Rwanda - Agricultural Services Project

Document of The World Bank FOR OFFICIAL USE ONLY Z~~~a &s R- RS- Report No. 7599-RW STAFF APPRAISAL REPORT RUANDA AGRICULTJRAL SERVICES PROJECT MAY 8, 1989 Agricultural Operations Division South-Central and Indian Ocean Department Africa Region This docweut has a residted dlsubudon and may be used by redslnt only in te perfonnaDof thdwr a * Ih v ' s a I  CURNCr EOUiv S Currency Unit - Rwandese Franc (RF) US $1.00 - RF 77 RF 100 - US $1.30 US $1.29 - 1 SDR The U.S. dollar/Rwandese franc rate is subject to change. Conversions in the Staff Appraisal Report were made at US $1.00 - RF 77, the average exchange rate at the time of appraisal in October 1988. WEIGHTS AND MEASURES Metric British/US Equivalents 1 meter (a) 3.28 feet (ft.) 1 cubic meter (m3) - 35.3 cubic feet 1 hectare (ha) - 100 ares 2.47 acres 1 kilometer (km) - 0.624 mile 1 square kilometer (km2) - 0.386 square mile (sq. m.) 1 kilogram (kg) - 2.20 pounds (lb.) 1 liter (1) D 0.26 U.S. gallon (gal.) 0.22 Imperial gallon (imp. gal.) 1 metric ton (m t) 2,204 pounds (lb.) GOVERNMENT OF RWANDk FISCAL YEAR January 1 to December 31  FOR OFFICIAL USE ONLY GLOSSARY OP ABSREVIATIONS ASP - Agricultural Services Project BEA - BSJic Exten ion Agent BGM - Bugesera-GisAka-Migongo Project BNR - Banque Nationals du Rwanda BP a Banques Populaires (also UPB Union des Banques Populaires or Cooperative Savings Bank) BRD - Banque Rwandaise do Doveloppement (Rwandan Development Bank) CCDFP - Centre Comual de Developpement et de Formation Permanente (Centers for Coumunal Development and Permanent Trainig of the Ministry of the Intorior) CFRC-IWACU - Centre de Formation *t de Recherche en Cooperatives (Cooperative Research and Training Center an NGO) COOPIBO - Compagnons Batisseurs (an NGO involved in development of cooperatives) DCE - Direction de Controle et Evaluation (Directorate for Control end Evaluation in the Ministry of Agriculture) FAO/CP a Food and A4riculture Organlzation Cooperative Program with the World Bank, Rome PD Fonctionnalre Dirigeant (higher level civil servant appointed as project administrator) GOR - Government of Rwanda IBRD - International Bank for Reconstruction and Development IDA - International Development Association IFAD - International Fund for Agricultural Development (Rome) IRD - Integrated Rural Development ISAR a Institut des Sciences Agronomiques du Rwandi (Agricultural Research Institute of Rwanda) KORA - An NGO whose name means 'to work MINAGRI - MiNestere d l'Agriculture, de l'Elevage et des For;ts (Ministry of Agriculture, Livestock and Forests) MINFIN - MInistire des Finances (Ministry of Finance) MINITRAPEE - Minist4re des Travaux Publics, Energle et Cau (Ministry of Public Works, Energy and Water Resources) NCC - National Coordination Comnittee NGO - Non-Governmental Organisation NSC - National Seed Comittee OCIR-Cafe - Office des Cultures Industrielles du Rwanda Cafe (Office for Industrial Crops Rwanda - Coffee) OCIR-The - Office for Industrial Crops-Tea OPROVIA - Office National pour le Developpement et 1B Commercialisation des Prodults Vivriers et de 1- Production Animal* (National Agency for Development and Marketing of Foodcrops and Animal Products) OVAPAM - Office pour la Valorisation des Produits Agricoles de Mutara (Offiee for the Development of Agricultural Products in Mutara) PCCV - Projet Cafe et Cultures Vivriers (Lake Kivu Coffee and Fooderop Project) SESA - Service des enqu;tes et des Statistiques Agricoles (Agricultural Statistics Service) SMS - Subject Matter Specialist SSS - Service des Semo'nces Selectionnees (Selected Seed Service) TRAFIPRO m Cooperative de Vente et d'Achat Travail-Fidelite-Progres (Transport and Marketing Cooperative) T&V - Training and Visit Method of Extension UNDP u United Nations Development Prograame U.O - Unlt; Operationelle (Operational Unit) USAID * United States Agency for International Development 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 authorizatiorm  RWA AGRICULTURAL SERVICES PROJECT TABLE OF CONTENTS CREDIT AND PROTECT SUMOARY............................................ i A. Itouto B. Sconomic and Agricu_tural Setting ........................... 2 C. Government's Sectorael Objectives and Strttegy 3 D. IDA,s Assistsnce and S4rategy.............................. . 4 - The 1GC II and PCCV Projets ............................. 4 - The Gitar_ma Agricultural Production and MINAGRI Institutional Stre nthening Project .............0....... 5 II. AGRICULTURAL SERVICES.................*......................... 8 A. The HMIistry of Agriculture, Livestock and Forests (MI&GRI) .................................................. 8 |~~~~ Bx.nRseacion.......................................................... 10 C. Extesin ........10 - Structure ................................................ 10 - Methodology and Content .................................. 10 - Donor-Financed Projects ................................. 11 D. Agricultural Inputs ...................................... 12 E. Agricultural Credit and Finance ............................. 13 - Government Funding of Agriculture ........................ 14 F. Markets and Prices .......................................... 15 G. Cooperatives and Farmer Groups .............................. 15 This report is based on the findings of a pre-appraisal mission which visited Rwanda in July 1988. The mission comprised of Mrs. V. Okonjo-Iweala, Messrs. D. Aw and S. Sissoko (IDA), Messrs. T. Marchan., A. Edery, Y. Clouet and J. MacAlister (Consultants). An appraisal mission comprised of Mrs. Okonjo-Iweala, Messrs. Aw and Sissoko visited Rwanda in September/October 1988. Ms. Robb provided secretarial support.  TABLE OF COWTENTS Pafte III. THE PROM4ECT. o. .. ..................................... 16 A* Rationaleo .a................. 16 B. Summary Project Description............ ................... .. 16 C. Geographical Coverage ....................... 17 D. Detailed Fetrs................................18 - Extension Structure ..................... 00.................. 18 - Methodology and Content of Extension ..................... 21 - Adaptive and On-farm Research ...... ... 24 - Monitoring and Evaluation (M&E) .......................... 25 - Strengthening of Seed Production......................... 27 - Prouotion of Agricultural Marketing, Input Supply, and Cooperative Development ................................... 28 - Improving Service Delivery to Wom en ...................... 30 - Training ................................................. 31 - Technical Assistance and Studies......................... 32 S. Project Costs .......... 321 F. F4U&nc4nge ........%*%** 34 H. Procurement ................................................. 35 R. Dsusmns.................................. 36 I. Financial Management, Accounts and Audit .................... 38 IV. ORGANIZATION AND ..................................... ... 39 A. ProJect Management ................... 39 B* Annual Work Plan 40 C. Review and Reporting Reqr nreme.. . 41 D. External Project Monitoring and Evaluation.................. 41 V. BENEFITS, JUSTIFICATION AND RIS R S ........ .......... 42 A. Benefits and Justification........................ . ........ 42 - Financial Benefits and Sustainability... 43 B. Economic Analysis........................................... 44 - Sensitivity A nlss4 C.en ivt Rnlysks .... .......... ....................... 44 D. Environmental Impact ......................................... 45 VI* AGREEMENTS RAHD.................................. 46  TABLE OF CONTENTS ANNEXES e ANNEX 1 - IMPLEMENTATION SCUEDULE..................................... 50 ANNEX 2 - Table 1 - Staffing Plan..................................... 51 Table 2 - Schedule of Disbursements......................... 52 Table 3 - MIM&GRI's Proposals for Donor Support to Agricultural Services Structure.. ............... 53 Table 4 - lMNAGRI Budget ................................... 59 Table 5 - The Agricultural Services and Gitarama Projects Basic Structure/Staffing Compared............... 60 Table 6 - Roads and Bridges to be Rehabilitated or Constructed in Cyangugu Prefecture .............. 61 Table 7 - Proposed Training Progrsm........................ 62 Table 8 - Project Prefectures, Sous Prefectures (U. Os) and Cemns..............................65 ANNEX 3 - PROJECT COST TABLES......................................... 66 Table 1 - Project Components by Year (Showing Taxes & FE).. 66 Table 2 - Project Components by Year Including Contingencies.................................. 67 Table 3 - Susmary Accounts Cost Sumay.................... 68 Table 4 - Summary Accounts by Year......................... 69 Table 5 - Suinary Accounts by Year Including Contingencies.................................. 70 Table 6 - Suumary Account by Project Componento............. 71 Table 7 - Breakdown of Summary Accountsc.... 72 Table 8 - Financing Plan by Project Components..ents******* 73 Table 9 - Financing Plan by Summary Accounts.ounts*........ 74 Table 10 - Financing Plan by Disbursement Categ gory**%****** 75 Table 11 - Disbursement by Financier per Semester..... ..... 76 ANNEX 4 - Table 1 - Recurrent Costs of Extension/Adaptive Researclt Projects (1987) in ASP Project Area......... * 77 Table 2(a) - Economic Analysis............................ . 78 Table 2(b) - Economic Analysis with Explanations............ 80 Table 2(c) - Economic Analysis with Explanations ............ 81 ANNEX 5 - Chart 1 - Present Organization of the Ministry of Agriculture, Livestock and Forests.............. 82 Chart 2 - Organization Agricultural Services Proect ............... 83 MAPS IBRD 21291 - Project Area IBRD 21406 - Donor-financed projects with Extension Components  AGRICULTURAL SERVICES PROJECT Credit and Proiect Summarv Borrowers Government of Rwanda Beneficiaries: Ministry of Agriculture, Livestock and Forests, Farmers in Cyangugu, Gisenyi, Kibuye, Kigali and Kibungo prefectures Credit hmounts SDR 15.5 million (US$ 19.9 million equivalent) Torms: Standard IDA terms Prolect Descrintion: Project objectives are to: (a) strengthen adaptive and on- farm research; (b) improve the delivery of axtension services to farmers with a particular emphasis on reaching women farmers; (c) strengthen the production of selected seed and (d) promote greater private sector participation in the marketing of agricultural inputs and produce. The project would be implemented over a five year period and would comprise: (a) reorganization and streamlining of extension services in all or parts of f4ve prefectures including implementation of a new structure integrating extension, adaptive research and monitoring and evaluation. The structure has been tested in the IDA-financed pilot Gitarama Agricultural Production and MINAGRI Institutional Strengthening Project (Cr. 1669-RW). The project would upgrade extension staff through training and support improvements in extension methodology and content; (b) strengthening of adaptive and on-farm research through better diagnosis of problems at the farm level, inclusion of farmers and local authorities in problem identification and solving, and implementation of on-farm and multi-local trials of new techniques and technologies provided by research; (c) support for monitoring and evaluation through the implementation of a decentralized management information system; (d) strengthening of seed production and distribution through support to the Selected Seed Service to improve seed quality control and certification, reactivation of the National Seeds Committee for supervision of seed regulation; improved cost reccvery for seed production, and rationalization of production including improved management of ex-BGMIPCCV project se..d farms. The project would support the development of a multipurpose farm for seed multiplication, and adapt .ve research trials in Bugarama, Cyangugu prefecture. The farm would also serve as a base to encourage the contracting of  ii small farmers as outgrowers for seed multiplication purposes; (e) promotion of private sector participation in agricultural input supply and marketing through assistance to cooperatives and private enterprises to procure inputs, obtain credit and improve enterprise management. NGOS and the Banques Populaires would be contracted to provide the needed services. Benefits and 2isks: Project benefits are expected to be: (a) increased productiv- ity and output for an estimated 168,000 farm families, and, as a result better nutrition and incomes. An additional 400,000 farm families will also benefit from improved supervision and coo-dination of prefecture wide agricultural activities by the prefecture teams to be financed under the project; (b) a strengthened field service expected to provide the framework for future coordinated donor support to extension; and (c) savings in the running costs of a more streamlined extension and adaptive research service. The main project risk is the limited availability of university trained personnel for the adaptive research service and the higher echelons of the extension service. To minimize this risk, a great deal of emphasis will be placed on continuous on-the- job training to upgrade post-secondary staff and render them capable of performing specialized tasks. Estimated Proiect Costs: Local Foreisn Total <---------US$ million--------> Extension, Adaptive Research and Monitoring & Evaluation Services (including support investments in Livestock and Forestry) 11.0 6.0 17.0 Strengthening of Seed Production and Distribution 1.0 0.8 1.8 Promotion of Private Sector participation in Marketing, Input Supply, and Cooperative Develop- ment 2.3 0.9 3.2 Central Administration 0.6 1.2 1.8 Studies 0.4 0.1 0.5 TOTAL BASELINE COSTS 15.3 9.0 24.3 Physical Contingencies 1.0 0.9 1.9 Price Contingencies 2.6 1.3 3.9 TOTAL PROJECT COSTS1' 18.9 11.2 30.1 1/ Includes direct taxes of about US$ 1.3 million.  iii Financing Plant Local Foreitn Total -------USS million--------> Government2l 9.0 1.2 10.2 IDA 9.9 10.0 19.9 TOTAL 18.9 11.2 30.1 Estimated Disbursemente: (US$ million) IDA Fiscal Year 1990 1991 1992 1993 1994 1995 1996 Annual 0.4 1.9 3.9 4.8 4.3 3.8 0.8 Cumulative 0.4 2.3 6.2 11.0 15.3 19.1 19.9 Rate of Returns Based on an indicative calculation of the order of magnitude of benefits needed to produce a 15% ERR and the likelihood of attaining these benefits. If the net incremental bernefits due to improved agricultural services were to reach RF 1020 million (US$ 13 million) per year by project year 15 i.e. 9% above the estimated "without project case" discounted economic benefits would equal discountea costs and the ERR would be 15%. Conservative adoption rates of new cultivation practices (an average 4% yearly) and modest production increases (11 annually would lead to a 15% ERR. Evidence from Bank financed extension projects in other sub-saharan African countries (e.g. Togo and Central African Republic), indicate adoption rates for new extension themes of 10-50% among farmers sampled. 2/ Includes the US$ 1.3 million in taxes.  RWANDA AGRCULTURL SERVICES S I. BACRGROUND A. Introduction 1.01 The project would support the reorganization of extension services on a national scale beginning in five of Rwanda's ten prefectures. This reorganization would include specific measures to improve extension delivery to women farmers responsible for the majority of agricultural tasks in Rwanda. The project would develop an adaptive research structure and strengthen the production and distribution of improved seed. An extension monitoring and evaluation system would be developed and increased private sector participation in input supply and produce marketing promoted. 1.02 The proposed project would build on lessons learned in the pilot Gitarama Agricultural Production project (Cr. 1669-RW) designed to test a new approach to the delivery of agricultural services in the Gitarama prefecture. It would consolidate and improve on extension work undertaken in two IDA-financed projects, the Lake Kivu Coffee and Fooderop Production (PCCV) Project (Cr. 1126- RW) and the second Bugesera-Gisaka-Migongo (BGM II) project (Cr. 1283-RW). The PCCV project closed June 30, 1988 and the BGM project is due to close December 31, 1989. The project will be a continuation of Government of Rwanda (GOR) and IDA's attempts to build national agricultural services as a response to emerging constraints encountered with integrated rural development (IRD) projects. By developing needed infrastructure, the latter have been successful in laying the foundations for future agricultural development. However, they have, on balance, proven costly in manpower and financial terms leading to doubts about GOR's capacity to sustain the investments in the long-term. Furthermore, they have led to the creation of parallel institutional structures that are difficult to coordinate, and by channelling resources to these structures, have contributed to a weakening of agricultural management capacity. The development of national asricultural services aims to streamline service delivery to make it cost effective, easier to sustain in the long-term, and capable of responding to the dif.icult challenge of increasing productivity in the agriculture sector. The Agricultural Research project (Cr. 1546-RW), which became effective in Ncvember 1985, was the first in this series of institution building p-ojects, followed by the Gitarama Agricultural ProductionJMINAGRI Institutional strengthening project. The proposed project would modify and expand, on a national scale, the Gitarama extension and agricultural services model. 1.03 The idea of a national Agricultural Services project was first discussed at a joint GOR-IDA programming workshop held in Mey 1986. GOR and IDA recog;aized that the BGM and PCCV projects would soon close and there would be a neei to continue the extension activities begun under those projects. It was accepted that continuation of extension activities could no longer be on the basis of the dispersed efforts that characterized Bank financed and other donor-supported IRD projects but would have to be within the framework of a national extension policy and structure along the lines of Gitarama. It was also recognized that the lack of well adapted technological packages and  - 2 - extension themes, and the limited use of modern inputs would constitute serious bottlenecks for a meaningful extension effort. An Agricultural Services project was proposed to assist in tackling these problems. An identification mission visited Rwanda in December 1986 to initiate project prepLration. The project was prepared by MINAGRI staff with some assistance from FAO/CP and follow-up by IDA missions. Preparation was completed in May 1988. IDA preappraised the project in July 1988 with appraisal in October 1988. This report is based on the findings of the IDA missions. B. Economic and Agricultural SettinR 1.04 With an estimated population of 6.0 million growing at a rapid annual rate of 3.7%, Rwanda is one of the few African countries with a record of sound economic management in the post-independence period. As a result, economic distortions have been less severe than in most other countries. According to official statistics GDP growth has been strong since the early 1970s averaging about 5% per annum (if the drought year of 1984 is excluded). However, economic growth appears to be slowing down and estimates for 1987 indicate a 2% decline in real GDP compared to 1986 levels. Agriculture sector management has been zelatively non-interventionist and the good policy environment contributed to a steady 5.82 per annuLm growth rate between 1974-1980. From 1980-83 agricultural output growth slowed to an average annual rate of about 3.5%. Since then, it appears to have slowed further averaging only about 2% per annum (again excluding 1984). Agriculture is the mainstay of the economy contributing about 80% of the total value of exports and 40% of GDP at current prices. As much as 957 of the population lives in the rural areas. Per capita GDP (1986) is about US$ 298 of which US$ 130 is attributable to agriculture. 1.05 Although statistics on food output are questionable, food production appears to have kept up with population growth up until 1985. In 1986 the growth rate of the fooderop sector fell to about 3%, and the estimate for 1987 is 2% well below the population growth rate of 3.7%. Commercial cereal imports have increased substantially within the past twelve years from 3000 mt in 1974 to 24000 mt in 1986 partly as a result of changing tastes. Production of the main export crops (coffee and tea) is cf vital importance as a source of foreign exchange and rural cash income. After an extended period of growth, value added in the export crop sector declined in 1986 as a result of falling world market prices. Fux hermore, there have been recent adverse developments in the exchange rate. The Rwandese franc has appreciated by about 30% relative to the carrencies of its major trading partners and this is threatening the competitivity of agricultural exports. The Government also appears to be moving toward more intervention in the setting of agricultural input prices and producer prices for food crops. These latter have generally been in the favor of farmers, sometimes even excessively so. So far, however, there has been little enforcement of such prices on the market. The Government of Rwanda has a reputation for realistic economic management and there are indicatioz.s that it is willing to adjust its economic policies to take account of present realities and to continue its policy of little or no enforcement of minimum prices. 1.06 The agriculture sector faces serious challenges which have an adverse impact on the pace of agricultural output growth. The most important is the pressure of population on scarce land reso.rces. The country's population  density of 400 per kmn of arable land is the highest in continental Africa. As evidence of this pressure farm size is diminishing and holdings Increasingly fragmented. Over half the one million farm holdings are less than one hectare. Fallow land is in limited sLpply and there is overgrazing by livestock. Farmers have so far managed to increase overall productio;J through area expansion and increased cropping intensities (double cropplng and intercropping) but with limited use of modern inputs. All this poses a threat to soil fertility,and there is indeed evidence that yields for most crops are at best stagnant. With the possibility of increasing agricultural output through area expansion and increased cropping intensity coming to an end, indications are that future increases must come from productivity gains. To achieve this goal the deivlopment of relevant agricultural technology and its dissemination should be facilitated through better agricultural support services. Agricultural exten- sion, and adaptive research are central to achieving results. The poterntial effectiveness of these services would have to be related directly to the provision of other ingredients required to improve farm productivity. These include a reliable supply of inputs and access to credit when needed. On the output side, greater attention to markets would be required to ensure that a greater part of the increased production is sold and cash is available for the purchase of inputs. C. Government's Sectoral Objectives and Strategy 1.07 The GOR has three basic objectives in the agriculture sector. In order of importance they are: food self-sufficiency, development of import substituting crops, and export diversification. The development of the foodcrop sector has received overwhelming attention (most donor and government projects are in this area) while export crop development has been somewhat neglected. Government's objectives translate concretely into development priorities which are re-articulated in the preparatory documents of the fourth five-year plan (1989-93) as follows: (i) intensification of crop and livestock production; (ii) soil conservation and improvements in soil fertility; (iii) improved management and rational exploitation of the little remaining unused agricultural land; (iv) regional specialization in production as a function of agro-ecological iotential; (v) implementation of agricultural inrerventions within the framework of a sectoral or crop specific approach; and (vi) progressive decentralization of the structures for intervention to the prefecture, sous-prefecture and commune levels. To achieve its sectoral objectives the Government has relied extensively on donor-financed integrated rural development projects many of which have helped carry out government programs to conserve soils, improve farming methods and develop new crop technology. There are over 60 donor-fVaanced projects in the agriculture sector covering more than 100 of the count- 's 143 communes. The IRD projects have performed well in the area of soil conservation, road and infrastructure development and this has no doubt contributed to good agricultural sector performance. However, they have had less impact on improving crop yields, are expensive, and are increasingly difficult for the understaffed Ministry of Agriculture to monitor, supervise and coordinate. 1.08 Faced with the need for productivity increases, and the lack of an effective strategy to achieve them, the GOR has since 1986 held national brainstorming sessions to develop policies to reorient agricultural projects, and to strengthen adaptive research, extension and inputs supply. The means for  - 4 - reaching women more effectively through the extension system have also been debated and discussed. An important concern has been the need to make agricultural development a bottom-up process which reaches out and includes farmers in problem definition and solving rather than talking down to them. This concern is based on the recent Government decision to decentralize development down to the communes (the level of local administration) to involve local populations more effectively. The culmination of these debates is a series of decisions to stress "recherche-developpement" (adaptive and on-farm research), to reorganize extension to respond better to the needs and realities at the farm level, to promote private sector participation in marketing of inputs and produc-ea, and to rationalize the patchwork of donor-assisted projects by developing a policy framework and a sustainable national structure within which these projects could eventually be integrated. The Agricultural Services project (ASP) would enable Government to begin Implementing this new approach. D. IDA's Assistance and Strateav 1.09 Since 1973 IDA lending has amounted to US$ 118.3 million for ten operations, four of which are ongoing. The operations are as follows: (a) four area development projects in the eastern Savannah zone (US$ 42.2 million); (b) two area development projects in the Lake Kivu and Zaire-Nile Divide zones (US$36 million); a Cinchona Development project (US$ 1.8 million); an Agricultural Research Project (US$ 11.5 million); a pilot extension and institutional develop- ment project--the Gitarama Agricultural Production and MINAGRI Institutional strengthening project (US$ 12.7 million) and a second Integrated Forestry Project (US$ 14.1 million). 1.10 Completion reports have been prepared for the first Mutara Agricultural Development project (Cr. 439-RW), the BGM phase 1 project (Cr. 668- RW) and the Cinchona project (Cr. 656-RW). The experience from these and other projects highlight a number of issues affecting project performance namely: a) the scarcity of technical messages for agricultural improvement at the farm level; b) the complexity of Rwanda's farming systems and insufficient knowledge of these systems; c) weak links between extension and research; d) project complexity ano excessive diversity of project institutional and organizational systems; el lack of experie-ced Rwandese managers and a weak technical ministry; f) inadequate monitoring of project progress; g) inadequate attention to operating budget implications of projects; h) unclear subsector policies and investment priorities and poorly coordinated donor support. The BGM II and PCCV Proiects 1.11 The BGM II and PCCV projects (Credits 1283 and 1126 RW respectively) were conceived as multi-component area development projects focusing on extension and, in the case of BGM II in particular, including activities in nutrition, family planning and infrastructure (road) development. The two projects have been successful in extending themes on erosion control (up to 80% of the PCCV area now has erosion control devices compared to 32% before the project), interplanting of banana trees with beans to extend area cropped, and introduction of new crops such as soybeans. Coffee quality has improved in the PCCV area with the introduction of better cultivation practices, construction of more pulping  centers and provision of new pulping machines. The BGM II project has successfully organized marketing cooperatives and instituted cost recovery for forest nursery plants, and some veterinary products. 1.12 The principal constraint has been the paucity of adaptive and on- farm research and this has manifested itself in the lack of new themes to extend and the monotonous repetition of old themes. There has also been limited involvement of the population in problem identification, solution and follow- up. The links between ISAR, the national agricultural research institute, and extension have been weak. In addition, the introduction of new varieties of improved" seeds has not been a success because of poor quality control of seeds multiplied by the projects and the weakness of certification services. The T&V system of extension appears to have been applied rather rigidly at the initial stages of the projects. Contact farmers were visited rather than groups and many people lost interest when the same farmers were visited repeatedly. Adjustments were subsequently made in the course of project implementation. Project management was too centralized in Kigali and proved unworkable. This was resolved by decentralizing responsibility to field level staff. The Gitarama Agricultural Production and MINAGRI Institutional Strenftthenint Prolect 1.13 The Gitarama/MINAGRI project (Cr. 1669-RW), which became effective in September 1986, was prepared as a vehicle to test a new approach to the provision of agricultural services (paras. 1.02 and 1.08) before extending this nationally. It was also designed to respond to some of the problems identified in past and on-going Bank projects in extension. The objective of the Gitarama Agricultural Production component is to use the Gitarama prefecture as a testing ground for a model of agricultural extension, adaptive and on-farm research, and private sector promotion which would be more cost-effective, replicable at the national level and financially and institutionally sustainable. 1.14 The Gitarama extension model is essentially T&V adapted to local social and institutional structures with an extension organization paralleling the presenz administrative structure. At the base is the extension agent responsible for an average 750 families. The agent is assisted and supervised by a commune level agricultural officer supported by other commune level technicians essentially, a livestock officer and a forestry officer. At the sous-prefecture level is the operational unit (Unite operationelle) charged with implementing adaptive research, preparing extension themes, and undertaking training of extension personnel. At the prefecture level, the prefecture agricultural coordinator assisted by a deputy and chief accountant is responsible for coordination of agricultural activities in the prefecture. He reports to the head of the Agricultural Promotion Division in MINAGRI who also doubles as the ministry appointed project administrator or "fonctionnaire dirigeant". Under the Gitarama model extension agents work within the framework of a rigorous time schedule, including farm visits, training and reporting and cover all crop, livestock and agro-forestry activities. This is an important departure from the past when there were different agents for the different types of activities. When fully implemented, the Gitarama model should result in a 50% reduction in the number of basic extension agents. An important service begun by the project is the promotion of private sector activities in agricultural input supply and marketing. This is accomplished with the assistance of NGOS contracted to  provide training to cooperatives and private individuals in management techniques, and training to artisans in the manufacture of agricultural tools; and the Banques Populaires for provision of credit and mobilization of savings. 1.15 The MINAGRI Institutional Strengthening component, implementation of which is yet to begin, is designed to reinforce the Ministry's sector management capability. After effectiveness, GOR indicated its wish to expand the scope of the institutional strengthening to include all central MINAGRI services rather than the secretariat general only as was planned in the component. The Government also indicated its desire to seek grant funding to substitute for IDA's participation. Negotiations have been on-going with UNDP a co-financier in the project to finance the entire component. These are nearing conclusion and implementation may begin by mid-1989. IDA is awaiting GOR's request for a reallocation of component funds. 1.16 A mid-term evaluation of the Gitarama Agricultural Production component has recently been concluded. 1/ The new agricultural services structure appears to have improved extension m.thodology and approach. "Teething" problems of relationships and linkages ' hin the structure remain. The major findings are: (a) the apparent suitability (effectiveness) of the agricultural, services structure compared to institutional structures of previous extensionlagricultural services projects; (b) the success of the promotion component in cooperative training, marketing, extension of agricultural credit, and mobilization of rural savings; (c) the failure of the adaptive and on-farm research component due to an unclear definition of tasks, poor coordination with ISAR, poor linkages with extension personnel, and failure to involve farmers in problem definition and solving. A's a result there has been a continued paucity of new area-specific extension messages, and repetition of general extension themes; (d) the relative ineffectiveness of the monitoring and evaluation component due to a lack of understanding by project staff of the objectives of monitoring and evaluation, and a failure to adequately define the modalities for its implementation. In addition, the extension approach is still too top down and could clearly be further improved; some extension themes related principally to livestock and coffee have received less than adequate attention; the use of contact groups for extension sessions is too narrow and co0tld be broadened and improved to include groups with like interests such as womea, young people with little or no land, small farmers, large farmers with livestock, and local officials should be better informed on and more clearly involved in the project. Finally, the project's financial structure appears too centralized, the agricultural services structure still appears too demanding in terms of vehicles and equipment, and the distribution of higher level personnel within the structure appears to be unbalanced. 1.17 The evaluation provides little quantitative information on farmer participation and adoption rates largely because the monitoring and evaluation system which should have regularly collected this information failed. Indicative figures provided are as follows: a range of 44% - 67% of contact group chiefs regularly attend extension sessions. For other farmers (known as volunteers) the figures range from 15% - 98%. The difference was explained, during appraisal 1/ Project Agricole de Gitarama Evaluation i mi-terme: Document Provisoire Novembre 1988. Note that the document summarizing the findings of the evaluation mission is under review by MINAGRI and by IDA.  of the ASP, as being due to nearness to a town or commercial center. Groups living near such areas had important non-farm activities and participated less in extension sessions. Those living farther away were keen participants. The evaluation indicates that an average 6X of the population of the project area has been reached of which about 1% women. (It is not clear whether the 62 also indicates adoption of themes extended.) Partial information for the Karama Uftite Operationalle gives a further idea of participation and adoption rates. 21 The information is preliminary and it is still too early to measure repeat adoption rates. About 10 of the 49,760 farm families in the U.O had been reached by the extension service in that cropping season. Of this number 2,986 or 62 participate actively in extension aess ions. With regard to the adoption of themes, about 52 of farm families adopted themes related to the planting of soil fixing plants in anti-erosion ditches and 6X adopted the composting theme. This seems to indicate that once families are active in extension sessions they tend to adopt themes wholly or in part. Because there were not enough vegetable seeds or cuttings of the popular K51 sweet potato variety to distribute, onl!* 32 of families got the vegetable seeds and 1X the K51 cuttings. This underscores the need to have sufficient supplies of needed inputs to back up extension training sessions. 1.18 The Gitarama, BGM and PCCV experiences have been factored into the preparation of the Agricultural Services project which is a key instrument in IDA's lend'ig strategy. Attempts have been made to incorporate lessons learned from other projects. First, emphasis will be put on the development of adaptive and on-farm research an'd the strenztheninz of links with ISAR to tive momentum to the development oc new messaRes and technical recommendations. Second, linkages within the extension structure will be better defined, extension monitor.ng and evaluation will be better explained and emphasis will be placed on decentralization of responsibilities to the field level. Finally, the successes of previous projects in cost recovery, development of cooperative marketing, and other promotion activities will be continued and expanded. 1.19 Recent IDA financed projects in research and extension including the proposed operation, are the outcome of a long and concerted sector dialogue with GOR, and have benefitted from an active participation in their preparation and supervision by MINAGRI, based on full agreement between GOR and IDA on means to take into account the kessons learned from earlier efforts. However, the sector dialogue has so far produced fewer results in other areas of Rwanda's long-term agricultural strategy which are equally important. There is as yet little discernible interest in GOR to initiate a major push to improve both quantity and quality of existing export crops and/or to diversify into other exports. With present fooderop production largely consumed on the farm, and with small domestic markets, increased exports are important to generate the additional cash income needed b,y farmers to buy the inputs required for productivity improvement. Also, the expansion of non-farm employment in rural areas to decongest the agricultural sector is urgently needed. These elements of the sector dialogue, which would run parallel with discussions with GOR on macro-economic policies, will be emphasized in the future. Also to be emphasized is the dialogue on agricultural pricing which has been ongoing and is beginning to yield positive results (para 2.18). 2/ Provided to the project appraisal team for the first 1988 cropping season.  - 8 - II. AGRICULTURAL SERVICES A. The Ministry of Atriculture. Livestock and Forests (MINAGRI) 2.01 With a staff of 2,500 of whom 80S are in the field, MINAGRI's mandate is to manage the agriculture, livestock, forestry, land and water resources of Rwanda. The ministry comprises four technical directorates (Directions Generales), and a secretariat (Secretariat General) (Annex 5, chart 1). The technical directorates are: a) Agricultural Production which covers agricultural extension,selected seeds, plant protection, food crops, industrial crops and land distribution; b) Livestock comprising animal production and health, fisheries, small stock, beekeeping; c) Forestry covering natural forest protection and management, communal forestry, agro-forestry, industrial forestry, reforestation; d) Rural Engineering and Soil Conservation comprising irrigation, farm tools and mechanization, topography and soil analysis, erosion control and soil conservation. The Secretariat General is responsible for coordination and supervision of the technical directorates, monitoring and evaluation of agricultural projects, sector studies and planning, ministry management and administration. MINAGRI is also the supervising ministry for several services and parastatals namely the Selected Seed Service (SSS), the Agricultural Surveyi and Statistics Service (Service des Enquetes et Statistiques Agricoles SESA), several agricultural training centers, and four parastatals in charge of coffee (OCIR-Cafe), foodcrop and livestock marketing (OPROVIA), development of the Mutara region (OVAPAM), and national agricultural research (ISAR). 2.02 MINAGRI faces several constraints the most important of which is the insufficiency of trained and qualified staff. The most important gaps are in sector planning and studies, financial-and economic analysis, policy analysis, project monitoring and evaluation, and sector management. At the technical level the need for foresters is acute. As a result of the inadequacy of trained staff sector analysis is weak, and investment planning, programming and budgeting is inadequate. This is compounded by the overall inadequacy of Rwanda's budgetary process. Monitoring and evaluation of donor-financed projects is very weak. MINAGRI has little idea of how much is spent in these projects or of the impact of expenditures on agricultural production and incomes. Compounding MINAGRI's problems are the present inadequate physical facilities, office equipment and logistical support. 2.03 MINAGRI has lecome increasingly aware of its shortcomings and is taking steps to remedy the situation. However, it needs the support of crucial ministries such as the Presidency and Finance. After the initial set back with the institutional strengthening component of Cr. 1669-RW (para. 1.15) there is a possibility that implementation will begin in mid-1989 with UNDP assistance. Negotiations are still on-going between UNDP and MINAGRI. The objective will be to reinforce the Ministry's planning and investment preparation capability as well as its monitoring and evaluation capacity. This will enable better coordination and use of external aid. Additional in-country and external training is high on the list of the ministry's priorities and more than 200 donor financed scholarships are expected for the 1988-92 period. Furthermore, as part of a general public service restructuring planned by the Government, MINAGRI has drawn up a new organigramme which proposes to create (among others) a planning directorate and within it an agro-economy division giving new emphasis to  economic and financial aspects of sector management, and a directorate of extension, which would eventually supervise the implementation of the new national extension and adaptive research approach. The organigramme emphasizes planning, policy formulation, coordination and supervision as functions for the central level ministry which is to be kept light in terms of personnel and equipment with responsibilities for implementation decentralized to the field level and drawing on private sector services where available. The organigramme is under review at the Presidency which has held preliminary discussions with MINAGRI on its contents. Indications are that MINAGRI's proposals imply more far-reaching changes than those of other ministries and will therefore require time for debate, consensus building and implementation. In the interim the Presidency has approved the creation, wLthin the Agricultural Pxoduction Directorate, of an Adaptive and On-farm research unit to coordinate and supervise activities in this area. B. Research 2.04 The Institute of Agronomic Sciences of Rwanda (ISAR) is responsible for agriculture and livestock research. ISAR was supported mainly by Belgian aid up until 1981. It has traditionally concentrated on improving husbandry standards for industrial crops, and introducing and testing new varieties of food crops (maize, sweet potato, irish potato and beans). Research has in general been conducted in isolation of small farmer problems and farming systems, and this may have contributed to the limited number of attractive technical recommendations available for extension to the farmer. Coordination with the extension services has been weak and there has been some confusion between extension and research about the modalities for carrying out adaptive and on- farm research. The Belgians withdrew support in 1981, and as a result ISAR's activities were considerably reduced. The IDA financed Agricultural Research Project helped to remedy this situation. Its main objectives are to reinforce research programming, budgeting and financial management, reorganize ISAR and rehabilitate some of its research stations, prepare a research master plan and, in general, redirect research to become more problem oriented focusing on national priorities in agriculture. 2.05 Encouraging progress has been made in the reorganization of ISAR and the recruitment of higher level staff. Overall implementation is, however, slow because of remaining administrative weaknesses and scarcities in local specialist expertise. There are important gaps in the research program with i inadequate attention to research areas such as fertilizer trials, banana, coffee and tea research. The links between extension and research are still weak and research findings still fail to get transmitted to extension. Attempts have been i made during the course of supervision to refocus research priorities and there is some progress in this area. Findings related to disease control in the banana * tree, soil fertility improvement through agro-forestry species, and new varieties of soybeans, sweet potato and beans are likely to be available soon for field and on-farm testing and eventual transmission to the farmer. The creation of an adaptive research unit in MINAGRI (para. 2.03) and the subsequent support for a field service that will implement adaptive research and on-farm trials through the Agricultural Services project will help institutionalize links between extension and research and speed up field trials.  - 10 - C. Extension Structure 2.06 Extension services In Rwanda are provided by donor-financed projects, agricultural parastatals such as OCIR-Cafe and OCIR-The (the coffee and tea organizations), and in the absence of these, by MINAGRI's own services. After years of diverse interventions by agricultural projects MINAGRI in 1982 established a National Extension System (Syeteme National de Vulgarisation - SNV). The objective of SNV was, among others, to train and extend messages to at least 1,000 farmers per commune, to associate farmers closely in identification and solution of farm-level problems, to involve local officials in extension activities through the establishment of commissions, to establish work programs for extension agents, and to monitor and reorient, as necessary, extension. activities. The SNV did not work very well for several reasons, the most important of which is the absence of prior planning and analysis of how such a system would function and what means would be required to implement it. Interventions in extension have therefore continued to be varied according to funding source. As a result of the SNV, however, a basic extension structure was established along administrative lines. In the field, it consists of, at the prefecture level, an agronomist, veterinarian, and, depending on the Importance of forestry in the area, a forester. The prefecture level staff supervise commune level agronomists, veterinarians and foresters of which there are, in prin_iple, one per commune. There are an average of fourteen communes per prefecture. The commune agronomist supervises sector-level extension agents. There is on average one extension agent for each of the eleven sectors in a commune. In principle, there are also veterinary aides at the sector level supervised by the commune veterinarian. Total staff available in the field for extension amounts to 68 agronomists with some post-secondary education, and 2,075 basic extension agents. Based on the present structure, if all sectors were to be covered by basic extension agents (including veterinary aides), a minimum of 3,146 agents would be needed. If the Gitarama/Agricultural Services model is extended nationwide, only about 1573 basic extension agents will be iequired. Additional agents may have to be recruited where livestock, forestry and bee- keeping activities are important. 2.07 There is, at present, a great deal of variation among communes and sectors in terms of coverage by the extension structure. Where there are donor- financed projects staffing is good and sometimes excessive, and vehicles and equipment are available. Oftentimes, the basic extension structure is considerably modified (sometimes out of recognition) to suit the purposes of the project. As a result, there are as many extension services as there are projects. There are also many inequities in the existing extension services. Staff working for donor-financed projects or parastatals are better paid and equipped while those recruited by MINAGRI or the communes are less well paid, poorly equipped and trained. Methodology and Content 2.08 Methodologies for extension are varied. Some of the extension services use some form or other of T&V for delivery of extension messages. There is a wide variation in approach and very little monitoring and evaluation  - 11 - and thus, the impact of extension in many projects is difficult to assess. With regard to extension content, H1NAGRI issues yearly directives focussing on a specific priority (year of the tree, year of agriculture-livestock association, of erosion control etc.). These directives are backed by a national campaign in which the political and administrative structure ls active and usually include a number of general themes such as reforestation, stall feeding of animals, composting and so on. The directives are given to the prefecture level staff who pass them on to the field staff for transmittal to farmers. The whole system is very hierarchical and agricultural staff and farmers are legally obliged to adhere to the directives. There has been a measure of success in extending particular MINAGRI themes such as tree planting, erosion control, marsh cultivation,and coffee improvements. Centrally designed extension themes are however often ill adapted to the different farm conditions and agro-ecological regions. The Agricultural Services Project will seek to generate extension themes from the field level and to better adapt central themes to local conditions. Donor-Financed Proiects 2.09 Agriculture attracts a great deal of donor financing, about 50% in grants, and many of the 60 odd externally assisted projects have extension components. The most important donors in extension are IDA, Belgian, German, French, Swiss and Canadian aid, the European Development Fund, IFAD and UNDPIFAO (IBRD map 21406). In the past, extension and in general agricultural development projects have largely been donor-driven. The approaches to extension among these projects is varied (paras 2.07 and 2.08). Some are clearly top-heavy in terms of personnel, vehicles and equipment and will be difficult to sustain. The impact of the projects is not well known and MINAGRI is finding their coordination increasingly difficult. The Agricultural Services Project is designed to help Government establish a structure within which the numerous projects can eventually be integrated. Through a series of seminars and discussions with donors, MINAGRI has outlined the manner in which it wishes the projects to fit into the proposed structure (Annex 2, Table 3). Basically, when present donor extension projects end, financing in the next phase is supposed to be directed to support for the proposed national structure. 2.10 In theory most donors support the n. w approach and welcome the development of an organizing framework for extension and adaptive research. In practice, however, there is still confusion about how integration will actually work and some hesitation about the proposed structure. Reactions range from very supportive to skeptical. The Canadian and Swiss aid have already realigned some aspects of their projects to the new approach. French and German aid appear willing to do the same. Some of the others want more time to evaluate the Gitarama experience and watch the outcome of the Agricultural Services project before committing themselves. Despite the range of reactions there is a general willingness to give MINAGRI's new approach a chance and the ministry is confident of the goodwill and eventual support of all the donors. IDA has stepped up its dialogue with the donors to explain the proposed Agricultural Services approach. This is providing needed support to MINAGRI.  - 12 - D. Agricultural Inputs 2.11 There is very little use of modern farm inputs in Rwanda. The exceptions are the use of some fertilizer and insecticides on potatoes, insecticides on coffee, and some veterinary products. Fertilizer consumption, at 1,400 grams of plant nutrient per hectare of arable land, is among the lowest in subsaharan Africa. The limited use of modern inputs is due both to the low level of farmer cash income and the lack of knowledge on supply response unde:- farm conditions. The benefits of chemical fertilizer are still largely unproven. Few fertilizer trials are considered valid enough to be generalizable. Clearly, this is an area for the immediate attention of the research and extension services. The Danish/FAO financed National Fertilizer project is conducting fertilizer research and has released some recommendations relative to the use of NPK, DAP, and urea on maize, potatoes, sorghum, groundntts and other food crops. !!ore field and on-farm trials are needed. Farm inputs are imported by a few private merchants in Kigali, by OCIR the and OCIR Cafe, by donor assisted projects (much of it in grant aid), and by a large cooperative organization TRAFIPRO. Fertilizer pricing policy is confused. In principle, MINAGRI requests projects to sell fertilizer to farmers at cost (i.e. CIF Kigali). Most projects that sell directly to farmers comply, although, this leaves little or no margin to cover retail costs. Some projects sell the fertilizer to private retailers who are permitted to add a margin (usually 5% - 10%) to cover costs. In some cases, fertilizer has been given free to farmers by NGOS. Such a situation may make it more difficult to encourage private sector participation in input supply. MINAGRI is aware of these anomalies and plans to review fertilizer pricing policy in 1989 with a view to harmonizing the different approaches. During the appraisal of the ASP MINAGRI agreed to the principle of recovering costs (including retail costs) for fertilizer where there is effective demand. At the proiect mid-term review (para. 4.07) GOR and IDA will review the impact of fertilizer use on crop yields in the project area, and will, further, evaluate fertilizer pricint policy with a view to improving cost recovery including recovery of retail margins. At negotiations assurances were obtained to this effect. 2.12 The quality of selected seed supplied to most farmers is poor. There are few varieties superior in performance to farmers' seeds and therefore demand is limited. Rwanda's seed industry is at an early stage of development, and thus far, private sector interest or participation in the industry has been limited. Some elements of a seed production and certification structure are in place but function in an uncoordinated and confused manner. Other elements, such as seed legislation have, until recently, been missing. A National Seeds Committee (NSC) exists but rarely meets. Its mandate is to monitor the enforcement of seed regulation, approve release of new varieties, review national seed requirements and promote distribution of improved seed. ISAR's mandate is to produce pre- base seed to be sold to the Selected Seed Service (SSS) for production of base seed. Development projects and private entrepreneurs are to produce certified seed. A division of the SSS is charged with seed certification. In practice, ISAR produces not only pre-base seed but also base and certified seed (the latter certified by itself). SSS produces both base and certified seed. Seeds produced by the development projects are not certified because: a) SSS has limited capacity to perform this function and b) the certification function is not backed up by any legislation. Other problems of seed production are its relatively high cost due to small scattered uneconomic fields, and the low sale price for some  - 13 - types of seed resulting in inadequate cost recovery and in some instances diversion of seed to consumption. Of recent, there have been some encouraging developments in the seed industry which should facilitate the production of good quality seed. The FAO has financed a well-equipped laboratory to be used for seed certification. With FAO assistance, the Government has prepared seed legislation which is presently under examination at the Ministry of Justice. The legislation is an important step in the development of the seed industry. It sets out clearly the role of each actor in the industry. The NSC is to maintain its policy and review role including advising the Government on seed policy and setting guidelines for seed prices. SSS is to be the main implementing body for seed policy. It will: a) serve as the secretariat to the NSC, b) in cooperation with ISAR develop and maintain a catalogue of varieties that can be multiplied, c) develop technical criteria to guide seed production and facilitate certification. These criteria will be submitted to the NSC for approval; d) through its certification division maintain its certification function; and, through the production division, produce base and certified seed. Development projects and private individuals can produce certified seed provided they are registered with the SSS and submit to its certification requirements. As the industry develops, SSS is expected to pull out of production, leaving this to the private sector, and focus on certification. ISAR is to be responsible for the production of pre-base seed and for maintaining the characteristics of varieties published in the seed catalogue. The legislation empowers MINAGRI to publish decrees to guide seed production based on technical proposals made by SSS and approved by the NSC. E. Atricultural Ctedit and Finance 2.13 Rwanda's financial rystem consists of the central bank (BNR), three commercial banks, a network of cooperative banks (Banques Populaires), a Savings Association (Caisse d'Epargne du Rwanda), a mortgage bank (Caisse Hypothecaire), and a development bank (Banque Rwandaise de Developpement BRD). The BNR supervises Government owned financial institutions, regulates equity investments by banks and exposure to any single enterprise. The financial system is highly specialized. Unless authorized by the BNR, commercial banks may not engage in long-term lending (more than four years) because of the short term nature of bank resources. The BRD and the Caisse Hypothecaire are the long-term lending institutions. Commercial banks have very little exposure in the rural sector and are mainly active in import and export financing and overdrafts to local companies. 2.14 The Union des Banques Populaires (UBP), set up with the assistance of Swiss aid, is Rwanda's dominant rural banking establishment. The cooperative bank union has a network of over 100 rural branches with an expansion rate of five to ten new branches a year. Financial management is conservative and sound, and the bank holds about 23% of national savings and 4% of the total national credit outstanding. Most rural loans are for housing and short-term commercial operations and are generally very small--less than US $500. The UBP is eager to expand lending to agriculture and has cooperated in the Gitarama project to make loans to farmers. The experience has been gratifying for both the farmers and the Bank. Loan recovery has been high--more than 80%. Continued cooperation with the UBP is expected under the proposed Agricultural Services Project.  - 14 - Government Funding of AMriculture 2.15 Government funding of agriculture has averaged ITS $16 million in the past four years (Annex 2, Table 4). About 56% of this has gone to the operating budget and the balance to the development budget. This distinction, however, underestimates the funding of operating costs since the development budget also contains some recurrent expenditures in the form of counterpart financing for externally funded projects. Usually, recurrent cost items of a permanent nature, e.g. personnel, are financed under the operating budget. Agriculture's share in the overall budget has remained fairly constant in the past five years at about 5% of the operating budget and 16% of the development budget. Other contributions to agriculture come from the communes through locally raised taxes. The amount is difficult to quantify but is not thought to be very high and is variable from commune to commune depending on whether or not an externally funded project operates in the commune. Local financing is mostly in the form of payment of some field level staff such as basic extension agents. A significant contribution comes from the use of collective labor known as umupanda organized by the communes. Umuganda is used for a variety of tasks such as the digging of erosion control ditches, tree planting, establishment of tree nurseries, maintenance of communal forest plantations, and road maintenance. The amount of the contribution to agriculture is difficult to quantify, however, it has been estimated that up to 50% of the cost of establishing tree nurseries cc-,es from umuganda. On a national level, total contributions from umuganda were estimated at a sizeable RF 1.44 billion (US $14 million) in 1985. 31 In the Gitarama project, the use of umuganda for some of the agricultural tasks, e.g. nursery establishment, has been formalized in a contract between the communes and the project. 2.16 The Government has been very effective in attracting external aid to the agriculture sector. An estimated US$ 39 million or about 22% of total official development assistance to Rwanda went to agriculture in 1985. For 1986 the figure is US$ 45 million. This is more than double Government's own financing and constituted 73Z of total funding for agriculture in that year. About 30% of external aid in 1986 was food aid. A great deal of external assistance is given in the form of technical experts a phenomenon which MINAGRI is eager to change. There are an estimated 177 technical assistants in the agriculture sector compared to 122 cadres at the central level of MINAGRI. Non- Governmental organizations are becoming increasingly important. Although figures on recent contributions are not available, they are thought to have contributed up to US$ 700,000 in 1983. The importance of donor financing has led to questions about the future sustainability of investments. Despite the likelihood of continued external support, the Government is worrying about its ability to sustain recurrent costs if a donor withdraws. As a result, the operating cost implications of projects are receiving more attention. In this regard, GOR has just initiated a Public Expenditure Review with IDA assistance. There is a better awareness of the need for cost recovery and reduction of subsidies. Although progress on this front is still slow some recovery of costs is taking place. Some veterinary products and services are sold to farmers at cost, And the subsidy on tree seedlings is being reduced through umuganda ;:ontributions. In addition, there is a move to attract the private sector to take over some of 3/ World Bank, Rwanda: The Role of the Communes in Soclo-economic Development 1987.  - 15 - the services to reduce Government's exposure. These initiatives will be fully supgorted as part of the proposed proiect. P. Markets and Prices 2.17 With increased intensification, the expansion of local, inter- regional and external markets will become important. Except for export crops which are handled both by Government and private concerns, the marketing system in Rwanda remains largely undeveloped. The main constraints are: low purchasing power of the population; the lack of entrepreneurs and managers; lack of credit; an inward looking (subsistence oriented) orientation both by farmers and Government. A fairly good network of national and secondary roads exists. However, some areas are still enclaves where rural road development and maintenance is a top priority. Fooderop marketing is largely in the hands of the private sector. A Government parastatal OPROVIA intervenes from time-to- tlme to stabilize prices through crop purchases. 2.18 Agricultural price policy used to be the domain of the Ministry of Finance. Recently, a new Ministry of Commerce and Consump'ion has been created which in consultation with MINAGRI and other ministries will oversee issues of pricing and marketing. The lack of economic and financial staff in MINAGRI means that its input is limited in an area in which it should in fact be the principal actor. There is io official consensus on what the extent of Government involvement in agricultural pricing should be. In practice, Government interventions have been limited to setting seasonally adjusted minimum producer prices for fooderops and retail prices for fertilizer, insecticides and some veterinary products but paying little attention to their enforcement. Export crop prices are set and have, in general, been favorable to farmers. Since 1987 there has been an increasing tendency for Government to insist that projects, and in some cases cooperatives pay minimum producer prices (particularly when these are bigher and therefore more favorable to farmers than market prices). Furthermore, retail prices for fertilizer are such that retailers find it difficult to cover costs (para. 2.11). Through the Second Integrated Forestry project (Cr. 1811-RW), the Gitarama project and the proposed project there has been a continuing and difficult dialogue between IDA and Government on agricultural pricing. The inherent contradiction in wishing to encourage increased private sector activity in the marketing of inputs, for example, when marketing margins are unfavorable have been pointed out. Opinions are changing. Recent (February 1989) discussions with the ministries of Commerce and of Agriculture indicate that the minimum producer prices regularly published by the Ministry of Finance, and to be continued by the Commerce Ministry, are to be considered as reference prices. There is also a move to review fertilizer pricing policy, to harmonize approaches and improve cost recovery (para. 2.11). G. Cooperatives and Farmer Groups 2.19 There are an estimated 3,240 registered cooperatives and farmer groups in Rwanda. Most have less than 50 members. In addition, informal groups (including farmer groups) have been estimated at 30,000. The cooperative and mutual help movement is strong in Rwanda and is an important ingredient of local  - 16 - communities. The activities of cooperative groups are diverse. Some are involved in farming--land preparation and sowing, others in marketing, processing, savings and credit. The Directorate for Cooperatives in the Ministry of Youth and Cooperatives provides support through its front line agents at the commune level. Donors are also active in this area. The Cooperative Research and Training Center (CFRC-IWACU), which is an NGO, is supported by USAID and Swiss aid. It has become an important training and research organization focusing on economic, social and management aspects of cooperative development. IWACU and a Belgian supported NGO the Compagnons Batisseurs COOPIBO, were contracted to provide management and organizational training and consultancy services to cooperatives under the Gitarama project. A federation of artisans KORA (meaning 'to work") was also contracted to provide training in the manufacture of agricultural tools to artisans in Gitarama. These activities in cooperative development are Also envisaged in the proposed project. III. THE PROJECT A. Rationale 3.01 Improving agricultural performance in Rwanda requires: a) the strengthening of essential agricultural services and their development on a sustainable basis; b) the expansion of markets with increased private sector participation in marketing; and c) an enabling policy environment. The Government has shown a willingness to create the appropriate conditions for agricultural development and IDA has provided assistance through the Agricultural Research Project (Cr. 1546-RW) and the Gitarama project (Cr. 1669-RW). The proposed project would be a further instrument to implement Government's strategy to increase productivity and output in the agricultural sector. B. Summary Project Description 3.02 The project would include: (a) reorganization and streamlining of the extension services in all or parts of five prefectures including the implementation of a new structure integrating extension, adaptive research, and monitoring and evaluation; upgrading of extension staff through training; and ftprovements in extension methodology and content. The Gitaraxa extension and agricultural services model would provide the basis for the reorganization but with modifications related to the streamlining of personnel at certain levels of the structure and reductions in the number of vehicles; strengthening of the role of management information systems; and improvements in the links between research and extension. (b) strengthening of adaptive and on-farm research through better diagnosis of problems at the farm level and inclusion of  - 17 - farmers and local authorities, in a participative manner, in problem identification and solution; implementation of on-farm and multi-local trials of new techniques and technologies provided by research and addressing farmers' priority problems . (c) support for monitoring and evaluation through the implemertation of a decentralized management information system in which extension staff at all levels would participate in monitoring and evaluation activities as part of their regular work program; sensitization and intensive training of staff in monitoring and evaluation (M&E); development of key indicators by technical staff for monitoring the internal functioning of the agricultural services, and the impact of extension; regular reviews (evaluations) by technical staff supplemented by more detailed outside evaluations as needed. (d) strengthening of seed production and distribution through support to the SSS to improve seed quality control and certification, reactivation of the National Seed Committee for supervision of seed regulation, improved pricing policy for seeds, a_.d rationalization of seed production including improved management of ex-BGM/PCCV project seed farms and support for the development of a multi-purpose farm for seed multiplication and adaptive research trials. The farm would serve as a center to encourage and develop the practice of contracting small farmers as out-growers for seed multiplication purposes. (e) promotion of private sector participation in input supply and marketing through assistance to cooperatives and private enterprises to procure inputs, obtain credit and improve enterprise management. NGOS and the Banques Populaire would be contracted to provide the needed services. C. Geozraphical Coverae 3.03 The project would be implemented in the five prefectures of Kigali, Kibuye, Kibungo, Gisenyi and Cyangugu covering 50% of the national territory 41 IBRD map 21291. Direct project activities would be implemented in seven sous- prefectures (three of which are yet to be created), covering twenty-two communes and comprising 239 sectors (Annex 3, Table 8). The total population of these communes is about 840,000 (168,000 families) or 14% of the national population. The communes cover an area of about 5900 km2 (23% of national area) located principally in the north west, south west and east of the country. They can be 4! The prefecture level teams to be financed under the project, will coordinate and supervise prefecture wide agricultural activities including those of other donors. In this sense, the project covers all 70 communes in the five prefectures or a total of about 570,000 farm families.  - 18 - represented by three agro-ecological zones vizt Lake Kivu shores in the northwest (former PCCV project area) with elevations of 1450 - 2200m and fertile volcanic and semi-fertile ferrisoils. Farm fragmentation is far advanced here and over 50-70% of holdings in this area are less than 1 ha; Imbo in' the southwest--this is the lowest elevetion in the country 900m, with good rainfall and fairly fertile soils. Holdings ate becoming increasingly smaller with about 22% less than 0.5 ha according to the 1984 agricultural census, (there is no donor financed project in this area); Bugesera-Gisaka to the east--this is an area of recent settlement assisted under the BGM I and II projects with fairly poor sandy soils, low and erratic tainfall, and relatively low elevations 1300- 1500m. Large farms are necessary to meet subsistence and other needs and up to 70% of the farms have 2 ha or more. 3.04 The principal fooderops in the three agro-ecological regions are bananas, beans, maize, sweet potato, and sorghum. Manioc and groundnuts are important in Imbo and Bugesera-Gisaka. Soybeans is an emerging crop in all three areas. Coffee is the main cash crop and particularly important in Inbo and Lake Kivu Shores (Bord Lac Kivu or BLK). Up to 60% of families own small stock principally goats. About 11% of farms have cows in the Bugesera and BLK zones. D. Detailed Features Extension Structure 3.05 The project would finance a field service structure that parallels the administrative organization, as in Gitarama. 5/ Its f1nctions would be extension delivery, implementation of adaptive and on-farm research, and monitoring and evaluation of the functioning and impact of the agricultural services to permit modifications and improvements when required. As such, the structure should be regarded as flexible. Since it is new and there are some functions not previously performed under existing services, some incremental personnel would be required, particularly at the post secondary level (Annex 2, Table 1). On the other hand, there would be reductions of personnel in some functions and redistribution from over staffed to understaffed areas. At the base of the structure would be the basic extension agent of which there would be one per sector (approximately one per 750 families) except in some very large sectors or those with difficult terrain where there would be two. A total of 248 basic extension agents (BEAs) (down from 310 at present) would cover the 239 sectors in the project. Of the 310 agents, 74 are presently financed by MINAGRI through its operating budget, 62 are financed by OCIR cafe, 61 by the communes, and the rest by the BGM and PCCV projects resulting in differing conditions of service (para. 2.07). The financing of the BEAs would be harmonized under the proposed project by including them in the MINAGRI budget to assure fairness and sustainability. Of the 248 BEAs, 91 would be financed under the operating budget. The remaining 157 would be covered under the development budget for a period of three years after which they would also be transferred to the operating budget. Similar arrangements would pertain to other personnel. The three year period is judged necessary to allow the Ministry of Finance time to make 5/ A description of tasks within the structure and details of the technical aspects of the project are contained in the Working Papers in Volume 2.  - 19 - adjustments in its operating budget procedures. Adjustments will also have to include allocations for other operating costs to be financed by MINAGRI under the project (para. 3.40). Assurances were obtained on these arrangements at negotiations. Contributions made by the communes would be channeled to other project activities of a more temporary nature such as establishment and maintenance of forest nurseries. The project would finance 171 motorcycles for BEAs working in the hilly west of the country and 77 bicycles for those in the east. Motorcycles and bicycles would be provided on credit. Rep&yment amounts calculated on the basis of replacement value would be paid to MINAGRI under loan agrealments satisfactory to IDA. The money would be deposited in an account to be managed by MINAGRI and used to purchase new vehlales as required. Assurances were obtained on this at negotiations. In line with present practice, transportation allowances would be paid to cover vehicle maintenance. With the proposed project, staff salaries and allowances for personnel in the structure would be harmonised. This will result in increases for some staff. For example, salaries of BEAs not previously working in donor financed projects would increase by 8% while salaries of A3 officers under the BGM II and PCCV projects would go up by about 10%. Salaries of some higher level personnel (AO staff) in the BGM II and PCCV projects would decrease by about 15% due to a reduction in allowances paid over the basic salaries. The number of people affected by the reductions is small. MINAGRI has agreed to proceed with the salary harmonization, and intends to extend the salary and allowance grid gradually to all agricultural services personnel countrywide. 3.06 At the commune level there would be a team of technicians made up of middle level post secondary but pre-university personnel of A2 level comprising a communal coordinator, an agricultural officer, a livestock officer, and an officer in charge of forestry services. The agricultural officer supported by the livestock and forestry officers would supervise and assist the BEAs and provide them with technical advice. The commune team would be answerable to the U.O coordinator who together with his team would provide them training, and assistance in programming and planning (para. 3.07). Where there is a lot of forestry activity (communal plantations, agro-foreatry, reforestation), the forestry officer could be assisted by up to two forestry monitors to supervise cutting, replanting and payment of tariffs in line with the new forestry law. Such monitors should eventually be paid through the revenues collected by tree sales. The Forestry Directorate fully agrees with this principle and will implement it in the communes once the forestry law becomes operational. Also, in communes where there is a large number of livestock, the livestock officer would be assisted by up to two veterinarians for animal health care and training of livestock owners to take over some of the veterinary services. As livestock owners become trained, the need for additional veterinarians should diminish and they should be let go. Since MINAGRI wishes to encourage bee-keeping for increased production of honey, an agent specialized in this area would be provided in those communes where there is an interest. The project would finance the construction of some offices and storage space (for inputs) in 17 communes, the necessary office equipment and materials, incremental staff, and vehicles (motorcycles and bicycles). 3.07 The "Unite Operationnelle" (Operational Unit or U.O) is a very important part of the new structure, begun and tested in the Gitarama project, and a new concept in extension and adaptive research. The U.0 is to be based at the sous-prefecture level. It will be staffed by subject matter specialists (SMS) responsible for training of extension personnel, conception of area-  - 20 - specific extension themes, continued dialogue with research and participation in the definition of areas for adaptive research, provision of technical advice to the commune team and the BEAs, and definition of issues for further follow- up through monitoring and evaluation. The U.0 team will consist of an AO (university trained) coordinator who, in addition to administrative tasks, will be the primary analyst in the team, an experienced A2 officer for agriculture and one for livestock. Selection of U.0 staff will be done with great care to ensure that the officers have long-term field experience, and possess the analytical capability to perform the required tasks. The team is expected to spend about 1/3 of its time on training for which they will receive assistance from the prefecture level training speciallst, 1/3 on research, and the rest on the remaining tasks. The number of BEAs per U.0 team will range from 33-55. The team will be answerable to the prefecture agricultural coordinator. The project will finance a pick-up van and four motorcycles per team, the consl4ruction of offices, houses, and a few training rooms in five of the seven U. Os where such infrastructure is lacking, and office equipment, furniture and material. Houses would be rented to personnel for an amount sufficient to maintain such housing but in any event no less than 1%, on an annual basis, of the investment costs for such housing. Rent would be collected by MINAGRI on the basis of lease agreements satisfactory to IDA and would be deposited into an account managed by MINAGRI for said maintenance purposes. Prefecture coordinators would make requests to MINAGRI for maintenance funds as needed. Assurances were obtained on this at negotiations. 3.08 The staffing and equipment of the U. Os would be scaled down from that in Gitarama leading to some small savings. Instead of a core team of seven with four cars, there would be four staff with one pick-up van and four motorcycles, (Annex 2, Table 5). In view of the scarcity of university trained personnel they would be carefully distributed to strategic places within the structure. The proposed reductions at the U. 0 level in the Agricultural Services Project has led to suggestions to reduce vehicles in Gitarama initially from four to two. More importantly, the number of university trained personnel per U. 0 is likely to be reduced from three to one so they can be redistributed to other places where they are needed in the project. 3.09 The prefecture level team will be responsible for coordination and supervision of project activities as well coordination of all agricultural activities in the prefecture. It is expected that this supervisory team (acting as a common denominator for all prefecture agricultural activities) will be a force for the gradual integration of donor extension and adaptive research projects into the new structure. The team will comprise an AO Prefecture Agriculture Coordinator whose duties will be mainly administrative, assisted by an AO technical officer responsible for coordination of adaptive and on-farm research and monitoring and evaluation. The latter officer will be responsible for assuring that the linkages between research, extension and monitoring and evaluation are kept at the forefront and maintained through the structure. Since training will be an important feature of the agricultural services (para. 3.36), an AO specialist in training will also be part of the team. He will assist and supervise trainers at the U. 0. An AO forester will supervise prefecture forestry activities. In Kirehe, Kanazi and Bugumya, three sous-prefectures where forest activities are very important, the forester will be based in the U. 0. An experienced A2 accountant will maintain prefecture project accounts and supervise the other local level accountants in the structure. The team will have one pick-up van and four motorcycles. The project will finance incremental  - 21 - personnel (such as the AO asssistant to the coordinator and the training specialists), some construction of offices, training and storage rooms, and houses particularly in the prefectures of Cyangugu and Gisenyl where these are lacking. Vehicles, equipment and furniture will also be financed. Although the five prefecture coordinators already exist their formal appointment in the new structure is a condition of credit effectiveness. 3.10 Overall supervision of the structure and coordination of all prefectures will be carried out by a light team at the central MINAGRI level. This is in line with the project's philosophy to avoid a top-heavy structure and to emphasize field level activities. It also supports the Government's policy to decentralize responsibility to the local levels. The team will be an integral part of MINAGRI and will consist of an AO "fonctionnaire dirigeant" (higher iavel cadre appointed to administer the project), assisted by an experienced AO research officer, possibly on secondment from ISAR, who will serve as the main resource person for adaptive research and the primary link with the basic research structure (ISAR). This officer will be part of the new adaptive research unit to be created in the Agricultural Production Directorate of MINAGRI (para. 2.03). A total of nine man-months of technical assistance will be provided to assist the fonctionnaire dirigeant (FD) with the start up and implementation of the monitoring and evaluation system including a mid-term project review. A financial officer will be recruited for three years to set up and manage accounts for the agricultural services and train accounting personnel. He will be assisted by an AO chief accountant, and an A1/A2 accountant. The team will have three cars. Documentation, publication and information dissemination services will be needed. MINAGRI has indicated that this will be provided by their documentation service which is to strengthened under the proposed UNDP financed MINAGRI Institution Building project. About US$ 6,000 will, however, be provided under the project to assist with information collection and dissemination. Ihe project will finance office furnitu v and equipment including a portable computer for financial and other analyses. Appointment of the FD, and recruitment of the financial and adaptive research officers are conditions of credit effectiveness. Methodoloat and Content of Extension 3.11 The extension system will be essentially T&V modified to reflect local institutions and structures. Extension agents will work within a time schedule consisting of a two-weekly cycle of farm visits, training and reporting. Emphasis will be put on contact groups with the principle of trying to reach all interested farm families through these groups. At the start of the projec.: farmers will be consulted as to what type of contact groups would be most appropriate in their area. Existing groups, created for other purposes but suitable for extension, could be used. Examples are the "Centre Communal de Developpement et de Formation Permanente" CCDFP groups (Communal Development and permanent training centers groups) formed by the Ministry of the Interior to facilitate training of rural people in differ'.'nt areas and skills; informal groups formed by farmers themselves for agriculture or related purposes. Some of these groups consist of or are formed by women. As part of the project's objective of increasing the outreach to women farmers these groups will receive special attention. Actions designed to improve delivery of extension services to women farmers are described in paras. 3.33 - 3.35. Where there are no groups, local people will be consulted on the possibility of forming groups based on  - 22 - geographical zones or topology of farms. Group sessions will be held at farmers compounds or farms in turn, and groups visits will be alternated with individual visits every two weeks. Extension agents will demonstrate new ideas or techniques on small plots (parts of farmers' fields) and other farmers will then be encouraged to try these ideas on a 3mall part of their fields before applying them on a wider scale. 3.12 The BEA will be the primary point of contact with the farmer. He will be responeible for transmitting all extension themes and sending up information gathered from the farmers. Farmers will be active in the generation of extension themes and the choice of themes to be discussed at group sessions. Instead of extending one theme per month as is done in Gitarama, several themes or sub-themes related to the farmers' activities in that particular part of the farming cycle will be discussed. The project will continue and expand the practice of the concours aRricole (annual agricultural competition) for the best farmer, best extension group, etc. Prizes are usually small agricultural tools, or other inputs. 3.13 The content of extension will be based on small, incremental, but highly beneficial improvements to existing practices. Revolutionary discoveries are unlikely given existing country and sectoral characteristics. The extension service will continue with themes or sub-themes that have been successfully extended in the past bul which have yet to reach many farmers, and with other themes that have recently become available. New themes or improvements in existing themes are expected to become available through the adaptive and on- farm research process. Themes that are available for extension through the project include: a) composting; b) legume culture--dissemination of recommended varieties along with appropriate cultural techniques and the use of compost. (this theme seems particularly appreciated by women); c) sweet potato--diffusion of popular and more productive varieties e.g. K 51 accompanied by rapid multiplication techniques; d) post harvest conservation and treatment of cereals and legumes and processing of manioc (again of interest to women); e) erosion control--digging of drainage ditches, terracing, cultivation of trees or appropriate plants along contours; f) soil fertility improvement--interplanting of soil improving agro-forestry species, use of compost and chemical fertilizer where applicable; g) coffee cultivation--plantation density, pruning, mulching, use of chemical fertilizer, disease control, better harvesting and processing techniques; h) introduction of soybeans;  - 23 - i) agro-forestry, communal forestry; J) banana cultivation--interplanting with beans, use of by- products, disease control; k) animal health--preventive care techniques including vaccinations; 1) treatment of seed before planting; and m) animal production--stabling of animals, genetic improvement, fodder crop cultivation, chicken and fish farming, beekeeping. 3.14 The project will finance some investments in support of the above themes. Material, equipment, and some operating costs will be financed for forestry and coffee nurseries at the sector level. Through umusanda farmers will provide labor for establishment where necessary. The idea is to gradually get farmer groups to take over the nurseries as cooperative concerns to produce plants for themselves. It is expected that by the third project year one-third of the 239 sector nurseries will be completely operated by the farmer groups, two-thirds by the fourth year, with complete coverage by the fifth year. Assurances were obtained on this at negotiati,ns. In support of communal tree planting a lump sum of about $1,500 per comwrAne per year will be provided for the purchase of plants contracted from private nurseries. Maintenance costs for existing communal plantations will be financed for two years after which the communes will take over maintenance themselves with funds recovered through the sale of trees in the plantations. The implementation role of the communes in the project including farmer groups within these communes, financing arrange- ments, and the responsibility of the agricultural services will be clearly spelled out in agreements (satisfactory to IDA) to be signed between the services (MINAGRI) and at least seven communes (or farmer groups in the comwznes) no later than June 30, 1990. Assurances were obtained on this at neRotiations. Implementation of contractual arrangements will be monitored by the communal coordinators and bourgmestres and problems discussed during the regular meeting of the commune technical and development committees (para. 4.04). 3.15 Some dispensaries, vaccination crushes, watering holes and slaughter houses will be constructed in support of livestock activities. These will be located mainly in Cyangugu where this type of infrastructure is seriously lacking. A transit center for young chicks will be constructed in the Gishiyita U. 0. No later than six months after credit effectiveness GOR will provide plans, satisfactory to IDA, for the management of this center including its eventual transfer to farmer groups or the private sectco. Assurances were obtained on this at negotiations. The project will finance material, equipment and breeding stock. A $78,000 revolving fund for veterinary inputs and $62,000. for the purchase of young chicks will be provided. Animal health services will be provided at cost (excluding salaries). Considerable progress has already been made in the BGM project areas in recoveriag non-salary direct costs. This practice will be extended to other project areas.  - 24 _ Adaptive and On-farm Research 3.16 This component is designed to identify and address farmers' concerns in a participative manner, and to strengthen the linkages between farmers, extension and research by establishing a two-way flow of information. It is expected that this will provide the basis for generation of new extension themes or improvement of existing ones. Strong linkages are expected between adaptive research, extension and monitoring and evaluation. To achieve component objectives several activities are envisaged. In the initial stages the project would finance diaznostic surveys to be carried out by a joint team of extension staff and ISAR. The aim would be to better assess farmers' constraints and resources. Stock would be taken of existing technologies which could be translated into extension messages to overcome constraints. Constraints for which there are no ready solutions would be fed back to research for action. Following this, a more detailed program of on-farm research would be pursued. Between seven to fourteen experimental groups (1-2 per U. 0) would be set up for more detailed study. In addition to an in-depth look at the technical aspects of the farming enterprise, on-farm research would strive to assess economic, social and environmental impact of recommended technologies. Results would be fed back to the extension service and to ISAR. The experimental groups would be made up of farmers representing different farming systems and agro-ecological conditions. The farmers would be selected on a voluntary basis and would carry out on-farm trials and regularly provide data on their farms. Exchange of information would be promoted between group members. 3.17 As part of the effort to identify constraints at the local level and find solutions the project would support communal agricultural planning. This would be initiated in three communes and the results assessed before expansion to an additional four communes. The project would provide up to US$ 10,000 per U. 0 to set up a participative process that would enable local people to define problems and propose action plans to solve them. Activities in the action plans could become part of the adaptive research program in the projects annual work plan (para. 4.06). The commune agricultural plans would also feed into the overall commune plans to be developed as part of the fourth five year plan exercise. Commune plans would themselves become part of the prefecture and national plans in support of the bottom-up planning process initiated by the President in 1987. 3.18 The extension service will cooperate with ISAR to expand the network of multi-local trials in order to cover the main agro-ecological zones. An estimated 14 multi-local test sites (2 per U. 0) of about 2 ha each would be established. The trials would be designed, visited at critical stages, and analyzed by ISAR researchers in cooperation with the U. 0 agricultural subject matter specialist. Regular supervisions would be carried out by U. 0 staff with routine operations implemented by commune level staff. Great care would be taken that only trials addressed to farmers genuine priorities would be implemented. Some of the experiments will be in farmers fields. Researchers supervising multi-local experiments and extension staff would meet at least once a year at the U. 0 to evaluate research results and prepare the collaborative work plan for the next year. 3.19 Technologies successfully concluded under controlled conditions would be systematically tested in simple on-farm trials before extending them on a large scale. The project would provide seeds, fertilizers and pesticides for  - 25 - these tests. Results would be shared with research and discussed with participating farmers. A priority area for tests would be fertilizer trials on coffee and on different food crops. Other areas would include testing of different higher yielding, and disease resistant varieties of beans, soybeans, sweet potato, and coffee; experimenting with erosion control techniques involving use of agro-forestry species; alley cropping; and improved cultural practices for coffee. 3.20 While ISAR would establish the protocol for multi-local and on-farm trials, and analyze results, the extension service would implement the tests. The agriculture SMS at the U. 0 would work closely with ISAR and with the prefecture level officer in charge of coordinating adaptive research on the conception of the trials. The SMS would supervise their installation by the commune agricultural officer who would associate the BEA in the installation. Given the importance of adaptive research, a full time experienced research officer will be appointed at the central level to coordinate research activities in the project (para. 3.10). About eight man-months of technical assistance would be provided to assist the research officer with conceptual work, training and the design of surveys and other test instruments. Part of this short-term consultancy could also be used to define tasks, job content and modalities of operation of the proposed adaptive and on-farm research unit in MINAGRI. The research officer may initially be based in one of the prefectures to facilitate contact with field level problems. The officer would be involved in the formulation and evaluation of ISAR programs in order to enhance their relevance to the farmers' constraints and resources of the project area. He would also liaise with other agricultural projects dealing with adaptive research to learn from and disseminate their experience to field staff. He would facilitate exchange of information on research results between prefectures. Prior to credit effectiveness ISA! and the extension service would-sign an accord, satisfactory to IDA, detailing the respective duties of each party. The Research Officer would monitor the respect of the contract by the two parties. Monitoring and Evaluation (M&E) 3.21 The objective of the monitoring and evaluation program will be to keep ministry and project staff at all levels regularly informed on three questions: a) Are the various project programs being implemented as planned? b) Are they leading to expected results, and c) If not, what should be done to revise or modify the original plan? The monitoring and evaluation program is to operate as a decentralized management information system and it is intended that technical and administrative staff at all levels in the structure w;t t themselves undertake monitoring and evaluation activities as an integral part of their regular work program. Information obtained through monitoring and evaluation will be fed back rapidly to extension and research for action and vice versa hence the tight linkage between the three. No special monitoring and evaluation unit will be established, however, short term technical assistance will be provided to advise  - 26 - and assist the different technical services of the project on the setting up and use of their own monitoring and evaluation system. The effectiveness and value of the monitoring and evaluation program rests entirely on the recognition by staff that monitoring is an essential ingredient of their work and that the program is designed to help them perform more effectively and is not to be used primarily as a controlling mechanism. To facilitate an understanding of raonitoring and evaluation, and to make the system operational and effective the project would provide technical support and training on a regular basis to staff at all levels through a program of seminars and short-term training courses (para. 3.37). 3.22 Internal prolect monitoring would be based on the development of a table of key indicators by each technical service. A set of monitoring guides will be used to help with the selection of appropriate indicators and targets. The indicators will be used to ensure that, on the one hand, project inputs (human, material and financial) are available at the right time, in the right quantities, and at the right place and that, on the other hand, the project is generating the expected outputs for the farming community in terms of services and products (extension, training, seed, fertilizer etc.). The monitoring guides will show, for each indicator, a target or objective against which actual achievements will be compared. These forms will be periodically updated and their contents evaluated throughout the year. Information for internal monitoring will come from internal project records kept by project staff. The output of the monitoring exercise will be discussed at regular fortnightly meetings at the commune level and monthly meetings at the U. 0. 3.23 The impact of extension and other atricultural services on the beneficiary population will be monitored. Three key indicators will be used to do this: a) participation ratess the proportion of the population that is maintaining regular contact with extension agents through meetings and demonstrations; b) adoption rates: the proportion of the population that is applying a recommended message at least on an experimental basis; and c) repeat adoption rates: the proportion of the population that, having once tried out a message, continues to apply it in subsequent seasons. These three indicators will be produced for the population of an area e. g. commune, sous-prefecture and for sub-groups of that population--male, female, large or small farmers so that comparisons can be made between sub-groups and even between different areas. Light, simple and timely surveys conducted 2-3 times a year on a random sample of farmers will be used to produce information on the three indicators. These surveys will frequently highlight problems which can be further investigated through diagnostic studies carried out by the extension staff themselves or by outsiders invited to do so. 3.24 The above monitoring system has a built-in evaluation mechanism. That is the chain of two-weekly, monthly, bi-annual and annual meetings that are scheduled to take place at the levels of the commune, U. 0., prefecture, and  - 27 - MINAGRI. Periodically, more complex evaluations by an external independent body will be undertaken. These may require some additional data collection, but a measure of the longer term impact of the extension service, for instance, will for the most part be obtainable through an analysis of the trends of adoption rates for different farmer groups over time. An important evaluation study will be the mid-term review in project year 3. This will look carefully at the impact of the extension methodology used. To facilitate the monitoring and evaluation work the project will finance, in addition to the 9 man-months of technical assistance, a team of enumerators to be called on for surveys as needed, and a computer for analysis. The 'fonctionnaire dirigeant' will coordinate the monitoring and evaluation effort. During project implementation, and in any event, no later than December 31, 1991, MINAGRI will identify an agro- economist, who will receive short-term training in monitoring and evaluation, methods so as to effectively assist with the coordination of monitoring and evaluation activities and with the analysis of data. Assurances to this effect were obtained at negotiations. Strengtheninx of Seed Production 3.25 The project would strengthen seed production and distribution and assist with the implementation of national seeds policy. The draft seed legislation now before the Ministry of Justice may take some time to approve into law. In the interim, MINAGRI plans to reactivate the seed industry along the lines described in the legislation. In this regard, the Agricultural Research project Cr. 1546-RW would support ISAR's efforts to improve production of pre- base seed. J4INAGRI would issue a ministerial decision empowering the certification division of the SSS to effectively control seed quality and certify seeds based on technical criteria to be prepared by the SSS and approved by the NSC. The ministerial decision would also spell out the role of each actor in the seed industry. At negotiations assurances were obtained that not later than June 30, 1990. SSS would prepare the draft technical criteria which should be sent to IDA for review and comment and then to the NSC for approval. Further, not later than January 31, 1991 (and in the event that the seed letislation has not yet become law), MINAGRI would issue the ministerial decision Ruiding seed production and certification. 3.26 The project would provide about US$ 20,000 to the NSC to enable it to seek consulting assistance relevant to fulfilling its mandate. In this regard, the NSC would review the present seed demand and supply situation and assess requirements for certified seed. 3.27 The project would seek to improve seed aualitv control and certification and thereby increase farmers' confidence in the quality of certified seed sold. Assistance would be provided to the SSS to develop a seed catalogue showing varieties that can be multiplied. The project would finance additional manpower and operating costs to facilitate field inspections by the certification division. This would complement assistance provided by FAO and by Belgian aid which finances four seed experts in the SSS. 3.28 To help rationalize seed production and improve the viability of former BGM/PCCV project seed farms support would be provided to the SSS to take over management of six of these farms totalling 50 ha. Once viable, the farms are expected to become potentially attractive to the private sector. The project  - 28 - will finance additional investments (minor civil works and equipment) on these farms and SSS's incremental overhead cost--salary of technical staff, transport and subsistence entailed by the transfer. The full direct cost of producing certified seed on the farms would be financed the first year. Thereafter, a decreasing subsidy would be provided until project year 3 when revenues from seed sales are expected to cover the direct costs of production. The project would also finance the gradual development of a 40 ha area in Bugarama, Cyangugu prefecture, set aside by the Government partly for seed multiplication and parti.y for adaptive research trials including small stock trials. Only a small part of the area (about 10 ha) would initially be developed for multiplication of base and certified seed. During the project mid-term review in year 3 (para. 4.07) an assessment of the total area to be developed for seed multiplication purposes would be made based on the availability of and demand for improved varieties, and the extent of implementation of national seeds policy. The seed multiplication activities would be managed by SSS which would use the farm as a base to encourage the contracting of small farmers as outarowers for seed multiplication purposes. Besides the establishment costs, the project will finance direct seed production costs on th.t same basis as the ex-BGM/PCCV seed farms to be transferred to the SSS. 3.29 The project will improve cost recovery and seed pricina nolicy. The present average direct production cost of base and certified seed is about RI 100/kg while the average sale price is RF 50/kg. The difference is financed through government subsidies. The project would aim to recover at least the direct cost of seed production within 3 years. This will be achieved by reducing production costs through better management and increasing the sale prlce for seed. During appraisal, the basic principle of selling improved seed at prices at least 50% above consumption prices was retained. On this basis, the prices of groundnut and soybean seed would be increased. A periodic (six-monthly) review and revision of seed prices on the above principle will be instituted. During negotiations assurances were obtained that GOR would recover 65% of the direct cost of seed production by December 31. 1990, 80% by December 31, 1991 and 100% by December 31, 1992. Full recovery of direct costs from thae ex- BGM/PCCV seed farms would save the Government an estimated US$ 54,000 (50% of direct production costs) a year on an expected production of 86 tons. An additional US$ 28,000 may be saved from seed multiplication activities on the Bugarama farm at full development. Total cost of the seed component is about US$ 2.2 million. Establishment of proper cost accounting procedures on the seed farms would be a top priority. This would improve tracking of costs and revenues and provide a basis for a proper financial analysis of seed production. Promotion of Azricultural Marketing, Input Supply, and Cooperative Development 3.30 The relative success of the Gitarama project in promoting private sector activity in input supply and marketing would be replicated in the project through the inclusion of a promotion component. Its objective would be to a) develop the provision of agricultural inputs--selected seeds, fertilizers, pesticides, and tools to farmers; and b) further assist the cooperatives and the private sector in the fields of processing, storage,and marketing of agricultural products. The component is designed to provide support to the project's extension activities by ensuring that inputs are available when needed and outputs can be marketed. An Agricultural Promotion Office (APO) would be established in each of the project's five prefecturas to carry out the promotion  - 29 - tasks including coordination of the activitizs of NGOs and tho Banques Populaires (BP) regarding training of cooperatives and artisans and provision of credit. The APO will be staffed with an AO Promotion officer answerable to the prefecture coordinator, an A2 assistant and an A2 junior accountant. The project would finance housing construction for three officers, office equipment and furniture, and a pick-up van and motorcycle per office. 3.31 In the initial stages of the project, the APO will act as a wholesaler responsible for procuring inputs (e.g. selected seed) and distributing to retailers in the project area with the objective of eventually stimulating the private sector to take ove; this function. At the end of the project the APO should work itself out of a Job. A revolving fund of about USS 218,000 would be provided to fund the inputs most of which would be purchased locally. The APO will also be responsible for identifying activities and studies necessary to develop agricultural production. Furthermore, the APO will be in charge of preparing and managing contracts with the following institutions: i) the Banques Populairess The BP will be responsible for providing and recovering agricultural credit including seasonal credit, briuging (consumption) credit, loans for purchasing cattle and paying farm labor, loans for start-up of trade, and purchase of processing and other equipment. The BP will also mobilize rural savings through the opening up of new branches in the project area. The project will support the BP's activities through the provision of a guarantee fund of RF 6 million (US$ 78,000) to cover lending to selected cooperatives in the project area. The fund will cover a maximum of 30% of any single loan to the cooperatives. ii) IWACU: This NGO will be in charge of training cooperatives in bookkeeping, inventory management, and preparing loan applications. IWACU is also expected to foster inter- cooperative relationships in the project area. Its activities will cover about 20 cooperatives in each prefecture beginning with Cyangugu, Kigali, and Gisenyi. Kibungo and Kibuye will be covered later. To this end, about RF 60 million (USS 780,000) will be provided in the project for a five year contract to be reviewed bi-annually. Special attention will be paid to womens' cooperatives interested in marketing, joint production of a particular commodity, and processing of agricultural products. iii) KORA: The training of artisan groups may be handled by this NGO with a special emphasis on the manufacture and repair of agricultural tools, processing and stocking equipment, wood works, and improved charcoal stoves for cooking. The contract with KORA will be contingent upon successful implementation of a similar contract in the Gitarama project. In the event that KORA's performance is unsatisfactory the contract will be awarded to another NGO. An amount of RIF 20 million (about US$ 260,000) has been included in the project for the training of artisans. Not later than March 31, 1990, MINAGRI would sign contracts with the above organizations (except for KORA whose work is still to be evaluated), to implement the activities  - 30 - described. Contracts should be approved by IDA. Assurances were obtainad on this at negotiations. 3.32 Some parts of the project area are isolated because of the lack of secondary roads, inadequate maintenance of existing roads, and broken bridSes. This poses considerable problems for marketing -rhich, if not resolved, could have serious adverse impacts on project production objectives. The worst areas are in the communes of Gishoma, Gafunzo, Cyimbogo, Gisuma, Gatare, Kirambo, Karengera, and Bugarama all in Cyangugu prefecture. A list of the roads and bridges in question is presented in Annex 2, Table 6. While the ASP project will not finance road development or maintenance, financing of the priority list of roads and bridges could be included in the proposed IDA financed first Transport Sector project. To facilitate this MINAGRI should reach an agreement in principle with MINITRAPEE on the importance of according priority to the infrastructure works needed under the ASP project through the communal roads component of the proposed Transport Sector project. The modalities for construction or rehabilitation of the roads and bridges identified and their future maintenance will be discussed with MINITRAPEE and with the communes in the context of the transport project. The ASP will finance the construction of three quays (points d'accostage) in communes along Lake Kivu to make these communes accessible to extension agents and for marketing purposes. Improving Service Delivery to Women .3.33 An important project objective is improving womens' access to agricultural services. While women do the bulk of sowing, weeding, and harvesting for fooderops, processing and caring for livestock, they typically do not feature in the groups of farmers that attend extension -cosions or demonstrations (para. 1.17). Extension agents tend to focus on the heads of households--men and the latter often fail to transmit messages to their wives or transmit them inaccurately. Failure to reach women may be instrumental in lowering adoption rates for messages particularly for fooderops. A start has been made in the Gitarama project to increase the number of women volunteers attending extension sessions. About 102 of the farmers in the Karama U. 0 participating in these sessions last cropping season were women. Considerable efforts still have to made. A study commissioned as part of the preparation of the Agricultural Services Project indicates that the most important constraint to womens' participation is time. During the day women are busy on the farm. In the evenings they are occupied with household tasks leaving little or no time for outside activities. Other constraints to reaching women include the social attitudes which cast the man in the role of 'point of contact' with the outside world; and the lack of awareness of extension agents of the importance of including women in their training sessions. 3.34 The project would aim to increase the participation of women in group training sessions bys a) better sensitizing extension officers on the importance of the role of women in agriculture and the need to reach them with the extension messages. In this regard specific topics on the role of women will be introduced in the training sessions for extension agents and agricultural officers throughout the extension structure (para. 3.36); b) scheduling extension training session3 at times convenient to both the women and their husbands. Initial consultations should be held with farmers on this at the beginning of the project. Extension agents will also try to reach women while they are  - 31 - working on their farms. In some cases informal women cooperatives working together on the farm would make good contact groups; c) sensitizing husbands on the need for their wives to attend extension sessions; d) introducing messages of special interest to women such as legume and fruit culture, nutrition and care of livestock; e) in parallel with the concours agricole the annual communal agricultural competition for the best farmer, the project would introduce a concours femme designed to reward women who have best put extension messages to practice. A lump sum of about US" 860 per commuine would be provided to purchase items of particular interest to women as prizes. Women cited their need for agricultural tools and processing equipment. The ultimate objective would be to raise womens' participation in group sessions up to 75% of that of men by the end of the project. The extent to which the extension service is reaching women will be carefully monitored by the extension service internal monitoring and evaluation system (para. 3.23). Without setting quantitative targets, the project will encourage employment of greater numbers of women BEAS and other officers within the extension structure. 3.35 Within the promotion component efforts would be made to identify and include at least five womens' cooperative Rroups out of the twenty to be trained in bookkeeping and other techniques in each prefecture. The groups would be assisted to obtain credit for joint ventures in marketing, farming and processing. Consultants would be brought in to teach new techniques on the processing, conservation and cooking of manioc, soybeans, and cereals. This would be complemented with regional study tours for the romen on the same or related topics. Trainina 3.36 Training is of the utmost importance for the success of the project and will be provided in the three areas of extension, adaptive research and monitoring and evaluation. Training will be geared to improving the technical knowledge of all staff and improving thei- communication skills. Some seminars and training sessions will be designed for local and other ministry officials to explain the rationale for the project, its expected mode of functioning and the role of the officials and others in its success. Much of the training will be in-country and on the job. For extension agents training will consist of fortnightly training sessJ.ons, monthly workshops and in-country study tours. Higher level extension and adaptive research officers will also undertake regional and overseas study tours. During the first project year, when a minimum 70% of the BEAs are in place, intensive orientation/training sessions would be provided. The project will finance 7 man-weeks of technical assistance to help design and implement the initial training sessions. A detailed training program for the first artd subsequent project years is attached in Annex 2. The content of the initial sessions on extension would include: the rationale for setting up the service, pr.nciples of the T&V system, description of posts and tasks, discussion of work plans, extension themes, communication methods, the role of women, and the need for continual training. For adaptive and on-farm research (for which ISAR staff would participate in the training) topics would include: translation of the products of research to the field, diagnosis of farm level problems, involving farmers in problem solving, setting up and follow-up of experimental groups. For monitoring and evaluation focus would be on explaining the rationale for monitoring and evaluation, ir'egration of monitoring and evaluation ectivities in the work program, choice of indicators, monitoring  - 32 - surveys, diagnostic studies, and evaluation. About RF 53 million (US$ 687,000) would be provided for in-country training while RF 27 million (US$ 348,000) would be spent on external training. Technical Assistance and Studies 3.37 The project includes very little technical assistance. A total of about US$ 1 million will be provided to fund a long-term position in financial management (three years), and several man-months of short-term assistance for monitoring and evaluation, adaptive and on-farm research, seeds, and training. Several studies on marketing, processing and storage of agricultural products, introduction of different fruits and legumes are planned. About US$ 534,000 has been provided for these studies. Consultants will be hired according to IDA guidelines. Terms of reference for the studies must be agreed by IDA and GOR. Assurances were obtained on this at negotiations. S. Prolect Costs 3.38 Total project cost (including direct taxes of about US$ 1.3 million) is estimated at USS 30.1 million (RF 2,320 million), of which US' 11.2 million or 372 is foreign exchange. Indirect taxes will be waived by GOR. Costs are baned on quantity and unit price estimates prepared by MINAGRI. Unit price estimates for vehicles, equipment, and civil works are in line with past project experience and recent civil works contracts. Prices are as of October 1988 updated to the negotiating date of April 1989. 3.39 Price contingencies amount to 15S c! total base cost plus phvsical continAencies and have been calculated on the basis of foreign inflation rates of 3.31 for 1988, 3.72 for 1989, 4.1% for 1990, 5.7X for 1991, 4.9X for 1992, 4.75X for 1993 and onwards; and local inflation rates of 62 for 1988, 51 for 1989, 41 for 1990-94. Physical contingencies of 101 have been included for selected project items such as civil works and equipment. Project costs are summarized in Table 3.1. Details are presented in Annex 3 and in the implementation volume.  - 33 - Table 3.1 Aatricultural Services L!roiect Proiect Cost Suumnarv