Ridge Natural Forest Conservation Strategy. This strategy aimed at achieving an equilibrium between demand and supply of forest  2 Part I: Project Reviewfrom the Bank's Perspective products, while maintaining an optimum amount of natural forest cover for ecological and watershed protection purposes. 2.2 Over the past decade, however, the ecological situation worsened due to a combination of socioeconomic and demographic factors. They include the following: Extremely high population growth and density. With a total population estimated at about 7.5 million, increasing at a annual rate of 3.7 percent since the 1940s, Rwanda is one of the most densely populated countries in Africa (290 people per square kilometer). At this growth rate, 1992 projections indicated a population of 25 million in the year 2030 or 1,373 persons per usable square kilometer. As a result, average farm size continues to decrease while fallowing has virtually disappeared. Apart from the increased utilization of land, unequal distribution of farmlands has pushed an ever larger number of people onto ecologically sensitive areas, such as erosion-prone highlands and the remaining natural forests.2 These trends were further exacerbated by the development of governmental or parastatal projects that have appropriated relatively large acreage for programs and that have produced few benefits for local farmers or the state. In early 1994, about 70-75 percent of the rural population was considered landless or nearly landless; * A decline in agricultural production and food availability. Since the mid-1980s, agriculture alone was unable to support the rural population; some parts of the country were either forced into a state of chronic food deficits or became highly susceptible to minor climatic perturbations. Agricultural intensification programs introduced few new profitable techniques, improved varieties or agricultural inputs, while soil degradation lowered the productivity of many farms. Government and project efforts to control soil erosion through terracing and agroforestry were ill adapted and insufficient to increase production enough to match the high population growth. Moreover, prices paid to farmers for agricultural products stagnated at low levels due to the unfavorable terms of trade (particularly for coffee on the international market since 1988) and limited national market outlets; and * Limited off-farm opportunities in rural areas outside the low-paying agricultural labor market. Very few non-agricultural programs to create rural employment opportunities were developed by the Government or the donor community, despite common acknowledgment that the future of Rwanda's rural areas could not depend solely upon agriculture. In addition, restriction of migration to cities, coupled with lack of economic opportunities, did not offer alternatives for young rural people. With few non-farm options, people had no other alternative than converting nearby pasture, valleys and natural forests into extensive cropping (or long distance migration to the East to settle on former pasture lands). 2 cf. Evaluation of the Zaire-Nile Forest Conservation Strategy, Kigali, November 1992.  Part 1: Project Review from the Bank's Perspective 3 2.3 Project Objectives and Components. The Second Integrated Forestry Project followed a First Integrated Forestry and Livestock Project (GBK Project; Credit 1039- RW), which was implemented from 1980 to 1987. Both projects were financed by IDA. The first project's loan of US$21.0 million was fully disbursed at its closing date on June 30, 1987. The Second Integrated Forestry Project, as appraised in 1987, was within the 1985 policy and priorities and responded to the views of most of the donor communities. 2.4 The first project --known in Rwanda as the GBK project-- had two distinct objectives: (i) to strengthen institutional capacity for policy formulation and to initiate an industrial plantation program for urban energy needs; and (ii) to support integrated forestry and livestock development in and around Gishwati to destock the natural forest reserve. Although effective in achieving physical targets for plantation and pasture establishment, the project suffered many shortcomings. Plantation site and species selection was inadequate, thus limiting the economic potential of the industrial plantation. In addition, work to encourage agroforestry was limited. In the Gishwati area, pasture establishment costs after natural forest clearing were higher than anticipated and the management of the dairy facility was hampered by higher operating costs due to poor choice of equipment and diseconomies of scale. Moreover, the destocking of cattle from the Gishwati natural forest has been met with only limited success due to the takeover of many pastures by members of the President's family and high government officials and project staff. Furthermore, project management was deficient, which resulted in misprocurement and misuse of funds. In addition, the focus on intensive forest clearing and pasture development in Gishwati prevented the development of surrounding communes and forest Batwa (Impunyu) were cut from their traditional way of life. As there was no resettlement program under the project, they migrated to other regions, Kigali in particular. 2.5 The Second Integrated Forestry Project was prepared by the FAO/CP in 1985 and appraised by the Bank in 1987. It was aimed at correcting problems arising from the first project; providing a major impetus to the implementation of Government policies for the protection and conservation of the remaining natural forests in Nyungwe and Gishwati and the promotion of agroforestry to increase fuelwood supply. The project was also expected to complete and maintain the investments in industrial plantations and pastures made under the first project on the assumption that project implementation responsibilities would be gradually transferred, where suitable, to the private sector and communes. As appraised, the project was to: * Develop an ecological inventory to complete the information base needed to sustainably manage and monitor the Nyungwe and Gishwati natural forests, create about 1,100 ha of buffer plantation at their periphery and establish a Forest Conservation Center to coordinate all forestry activities on the Zaire-Nile Ridge; * Manage about 3,600 ha of productive fuelwood plantation established under the first project and expand on-farm and agroforestry activities through the rehabilitation of 60 sector nurseries and the promotion of collective and private nurseries;  4 Part 1: Project Reviewfrom the Bank's Perspective * Support integrated forestry, agriculture, and livestock development in Gishwati by maintaining industrial plantations (2,100 ha) and pastures (1,600 ha) established under the first project and developing an additional 2,000 ha of new industrial plantations based upon the recommendations of a land-use study to support an integrated approach to land management in the degraded areas of the Gishwati forest; * Intensify agricultural extension services in five communes located on the northwest border of the forest; and * Support the rehabilitation of 1,500 ha of pastures established under the Second Mutara Rural Development Project (Credit 937-RW), which was closed December 31, 1986. 2.6 The project also included a capacity building program aimed at strengthening the Forest Department's (DGF) managerial, technical and monitoring capacities, as well as an applied research component to respond to problems directly linked with the implementation of the above components. 2.7 The project was financed by IDA Credit 181 1-RW with an amount of SDR 11.0 million (US$14.1 million). Under the Development Credit Agreement, the GOR was to provide about US$6.0 million, or 30 percent of project costs, including taxes. The credit was extended by a year to June 30, 1994 to: (i) complete the ecological inventory in the Nyungwe forest; (ii) launch pioneering actions to test new participatory plantation and forest management and buffer zone development options; (iii) formulate a new forest policy framework; and (iv) prepare a follow-up Natural Resources Management and Environment Project, integrating the recommendations and proposals of the National Environmental Strategy for Rwanda ("Strategie Nationale pour l'Environnement au Rwanda" or SNER/NEAP) as well as lessons learned from the project. B. Majors Factors Affecting the Project 3.1 The project was plagued with major implementation problems, due to a number of factors, among them: (i) specific disagreements between the Bank and the Government regarding the Gishwati component; (ii) the decreased availability of counterpart funding, which was exacerbated by higher military spending; and (iii) the upheaval and destruction that followed the outbreak of civil war. 3.2 Effectiveness and Suspension of the Livestock Component. The Credit became effective fourteen months after Board approval because of strong disagreement between the Bank and the Borrower regarding the scope and the content of the Gishwati component. From the Bank's perspective, a major bone of contention arose from the allocation of pastures and related issues of milk-pricing and cost recovery under the first  Part 1: Project Review from the Bank 's Perspective 5 project. As mentioned in paragraph 2.4, most of the pastures developed under the first project were allocated to small groups of beneficiaries belonging to the President's family, high government officials and project staff, in spite of the project objective to destock the local Ankole cattle from the Gishwati forest in exchange for higher yielding Brown Swiss dairy cows. In addition, as a result of land use conflicts between the local population, new immigrants and the project, the Bank decided that no further pasture development would be financed until the completion of both a land use and livestock sector study. This study would answer fundamental questions about future patterns of development and integration of small livestock owners. In spite of these disagreements, the Government decided to go ahead with further pasture extension using its own funds. As a result, the financing of the pasture management component and dairy activities was discontinued in February, 1990, when it became evident that the Government did not intend to undertake the proposed land use study. In spite of the recommendations of the Bank supervision missions,3 12,000 ha were fully developed from degraded natural forest stands and stocked with about 12,500 cattle units in early 1994. 3.3 1990/93 Civil War Impacts. The repeated conflicts of the civil war of the 1990s, and the difficult public finance situation had three major consequences for the project. First, the consequences of the civil war were a worsening of an already difficult public finance situation. As a consequence, the performance of externally-financed projects deteriorated and disbursements dropped dramatically, particularly because of inadequate budgetary allocations for counterpart funds. Moreover, the civil servants' commitment was weakened by lower salaries and expenses. Second, with the creation of new political parties opposed to the ruling party, the Government's power was reduced. For many farmers, this situation meant that they no longer owed allegiance to Government- designated officials. As a result, in 1992, farmers began to refuse enforced communal labor and, consequently, UMUGANDA was officially abandoned by the Government in late 1992. Resentment was such that some farmers even destroyed anti-erosion ditches that they had dug under communal labor and burned government-owned plantation and woodlots planted on communal (and private) lands. And third, about 900,000 people were displaced by the successive hatreds and forced to live in refugee camps located in the periphery of Kigali. As a result, a large part of the GBK fuelwood plantation was further encroached or illegally exploited by charcoal makers to supply the refugee camps with wood fuel. 4.4 Impact of the 1994 war. During 1994, a large majority of the project investments made under the first and the second projects was destroyed. Offices, buildings, documents and equipment were looted. All of the pre-war 12,500 pure-bred, cross-bred and Ankole cattle existing in Gishwati were slaughtered or taken out of the country. Deforestation and shifting cultivation are spreading rapidly in the Nyungwe and Gishwati natural forests, and the gradual clearing of immature buffer and industrial plantations for agriculture is increasing. Many bridges and segments of access roads and 3 See: Post appraisal mission's BTO of October 31, 1986.  6 Part L: Project Reviewfrom the Bank's Perspective tracks built under the project, particularly along the periphery of the Nyungwe forest, have been destroyed. However, ninety-five percent of the ecological information base established to prepare the Nyungwe forest conservation and management plans have been safeguarded on CD ROM and floppy disks. 4.5 Financial Management, Accounting and Disbursements. Throughout project implementation, the quality and rigor of the financial management were largely inadequate as highlighted by the successive supervision missions. Major delays occurred in either disbursements, procurement or replenishment of the three Special Accounts. Furthermore, the project was seriously undermined by govermnent delays in providing local funding needed to cover its share of operating costs, counterpart investment capital and expropriation fees. As a consequence, project units suffered from increasingly acute financial problems, and relations with the Bank deteriorated. The Gishwati component's financial situation became such that suppliers not only stopped delivery of supplies but undertook legal action against the project unit causing a virtual standstill of operations after the resurgence of the civil war in 1993. At the Credit's closing date, an amount of SDR 5,998,383 was canceled effective October 31, 1994. As a comparison, it is worthwhile to note that the first project loan of US$21 million had been fully disbursed at the project closing date. Table 5 of Part III shows the detailed breakdown of disbursements by financing category and component, which is summarized as follows: Category Credit Disbursed at % Disbursed Allocation Closing Date at Closing Date PASP Il/ Nyungwe 2,950,000 2,142,843 72.5 PASP II/DGF 1,277,000 516,352 40.4 GBK/Gishwati 3,398,000 1,416,796 39.6 Suspended Gishwati 1,604,000 PPF 1,166,000 635,369 54.0 Unallocated 605,000 290,2574 | Total 11,000,000 5,001,617 45.5 4 The amount of SDR 290,257 is the difference between the SDR value of the initial deposits and the SDR value of the accounts at closing, adjusted for any recoveries of the initial deposits.  Part l: Project Reviewfrom the Bank's Perspective 7 C. Achievement of Objectives 4.1 Ecological Inventory in the Nyungwe and Gishwati Forest. The Nyungwe and Gishwati forests were among the richest remaining high altitude pristine natural forests of the eastern rift and major centers for endemism, both at the regional and continental levels. Before the Bank's intervention, a number of research activities on forests were carried out over the years in the Nyungwe forest, but no real consensus on viable management and conservation options was ever reached. In particular, the knowledge of the forest ecosystems was insufficient to justify overall conservation as an alternative to unprofitable and unsustainable forest exploitation. The ecological inventory, carried out under the project, was able to provide information not only on the forest ecosystems' diversity, floral/fauna composition, structure and functions, but also on the spatial complementarity of the biota among habitats within sites. As a result, the project component has: (i) built up a national capacity that will be able to carry out biodiversity inventory in other high-altitude forests in Rwanda, in neighboring countries and elsewhere in Africa; (ii) established a long-term database for iteratively accumulating knowledge on African high-altitude pristine forest and social systems, allowing for the development of predictive monitoring and trends analysis, and the evaluation of management scenarios; and finally (iii) established a software tool (the Ecobase) that will considerably increase the efficiency of research by making available a large amount of data on the various species that form the biological communities of the forest ecosystems of the Zaire-Nile ridge. 4.2 Extension of the ecological inventory to the Gishwati forest reserve never materialized because of security reasons after the resurgence of the civil war in 1993. 4.3 The Buffer and New Industrial Plantation Program. The project achieved 240 percent of the target for buffer plantation in Nyungwe, and 66 percent in Gishwati as shown in Part III, Tables 6A to 6C. Since the beginning of the project, however, damage to young trees was widespread due to inattentive intercropping patterns, land tenure and/or expropriation conflicts and/or because farmers had no long-term rights to the land or trees (as they do in true "taungya systems")5. As a result, all buffer plantations were suspended in January 1993. From observations made during the December 1994 mission, the gradual clearing of immature buffer plantations located at the periphery of the Nyungwe and Gishwati forests for agriculture, already experienced since 1990, had dramatically increased as a result of the 1994 civil war. It is now estimated that about 60 percent of the buffer plantation in Nyungwe and 90 percent in Gishwati must be counted as a loss and subtracted from project benefits. In the first Gishwati project, industrial plantation, indiscriminate pruning and heavy thinning to open the plantation to livestock grazing, drastically reduced its wood production potential. In the absence of benefit 5 Lease of forest land to farmers for a maximum of three years.  8 Part 1: Project Review from the Bank's Perspective sharing arrangements and market outlets, most of the wood remained in the plantation, thereby increasing fire hazards. 4.4 These gloomy results are due to a number of factors which hampered implementation of buffer and industrial plantation. They include: (i) lack of local counterpart funds to pay expropriation fees; (ii) the unwillingness or the very limited ability of the project unit staff to solve the numerous conflicts with the local population; (iii) the failure to design and implement a comprehensive participatory strategy as outlined in the SAR; and (iv) the conflicts posed by the eviction of the resident population after the plantation of trees and the termination of taungya agreements. Despite the numerous recommendations of supervision missions, no tangible progress toward resolving these problems took place during project implementation. This situation led to persistent disagreements with the Bank, as well as, more recently, the systematic destruction of the buffer plantations and a dramatic increase in forest conversion for agriculture. 4.5 The Nyungwe Forest Conservation Center. The construction of the Nyungwe Conservation Center at Kitabi started in 1992, and was finished in 1993. It now includes, in addition to the Conservation Center and the project office, 23 houses, one barrack for forest guards, and one health center. Because of the 1994 civil war, work on electricity, telephone and water was interrupted, as were site development works, landscaping, access roads and fencing. No decision was ever taken with regard to the management of the Conservation Center, or on how to finance operating and recurrent costs of such a compound, despite the recommendations of Bank supervision missions and discussion with other donors involved in forest conservation activities. During the 1994 civil war, the Conservation Center was occupied and protected by the UNAMIR. 4.6 Agroforestry and Rural Afforestation. On average, about 2 million seedlings, mostly eucalyptus, grevillea and pines, were produced per year in the sixty sector/Government nurseries. Around mid-1993, the privatization of sector nurseries was officialized as a result of a Ministerial decree. In early 1994, the production of agroforestry seedlings started in 20 pilot private nurseries around Kigali and 12 around Butare. The resurgence of the civil war, however, disrupted all plant production activities after April 1994. 4.7 It was expected at appraisal that only 20 percent of the seedlings produced in sector nurseries were to be sold. The rest was expected to be distributed to the UMUGANDA members in return for their labor. However, the achievements of the component fell short of expectations. The distance of nurseries from farms prevented smallholders from getting seedlings for their own use. As a result, most of them were planted on communal land or sold to government employees, officials and/or large-scale land owners at subsidized prices. Furthermore, the lack of technical packages able to stimulate grassroots involvement in the proposed program was undoubtedly an important additional constraint. In particular, an assessment of farmers' attitudes towards trees and species as well as factors affecting the adoption of agroforestry innovations by small  Part L: Project Review from the Bank 's Perspective 9 farmers would have been needed to shape the packages in such a way that they could be adapted to the traditional cultural systems and local constraints. Moreover, it would have helped the project to identify more suitable species. Studies carried out in the project area by other projects showed that more than 100 species (75 percent native) are used in the traditional agroforestry systems while the species extended by the project represented only 15 percent. 4.8 Because of these shortcomings, small land owners and the rural poor did not perceive the need for getting trees from sector nurseries. Nor did the mutual mistrust that characterized relations between the forest technicians and the rural population foster increased grassroots involvement in private nursery development. In hindsight, however, in a country like Rwanda with so many land tenure problems, a more pragmatic approach to influencing farmers to accept agroforestry in their farming systems would have been to build on their traditions and, where necessary, attempt to adjust them gradually, without introducing drastic changes or new species. 4.9 Management and Maintenance of the Kigali-Butare Fuelwood Plantation. Management and maintenance of fuelwood plantations included tending operations, firebreaks cleaning and maintenance of access tracks. They were executed in conformity with project objectives until 1991 (3,400 ha). After 1990, however, political disturbances associated with counterpart funding problems led to an increasing incidence of forest fires, encroachments, and illegal felling for charcoal-making, which negatively affected the production potential of the GBK plantation. Another consequence of the political situation was the interruption of the plantation inventory and the completion of management plans. 4.10 In 1991, political, socioeconomic and budgetary factors affecting the management of the GBK fuelwood plantation prompted supervision missions to recommend the development of a new management strategy involving local groups in silvicultural activities under management contracts. To implement this strategy, the DGF decided to subcontract the preparation of participatory management plans to local consulting firms. However, the importance of illegal felling prompted the DGF to launch a salvage operation to control the supply of woodfuel from the GBK plantation to some 600,000 refugees established around Kigali. This initiative was taken in full accordance with the guidelines provided by the supervision missions. The Swiss Cooperation financed the transport of fuelwood and charcoal produced under "controlled" exploitation. 4.11 Implementation issues experienced under this component cannot be explained by budgetary issues and political disturbances only. In the past two decades, Rwanda's forest policy has focused on protecting a forest resource on land seized from the rural population and on a hypothetical development of new wood resources through industrial plantation. In the process, any grassroots initiative was stifled. The very fast spread of fires, encroachments and illegal felling for charcoal-making showed that the forest governance framework was no longer adequate. Most of the mature plantations in Rwanda, in competition with private and communal plantations and agroforestry also  10 Part : Project Reviewfrom the Bank's Perspective experienced wood marketing problems and land tenure conflicts with farmers and pastoralists.6 As a result, wood fuel prices remained far below actual costs of plantation7. Addressing these forest management and pricing issues had required governance reforms to promote decentralized decision-making, revenue-sharing arrangements and incentives to foster local participation and private sector involvement. 4.12 Monitoring of Rural Forestry Activities. The creation of a Rural Forestry Monitoring Unit was requested by the Bank in order to: (i) define qualitative monitoring and evaluation criteria to assess the impact of rural forestry activities; and (ii) formulate a national rural forestry strategy. Although the DGF drew up some preliminary proposals from the consultant's recommendations, the unit, set up in 1993, was never really operational in spite of numerous supervision recommendations. 4.13 Applied Research Program. The applied research program proposed at appraisal was expected to play a decisive role in the choice of management strategy and technologies in improving the knowledge of technical, financial and socioeconomic constraints affecting the management of GBK fuelwood industrial plantation. The contract with the National Research Institute of Rwanda was never materialized because of conflicting priorities, internal constraints and lack of staff and capacity. The program fell short of expectations, thereby curtailing prospective yield improvements through better selection of plant material, silvicultural treatments and management intensification. 4.14 DGF Strengthening. The main activities included the establishment of a seed center in Kigali and the training of five high level personnel in forest engineering (1), agroforestry (2), rural development (1) and remote sensing (1). Other activities included participation in seminars abroad and short-term training in forest management, legislation, agroforestry, nursery techniques, charcoal-making (in association with the UNDP/Bank Energy Program) and silviculture for DGF's staff, MINAGRI's agricultural monitors and communal authorities. To help manage and define project priorities during implementation, two basic studies were carried out by international consultants: a wood marketing study to enhance the value of existing plantations and a study to identify implementation modalities of the Monitoring and Evaluation Unit as well as DGF's training needs. Both studies were inconclusive and were not adequately followed up. 4.15 Livestock Development. On February 28, 1989, a partial cancellation of this component was proposed to the Government in view of its reluctance to meet the condition for disbursement. The conditions required an appropriate milk pricing policy that would ensure efficient production and marketing of milk. Another disagreement was 6 The supervision missions noted that forest fires might have also been initiated as a reaction against "umuganda" used to establish woodlot and plantation under the GBK project. 7 The low market price for woodfuels reflected the combination of generally low purchasing power of consumers relying on fuelwood for cooking, limited markets for commercial fuelwood and charcoal, and supply from free sources or sources where the total cost of the wood was not fully perceived by the supplier.  Part I: Project Reviewfrom the Bank's Perspective II related to the allocation of pastures and the terms of reference for the land use study. Since the Government never officially responded to this proposal, SDR 1.6 million was unilaterally suspended from the Credit allocation. As a result, the Government, without support from the Association, decided to go ahead with deforestation, further pasture extension and equipment of the dairy. From 1990 onwards, all supervision missions related to livestock activities were discontinued. By early 1994, 12,000 ha were cleared and 9,900 ha fully developed and stocked with 12,600 heads of cattle as detailed in Part III table 6E and F. 4.16 Pasture management was affected by many constraints throughout the course of the project. The livestock federation never materialized and cattle owners did not contribute to management costs through adequate pasture contracts. As a result, the Government continued to subsidize land development and recurrent costs. Since no adequate land use solution was provided by the project, local cattle continued to graze and further degrade the remaining natural forest. Attribution of pastoral units remained unregistered. There was no concern about the fate of the forest Batwa. 4.17 Other project setbacks were related to dairy and slaughtering activities. With the purchase of additional equipment by the Government, the dairy processing capacity was upgraded to 8,000 liters per day. However, inadequate and aging processing equipment purchased under the first project, poor milking and processing sanitation, and an inadequate transport system as well as high producer prices, low consumer prices and output in excess of market demand led to considerable wastage of milk throughout the project. The major constraints to cattle fattening and meat marketing included irregular supply of agricultural by-products (cattle cake, molasses), low selling prices of carcasses, transport difficulties and lack of market outlets in spite of meat export expectation and reduction of illegal imports from Uganda. In practice, government staff and military camps were the only regular customers. 4.18 In view of these poor achievements, another disturbing result of the component is related to the high claims on the budget to develop and maintain the pastures unplanned at appraisal. Between 1987 and 1991, an average of about Frw 200 million per year, or 12 percent of MINAGRI's total development budget, were allocated to the Gishwati component. 4.19 Agriculture Extension (APEF). The APEF extension program focused on three main activities in the three communes of Rwerere, Kanama and Mutura on the northwest border of the Gishwati forest: (i) extension of proven themes such as stall feeding of cattle goats and sheep, use of farm yard manure and composting, erosion control through on-farm terracing, improved techniques for potato cultivation, fodder edges, agroforestry and woodlot planting; (ii) support to animal health and nutrition by developing additional waterpoints, cattle crushes and bull sheds; and (iii) development of new themes focused on on-farn testing and introduction of multipurpose agroforestry species with some assistance from ISAR. Management responsibility of this program was transferred to the Agricultural Services Project (Credit 2026-RW) in 1992, two years  12 Part 1: Project Review from the Bank's Perspective behind schedule. The program fell short of expectations in the absence of monitoring and evaluation activities that not only prevented the evaluation of the impact of the program but also deprived the project unit of a management tool which might have enabled a more integrated approach focused on farmers' needs and constraints. 4.20 Mutara Pasture Rehabilitation. Past Bank experience in establishing ranches in the Mutara region has been hampered by the lack of organization of beneficiaries, and agreement with the livestock owners on user fees to cover recurrent costs. As a result, this component was first designed to compare the cost of manual versus mechanical range rehabilitation and maintenance over 1,500 ha; and second, to help develop range management associations and mobilize livestock owners to manage their own resources. To achieve these objectives, the project purchased two four-wheel drive tractors and one crawler. Due to the 1990 invasion of the region by the Rwandese Patriotic Front, these vehicles remained in Kigali and all activities were halted. D. Project Sustainability (i) Forestry 5.1 Economic Assessment. No economic analysis was calculated for comparison with the appraisal assessments due to the lack of reliable plantation productivity data and an adequate cost accounting system. From the above, it is obvious, however, that the actual economic rate of return is far below appraisal projections because of: (i) higher than estimated development costs; (ii) much lower plantation productivity; (iii) lower wood prices; and (iv) widespread illegal felling, forest fires and conversion to agriculture. In theory, it might have been possible to improve plantation value (for instance by increasing the production of poles and timber). This would, however, have required very expensive silvicultural tending to improve standing material. At the same time, competition from on-farm and communal forestry activities, low timber import prices from Zaire and low purchasing power of urban end-users were important limiting factors. 5.2 Buffer Plantations and Protection of Natural Forests. Despite buffer plantations, external pressures from the periphery continued to constitute the primary threat to the remaining natural forest. At the border of the Nyungwe forest, population densities have continued to rise due particularly to a high imigration rate. The socioeconomic study carried out in 1992 as part of the biodiversity assessment highlighted the contradiction that prevailed between alleged participatory strategies using the taungya method and actual repression-oriented field activities. As a result, buffer plantations remained an area of conflict between protection needs and local interests. In spite of the fact that project activities generated significant employment for thousands of local people, land expropriation related to the establishment of the buffer plantation have caused great resentment, reflected in the widespread damage and/or eradication of trees planted under taungya agreements. In the Gishwati area, where the Government completed the conversion of nearly two-thirds of natural forest into pastures between  Part 1: Project Reviewfrom the Bank's Perspective 13 1987 and 1990,8 buffer and industrial plantation sites were systematically encroached and seedlings eradicated by the local population for the production of potato seeds.9 By project's end, a population of about 10,000, mainly composed of local farmers, was employed by the project for forest clearing and expansion of pastures, and new immigrants were living in a narrow buffer zone between the forest and the pastures without clear land tenure rights. Supervision missions' proposals to provide the local population with clearly defined land-use rights associated with the development of a comprehensive soil conservation and agroforestry program were never seriously considered by the project unit. 5.3 However, the successful implementation of the Nyungwe ecological inventory contrasted with these gloomy results. The creation of a long-term ethno-ecological base would permit the future development of forest management and conservation activities and predictive monitoring and trend analyses if the present deforestation trends could be controlled. 5.4 Wood Supply from Fuelwood Plantation and Agroforestry. As a result of implementation constraints, including a very difficult political situation, institutional weaknesses, policy failures and design flaws, the components' objectives were not achieved. Unused and not integrated into local economies, a large proportion of the GBK fuelwood plantation, virtually transformed into an open-access resource following the 1990/93 civil wars, was illegally harvested for charcoal, cleared for agriculture or burned'0. Moreover, the agroforestry component fell substantially short of goals with the collapse and subsequent discontinuation of the UMUGANDA in 1991. Most of the forest seedlings produced in subsidized sector nurseries were absorbed by wealthy farmers and local officials. (ii) Livestock Development in Gishwati. 5.5 Pasture development. While the technical choices were good in theory as regards cattle breeds (with the exception of the Frisian), composition of pastures and the management strategy focused on dairy production, and beef as a by-product, the semi- intensive grazing production system on artificial pastures was economically unsustainable. In a country like Rwanda, under extreme land constraints, such a space- consuming range system should have been discarded in favor of dairy projects located at the periphery of the main urban centers and based on intensive smallholder stable systems without land development costs. It was, however, the only possible choice, since semi- 8 Leaving not more than 4,500 ha of military reserve in the north (Bigogwe), and an equal area of proposed nature reserve in the south (Kayove). 9 A large potato seed farm was established in the project area by ISAR in 1987, thereby attracting a large amount of migrant workers who remained in the project area after privatization in 1992. 10 A large amount of plantation programs supported by other donors experienced the same level of destruction. In 1993, the GBK fuelwood plantation around Kigali supported the bulk of the fuelwood and charcoal supplied to the refugees fleeing the northern battlefields.  14 Part 1: Project Review from the Bank 's Perspective intensive production was the only strategy left once the decision was taken by the Government and the Bank to clear the Gishwati forest and establish improved pastures. 5.6 Commercial Dairy. Furthermore, the choice of commercial dairy, with modem technologies, implying a central processing plant with modem equipment, efficient collection system, adequate road network and transport facilities, efficient veterinary and health services, hygiene, energy and, lastly, imported inputs for the packaging of the final product, is also questionable. In Rwanda, the dairy market is extremely limited, particularly for sour milk (80 percent of the project sales), a short conservation product requesting a low processing technology that cannot be sold far away from the processing plant. As a result, most, if not all, dairy and cheese processing facilities existing in Rwanda are facing financial difficulties, particularly because of the lack of market outlets. Again, in a country like Rwanda, where costly and lengthy transport should be avoided, the development of smallholder stable system with the dual purpose of producing milk for consumption and marketing, and manure for crop fertilization should have been preferred. E. Performance of the Association and the Borrower 6.1 Project Preparation Cycle. Project Preparation/Appraisal started with an FAO/CP mission and was completed in February 1986. A series of two appraisal missions with Bank staff and consultants finalized the appraisal document in May 1987 in line with 1985 forest policy. In retrospect, project design did not, however, pay sufficient attention to structural and institutional issues. While project justification and design were well elaborated, it is evident that forest, livestock and dairy productivity assumptions and price projections were over-optimistic. For the plantation components, the scarcity of basic data when the project was being prepared may explain the under-estimation of species selection and soil constraints on plantation yields. Not enough attention was paid during appraisal to (i) effective property and land use rights in the Gishwati area; and (ii) the problems affecting the marketing of plantation products, meat and milk. As regards project management, the capacities of the three project units to act on structural and sector issues and to adjust to the changing political, socioeconomic and budgetary situations were substantially over-estimated. These facts largely explain the setbacks recorded in each component. It should be noted, however, that (i) difficult negotiations took place during the preparation phase when the Bank team started to raise the right type of issues; and (ii) the appraisal report recognized the possibility of these setbacks, given the main sector issues identified at appraisal," despite the encouraging results recorded l The main issues included: (i) the lack of means to implement the 1985 forest sector policy as well as the delineation of responsibilities for the management of the sector; (ii) the limited capacity of DGF to monitor, evaluate and coordinate sector activities; (iii) the critical shortage of agricultural land; (iv) a very weak information base on the supply, demand pricing and commercialization of wood products, in particular regarding a possible over-supply of pine used as sawnwood in the medium term.  Part L: Project Reviewfrom the Bank's Perspective 15 by the first project. The Bank's error was to believe that the Government was open to progressive reform. 6.2 Supervision. The Bank generally did not respond adequately to GOR's development and budgetary constraints that affected project implementation. This is undoubtedly due to the lack of comprehensive sector work at appraisal addressing structural constraints, policy distortions underlying forest resource degradation and institutional weaknesses. Most of the problems faced by the project could have been overcome if an overall framework for sector development had been put in place before appraisal, clearly recognizing non-sector consideration in forest and pasture development (e.g. macroeconomic crisis, impacts of population growth and national migration policy, decrease of agricultural productivity, and other sector policies, especially the land tenure security needed to resolve forestry development issues). In spite of the major implementation issues having affected project development, the project was allowed to continue because of the importance of natural resource conservation in sustainable development of the country. In 1992, in an attempt to redress the situation, the Bank proposed a new pilot program to better integrate forest management and agroforestry into local economies. This program was, however, disrupted by the 1993/94 political events. A total of 11 supervision missions took place during the seven years of the project. On the Bank's side, two of the five task managers were forestry or livestock specialists. 6.3 Technical Assistance and Training. Considering the wide support allocated to the forestry sector by other bilateral donors, Switzerland in particular, project management was generally reluctant to recruit consultants from the credit allocation. Nonetheless, the performance of the expatriate technical assistant involved in the development of the ecological inventory12 was good due to both his technical capacity and his commitment to training and operational management. The quality of other TA varied considerably. Short-term consultancies utilized in the ecological inventory were in general fully satisfactory. The performance of technical assistants hired to implement and supervise financial accounting was, however, defective because of inadequate qualifications (Nyungwe) or unwillingness to live in remote areas (Gishwati). 6.4 Executing Agencies. Implementation responsibility was spread among three project units with compartmentalized mandates and competing interests. This situation led to inefficiency and waste of public resources. The fact that most project objectives were not achieved is due to the entrenched position of the Rwandese staff assigned to the project and the reluctance to adopt alternative participatory solutions. In retrospect, institutional strengthening was limited to vehicles, constructions and to providing some training and technical assistance. The strengthening of the MINAGRI was further limited by the low integration of the project components in its technical services. 12 He deserves a special mention for his initiative, his adaptability and his dedicated interest in forest ecology and participatory forest management problems.  16 Part 1: Project Reviewfrom the Bank 's Perspective 6.5 Donor Coordination. Coordination with other donor-funded projects was also limited. This situation led to duplication in the donor effort supporting the sector and compromised the potential development impact of external assistance. Although attempts were made to better coordinate investment activities in the Nyungwe forest, there was no overall consensus to guide investments and ensure maximum impact on protection. Communications between the Bank, the Government and/or other multi- and bilateral agencies was also disappointing on policy dialogue, institutional reforms and capacity building, and country-based coordination. In early 1993, two correlated policy initiatives were independently initiated at the same time under DGF responsibility: the formulation of a new sector policy framework supported by the Swiss Technical Assistance and the Tropical Forest Action Program under FAO financing. These initiatives narrowly focused on: (i) the value of natural forest, plantation and agroforestry as mere wood reserves; (ii) cross-sector policy and market failures; land tenure issues and land use conflicts; public participation in forest resource planning and management; and issues pertaining to private sector resources and skills; (iii) innovative institutional arrangements and financing mechanisms able to enhance private and community-based participation in forest management; and (iv) capacity building and institutional reforms given the precarious nature of the DGF, both in terms of its weak financial position and poor skills base. Unfortunately, little attention was paid to the comparative advantage of DGF, the private sector and local communities, given the unstable conditions of public expenditures to cover recurrent costs. F. Assessment of Outcome 7.1 For the past two decades, Rwanda's forest policy has focused on protecting its natural forest heritage on a hypothetical development of additional wood resources through fuelwood, buffer and industrial plantation. In the process, structural constraints and policy distortions that created negative consequences for forest resources were largely neglected. All grassroots involvement in forest and plantation management was disregarded. Moreover, the need to obtain land for plantation led to continuous conflicts with farmers, exacerbated by the Government's inability to pay for expropriations. Under these circumstances, the rural population was excluded from the decision-making concerning plantation management and from the benefits forests could yield. This could not help but contribute to forest destruction. The institutional weaknesses, both financial and technical, and the repressive attitudes of forest officers exacerbated these trends. 7.2 The project's failure demonstrates five things. First, in a densely populated country such as Rwanda, excluding rural population and private sector involvement in sector activities in favor of Government control ultimately undermined project sustainability and, moreover, created a dependence on an overstretched bureaucracy lacking the financial and staff resources required to maintain large plantation estates. State control also worked negatively by limiting the prospects of cost effective management behaviors and private entrepreneurs' involvement. As a result, plantations established under the first and the second projects became a continuous drain on public finance. Second, the 1985 forest policy flaws should have been assessed at the project  Part L: Project Reviewfrom the Bank's Perspective 1 7 design stage and significant reforms made more conducive to participation. Such reforms should have been introduced as conditionalities: decentralization of forest resource management; integration of resource utilization into the local economies; decentralization of management responsibility; trade reforms, pricing and tax incentives; etc. Third, ensuring the success of sustainable forest management, conservation and agroforestry related-strategies would have called for structural, institutional and organizational changes that ultimately would have led to repercussions on the role, responsibilities and functions of the DGF. In particular, the DGF's technical capacity should have been further strengthened and staff attitudes enhanced to support local forestry initiatives. Other points of concern might have included: realignment of mandate; institutional incentives; decentralization of the decision-making process; better capacity to deal with sector issues as well as better sector background information, particularly of mechanisms that govern participation and marketing of forest products. And fourth, disbursements should have targeted recurrent expenses related to forest management and conservation activities. G. Lessons Learned and Recommendations for Future Operations'3 8.1 Sustainable management of forest resources in Rwanda (including fuelwood, buffer zone and industrial plantation resources) will depend in large measure on the willingness of Rwanda's forest service to formulate a new integrated forest policy, correct structural constraints and policy distortions and obtain the administrative means and legislative tools to implement it. The project's set-backs contributed to a progressive understanding of the policy and institutional issues involved in forest resource conservation and management. The development of the new forest policy, under consideration since 1992 was, however, discontinued by the recent civil war. A synthesis of the above discussion suggests that the Bank and the Government should take note of the following key points when designing forestry and natural resource management operations in the future: * The need for a cross-sectoral approach which addresses constraints and policy distortions underlying the natural resource degradation andforest depletion. Policy reforms should be developed in response to sector issues raised by the major forest users and institutions in advance of the design of detailed investment operations. They should include economic policy reforms related to forest resource pricing and fiscal policies, and forest management reforms related to the nature of state intervention for resource allocation. Appropriate policy reforms should require a strong focus on sector work to help identify intersectoral policy flaws, and integrate 13 This chapter discusses lessons learned and recommendations from a pre-1994 situation. It is based on the hypothesis that the strong synergies and causality chains linking overpopulation, poverty, decreasing agricultural productivity and food security, lack of employment opportunities, and unadapted land tenure system were accelerating the process of environmental degradation and contributed to the violence that followed the death of the President on April 6, 1994.  18 Part 1: Project Review from the Bank's Perspective forestry objectives in a wider economic and environmental framework. The process of formulating forest policy should be open to the participation of all stakeholders; * Because Government's role in forestry development and environmental conservation will remain essential, greater attention needs to be paid to capacity building, particularly human resources development, and the long-termn horizon which this implies. To date, both, the DGF and the private sector are ill-prepared to face the new challenges of forestry development and conservation. Successful implementation of any project will, therefore, require strong externally-financed institutional support to DGF, local communities and the private sector that the Government is not in a position to offer.'4 Development and conservation of forest resources will also require drastic changes in the very structure of the DGF. As a result, there is an urgent need for a detailed institutional study'5 covering all sector institutions (including the DGF and other dependent agencies, training, research, etc.), in order to define a new governance framework and a long-term human resource development program to prepare the forestry staff to assume its new functions. With regard to the human impact of the recent civil war, this study should pay particular attention to existing capacities at the central and local level, cross-sectoral coordination, policy planning, programming and financial management capacities; * Full recognition needs also to be given to the opportunities for rural population and private sector participation in forestry, conservation and forest industry activities. This means fully taking into account stakeholders' views on project objectives and how they would be achieved starting with the preparation stage. By enhancing stakeholder participation, the aim would not only be to strengthen local ownership, but also reduce the risk of failure; o Full advantage should be taken of opportunities to make forestry (and conservation) activities self-sustainable on a financial basis, in order to avoid potential difficulties with counterpart funding. Because the country's financial situation will remain very weak in the future, alternative sources of funding should be identified (such as GEF, Debt for Nature Swap, endowments for nature conservation, etc.) so that forestry development and conservation recurrent costs need not depend entirely on Government counterpart funds; and * Improved Government/donor coordination in policy dialogue, project design and implementation is essential to maximizing potential development impacts. In particular, project implementation would benefit from: (i) a joint evaluation of policy distortion and agreement on policy reforms; (ii) donor agency participation in joint 14 Institutional strengthening activities required for project implementation should clearly be differentiated from the long-term institutional development program. 1 In-depth analysis of institutions' ability to respond to sector policy objectives should be a prerequisite for the design of all capacity building programs.  Part 1: Project Reviewfrom the Bank's Perspective 19 field reviews, monitoring and evaluation works; and (iii) increased exchange on thematic and technical matters. 8.2 Given the difficult political and economic situation prevailing in Rwanda at the time this report was written, considering rural poverty, rapid depletion of natural resources, natural forest resource management and agricultural sector development in an integrated fashion will be absolutely critical to promote sustainable development (the NEXUS approach). But in the Rwandan context of rapidly increasing population, limited resources, ethnic problems and lack of economic development opportunities, achieving both economic growth and reducing environmental degradation will require removing desincentives to sustainable management and developing specific sector reforms and projects to: secure land tenure, develop urban markets and access to them, enhance the distribution of agricultural inputs and appropriate technologies, improve access to health and education facilities (particularly for women) and family planning, develop off-farm and urban employment opportunities, and increase remittance from urban employment. Under such assumptions, integrated watershed management, agroforestry, and joint sustainable forest management are still promising ways to promote sustainable use of forest resources for the poor in Rwanda.   IMPLEMENTATION COMPLETION REPORT REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (Credit 1811-RW) PART II: PROJECT REVIEW FROM BORROWER'S PERSPECTIVE No comments have been received from the Borrower   IMPLEMENTATION COMPLETION REPORT REPUBLIC OF RWANDA SECOND INTEGRATED FORESTRY PROJECT (Credit 1811-RW) PART III: STATISTICAL INFORMATION Table 1: Summary of Assessment A. Achievement of Objectives Substantial Partial Negligible Not Applicable Macro Policies l Sector Policies Financial Objectives . V Institutional Development lV Physical Objectives Poverty Reduction V Gender Issues _ Environmental Objectives . Public Sector Management Private Sector Development B. Project Sustainability Likely Unlikely Uncertain State-owned Fuelwood Plantation V Industrial Plantation on the CZN Ridge V __ Buffer Plantation V Agroforestry V Extensive Livestock Development in Gishwati V Dairy/Meat Production in Gishwati V C. Bank Performance Highly Satisfactory Deficient Satisfactory Identification Preparation Assistance Appraisal Supervision V D. Borrower Performance Highly Satisfactory Deficient Satisfactory Preparation V Implementation V Covenant Compliance V Operation V E. Assessment of Outcome Highly Satisfactory Unsatisfactory Highly Satisfactory Unsatisfactory Nyungwe Biological Assessment Indigenous People - Resettlement V All other Components V  22 Part III. Statistical Information Table 2: Related Bank Loans/Credits Credit Credit Title Amount Purpose Year of Status Number (US$ Approval million) Cr. 0439-RW Mutara 1 3.8 Area Development 1973 Completed in 1979 Cr. 0656-RW Cinchona 1.8 Perennial Crops 1976 Completed in 1983 Cr. 0668-RW BGM 1 14.0 Area Development 1976 Completed in 1982 Cr. 1039-RW GBK (Agro-Sylvo 1) 21.0 Forestry/Livestock 1980 Completed in 1987 Cr. 1283-RW BGM II 15.0 Area Development 1982 Completed in 1991 Cr. 1546-RW Agricultural Research 11.5 Research 1985 Completed in 1991 Cr. 1669-RW Gitarama 12.7 Extension and related services 1986 Completed in 1991 Cr. 2026-RW Agricultural Services 19.9 Extension and related services 1989 In progress Cr. 2547-RW Agricultural Research II 15.0 Research 1993 Withdrawn in 1995  Part III: Statistical Information 23 Table 3: Project Timetable Steps in Project Cycle Date Planned Revised Date Date Actual Identification (EPS) March 1985 Preparation October 1985 Appraisal March 1986 Post-Appraisal October 1986 Negotiations April 1987 Board Presentation _ _ May 1987 Signing August 1987 Effectiveness December 1987 March 1988 October 1988 Credit Closing June 1993 June 1994 June 1994 Project Completion December 1994 June 1995 March 1995  24 Part 11.1 Statistical Information Table 4: Credit Disbursements A. Cumulative Estimated and Actual (US$'OOO) FY87 FY88 FY89 FY90 FY91 FY92 FY93 FY94 Appraisal Estimate 1,100 2,250 4,700 7,750 10,950 13,500 14,100 14,100 Actual16 829.5 1,895.0 3,371.3 4,341.6 5,069.2 6,692.9 6,776.7 Actual as % of Estimate n.a. 36.8 40.3 43.4 39.6 37.5 47.4 47.1 Date of Final Disbursement: July 1, 1994. B. Estimated and Actual Quarterly Disbursements 900 800 700 __600 500 cl 400 - 300 200 | a a a a a a a a a ~~~~a a a aY a a a |~~~~~- cos co a) a> 0 0 _ N N e M st V 1~~~C 0 0 0 0 a) 0 a) 0) 0 0 0 0 0) 0) 0 1 | i I Estirrrated Dsbursement*Actual Dsbursenent 16 Including unrecovered Special Accounts  Part III: Statistical Information 25 C. Cumulative Estimated and Actual Disbursements 16000.00 l 14000.00 12000.00 10000.00 4ft~