Malawi - Agricultural Extension and Planning Support Project

Document of The World Bank FOlR OFFCIAL USiE ONLY Report No. 13633 PROJECT COMPLETION REPORT MALAWI AGRICULTURAL EXTENSION AND PLANNING SUPPORT PROJECT (CREDIT 1626-MAI) OCTOBER 25, 1994 Agriculture and Environment Division South Africa Department Africa Regional Office 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 witbout World Bank authorization.  CURRENCY EQUIVALENTS 1985US$ 1.0 = MK 1.72 1986US$ 1.0 = MK 1.86 1987US$ 1.0 = MK 2.21 1988US$ 1.0 = MK 2.56 1989US$ 1.0 = MK 2.76 1990US$ 1.0 = MK 2.73 1991US$ 1.0 = MK 2.80 1992US$ 1.0 = MK 3.60 1993US$ 1.0 = MK 4.39 (Av. Jan.-Sept. 1993) 1994US$ 1.0 = MK 7.0 (Av. Apr.-June 1994) WEIGHTS AND MEASURES Metric System GOVERNMENT FISCAL YEAR April 1 - March 31  FOR OFFICIAL USE ONLY ABBREVIATIONS ACB Agricultural Communications Branch ADD Agricultural Development Division ADMARC Agricultural Development and Marketing Corporation AGREDAT Agricultural Economics, Statistics and Data Processing Unit ARC Agricultural Research Council ART Adaptive Research Teams ASP Agricultural Services Project BES Block Extension System CAETO Chief Agricultural Extension and Training Officer CAO Chief Agricultural Officer CARO Chief Agricultural Research Officer CAS Controller of Agricultural Services CIMMYT Wheat and Maize Improvement Centre CPO Chief Planning Officer CTL Commodity Team Leaders DAR Department of Agricultural Research DCA Development Credit Agreement DCAO Deputy Chief Agricultural Officer DCARO Deputy Chief Agricultural Research Officer DEVPOL Development Policy DO Development Officer DOA Department of Agriculture EP&D Department of Economic Planning and Development EPA Extension Planning Area ERR Economic Rate of Return FA Field Assistants FAO/CP FAO/World Bank Cooperative Programme FHA Farm Home Assistants FSNU Food Security and Nutrition Unit GOM Government of Malawi ICRISAT International Crop Research Institute for the Semi-Arid Tropics IFAD International Fund for Agricultural Development IITA International Institute for Tropical Agriculture INIBAP International Institute for Banana and Plantain Research ISNAR International Service for National Agricultural Research LLDP Lilongwe Land Development Progmrame MADD Mzuzu Agricultural Division MAEPS Malawi Agricultural Extension and Planning Support Project MARE Malavi Agricultural Research and Extension Project MOA Ministry of Agriculture MOF Ministry of Finance NARP National Agricultural Research Project NRC National Research Coordinators NRDP National Rural Development Programme OFD On Farm Demonstration OSU Oregon State University PCP Pilot Communication Programme PCR Project Completion Report PD Planning Division (MOA) PO Project Officer RDP Rural Development Project RMEA Regional Mission in Eastern Africa SACA Smallholder Agricultural Credit Administration SAR Staff Appraisal Report SDR Special Drawing Rights SFFRFM Smallholder Farmers' Fertilizer Revolving Fund of Malawi SMS Subject Matter Specialist T&V Training and Visit TA Technical Assistance TOR Terms of Reference TOT Training of Trainers TRFCA Tea Research Foundation of Central Africa UNDP United Nations Development Programme USAID United States Agency for International Development This document has a restricted distribution and may be used by recipients only in the perfornance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation October 25, 1994 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on Malawi - Agricultural Extension and Planning Support Project (Cr. 1626-MAd) Attached is the Project Completion Report on Malawi Agricultural Extension and Planning Support Project (Cr. 1626-MAI) prepared by the Africa Regional Office. Part II was contributed by the Borrower. While in general agreement with the conclusions of Parts I and II, the Borrower mentions some difference on points of detail and also correctly notes that the Staff Appraisal Report did not make specific recommendations concerning the poorer smallholders who constitute 55 percent of farmers and operate one third of arable land without access to significant credit; failure to assist this majority was mentioned as a disappointing result in Part I. The project aimed to improve the planning capacity of the Ministry of Agriculture (MOA), improve the planning, budgeting and management in its extension department, enhance staff quality through training, and improve extension services in a large administrative region (Mzuzu); the latter also received equipment, infrastructural development and additional credit. The MOA institutional improvement was reasonably successful but has been plagued by senior staff retention problems. Improvements in the extension department proved difficult to implement; at the end of the project a planning and budgeting system was installed but its impact is as yet unknown, and the intended improved career structure for extension staff was abandoned. On-farm demonstrations were successfully introduced to the existing modified T&V system of extension in Mzuzu, and even during the project period this innovation spread to the rest of the country. However, there was an unexplained reduction in the number of on-farm demonstrations in the last two project years. Mass media programs were enhanced with USAID support, and there was an increased involvement of women in extension and credit activities. The extension system, however, was affected by serious operational funding constraints; there is room for rationalization of staff numbers and services; research-extension linkage is deficient and site specificity in fertilizer recommendations needs to be improved; and little attention is being given to those without access to credit (as has been traditional in extension in rural development projects in Malawi). Data to quantitatively assess impact even in Mzuzu were unavailable, but the extension system did at least contribute to the substantially expanded use of hybrid flint maize and higher nutrient fertilizers. The project is rated as having a satisfactory outcome, with a modest institutional development impact. Sustainability is uncertain, especially considering the funding problems. A follow-on project is attempting to address the identified problems and should at least temporarily alleviate the operational support constraint. The quality of the Completion Report is satisfactory. An audit is not immediately planned, but the project is likely to be audited together with other research and extension interventions in Malawi in the future. Attachment This document has a restricted distribution and may be used by recipients only In the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   PROJECT COMPLETION REPORT MALAWI AGRICULTURAL EXTENSION AND PLANNING SUPPORT PROJECT (CREDIT 1626-MAI) TABLE OF CONTENTS PREFACE ....................................................... i EVALUATION SUMMARY ........................................... iii PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE ................. 1 1. Project Identity . .............................................. 1 2. Background ................................................. 1 3. Project Objectives and Description .................................. 2 4. Project Design and Organization .................................... 2 5. Project Implementation .......................................... 3 6. Project Results . .............................................. 11 7. Project Sustainability ........................................... 14 8. Bank's Performance .15 9. Borrower's Performance ......................................... 16 10. Lessons Learned . .............................................. 16 11. Project Relationship . ............................................ 18 12. Consultancy Services ........................................... 18 13. Project Documentation .......................................... 18 PART II: PROJECT REVIEW FROM THE BORROWER'S PERSPECTIVE .... ........ 19 PART III: STATISTICAL INFORMATION ................................. 27 TABLES 1. Related Bank Loans and/or Credits ..................27 2. Credit Time Table........................28 3. Credit Disbursements.......................29 4. Project Implementation ......................30 5. Project Costs and Financing.....................35 6. Project Results.........................37 7. Status of Covenants .......................38 8. Use of Bank Resources ......................40  ANNEX 1 - TABLES 1. Comparison of Target and Achievement Dates ........ ................... 43 2. Extension Service Demonstration Data ........... ..................... 44 3. Mzuzu ADD Credit Operations ..................................... 45 4. Inflators and Exchange Rates ...................................... 46 5. Incremental Maize Production (1983/84 - 1992/93) ........................ 47 6. Switch of Fertilizer Mix: Estimated Savings in Fertilizer Imports Expenditures in 1992/98 48 7. Investment Costs for Economic Analysis .......... ..................... 49 8. Smallholder Sector: Maize Area and Yields ........ .................... 50 9. Import Parity Prices ......................................... 51 10. Budget per Hectare for Local and Hybrid Maize ....... ................... 52 11. Mzuzu ADD Net Value of Maize (Local and Hybrid) Production in Economic Terms . .. 53 12. Economic Rate of Return Estimated: Mzuzu ADD ........................ 54  PROJECT COMPLETION REPORT MALAWI AGRICULTURAL EXTENSION AND PLANNING SUPPORT PROJECT (CREDIT 1626-MAI) PREFACE This is the Project Completion Report (PCR) for the Malawi Agricultural Extension and Planning Support Project (MAEPS) for which Credit 1626-MAI in the amount of SDR 11.7 million (USS 11.6 million) was signed on December 20, 1985. The credit became effective on 20 March 1986. Originally expected to be implemented over a five-year period and closed on September 1, 1991, the credit was actually closed, after two extensions totalling 22-months, on 30 June 1993. Actual disbursement totalled SDR 11.2 million (US$ 15.0 million) or approximately 96%, in SDR terms, of the originally approved credit amount. The last disbursement was made on October 1, 1993. Parts I and HI of the PCR were prepared by the FAO/World Bank Cooperative Programme (FAO/CP) for the Agriculture and Environment Division of the Southern Africa Department (AF6AE) of the World Bank. Part II is being prepared by the Borrower. Preparation of the PCR was started during a visit to Malawi by a teaml from FAO/CP in December 1993, and is based,.inter alia, on information gathered in Malawi from discussions with project staff and officials of the Planning Division and the Department of Agricultural Extension and Training within the Ministry of Agriculture and with officials from USAID which participated in project co-financing. Meetings were held with officials of the Department of Economic Planning and Development and the Smallholder Agricultural Credit Administration. The report also takes into account information and data gathered from the Staff Appraisal Report (SAR), the Development Credit Agreement (DCA), Supervision Reports, USAID Agricultural Research and Extension Project (MARE) final evaluation report, correspondence between IDA and the Borrower and internal IDA memoranda. 1 Messrs. R. Suppa, Economist and Mission Leader; G. Stem, Agricultural Extension Specialist (Consulant).   PROJECT COMPLETION REPORT MALAWI AGRICULTURAL EXTENSION AND PLANNING SUPPORT PROJECT (CREDIT 1626-MAI) EVALUATION SUMMARY Objectives The project, signed in December 1985, formed part of the fifth phase of the National Rural Development Programme (NRDP) which included the National Agricultural Research Project (NARP) (Credit 1549-MAI) and the USAID financed Malawi Agricultural Research and Extension Project (MARE), which provided parallel financing to the two IDA Credits. The project aimed at supporting Malawi's institutional development in the agricultural sector with the specific objectives to: (i) improve MOA's analytical and long-term planning capability; (ii) strengthen the National Extension Service including the establishment of a mechanism for better defining extension strategies, priorities and programs and rationalizing existing structures; and (iii) develop the human resources needed for an efficient extension programme. The following is a description of project components: (i) strengthening the Ministry of Agriculture's (MOA's) Planning Division (PD); (ii) improving the national agricultural extension system, including developing an extension planning and budgeting system; (iii) upgrading the Department of Agriculture's (DOA's) technical capacity by establishing a professional career stream and promotion system for extension staff, and in-service training program; (iv) producing and distributing of teaching and training materials; (v) improving extension services in the Mzuzu Agricultural Development Division (ADD); (vi) providing incremental credit for farmers in the Mzuzu ADD; and (vii) improving research-extension linkages through the participation of extension staff in Adaptive Research Teams (ARTs), research-extension workshops and the development and dissemination of locally-specific technical recommendations. Implementation Experience The project was originally scheduled to be completed by September 1, 1991 after an implementation period of five years. Although investment planned at appraisal and subsequent modifications were substantially completed on time, the project was extended to 1993 to bridge the gap between it and a follow-up Agricultural Services Project (ASP) which supports both national research and extension systems. Major implementation constraints included: (i) senior staff vacancies and changes; (ii) inadequate project management and coordination, with the Steering Committee established for that purpose meeting only infrequently; and (iii) weak procurement and financial services, that seriously delayed equipment and operating funds for extension work. The constraints were to some extent alleviated by implementation assistance provided by Bank staff, particularly in the fields of extension, disbursements, finance and procurement. Changes to project design agreed upon during implementation included: (i) a pilot operation to improve field procedures and practices of the extension services, that started on a very small-scale in Mzuzu ADD and spread to all parts of Malawi, this, rather than planning and budget improvements, became the iii  centerpiece of the project; (ii) Technical Assistance (TA) for the Food Security and Nutrition Unit, newly established in the Department of Economic Planning and Development; and (iii) provision of operating funds for new extension activities (i.e. expansion of the extension methodology used in Mzuzu ADD to the rest of the country), by amending the Development Credit Agreement (DCA) and using previously unallocated funds for the new disbursement category. Results and Impact Overall, project performance was generally satisfactory, and the outcomes were mixed. Observable results and impact suggest that the project made significant progress toward achieving many of the original objectives (training, production and distribution of teaching and training materials, infrastructure development, improvement of the extension system in Mzuzu ADD, provision of incremental credit to farmers in Mzuzu ADD) and was disappointing in other aspects (planning and budgeting system, professional career stream, research-extension linkages). Technical assistance and 12 masters degree fellowships have strengthened the Planning Division's analytical and planning capacity, but improvements were constrained by continuing senior staff vacancies and turnover. Of the various interventions to strengthen extension, an improved planning and budgeting system has been developed and adopted, but so late that the impact is not yet apparent. The two USAID funded components, the Women's and Communications Programs were successful. The Women's Programme helped to formulate policy and procedures, including the change of emphasis from home economics to equal attention being paid to agricultural production activities. The Communications TA, training and equipment gave the Agricultural Communications Branch (ACB) of the MOA a much strengthened desk top publishing capacity, and the ACB now conducts regular radio programs for a wide audience and produces advisory leaflets and other printed material on demand. A pilot scheme designed during implementation, to improve extension methods, procedures and field training, spread to all parts of the country and has been adopted nationally. However, there was no funding for vehicles, equipment and operating funds needed for that purpose. As a result, a better system is in place, but some shortcomings identified during implementation remain to be addressed, including funding and transport problems, failure of extension to reach the poorer smallholder sector and training weaknesses. With regard to upgrading DOA's technical capacity, the Government was not able to introduce an improved career stream, the fellowship and massive local training programme was implemented efficiently, but the latter proved so costly that it could not be completed for most field level staff. The Adaptive Research Teams, which received support from both this project and NARP, did not improve the research/extension linkage in Mzuzu ADD. The complementary credit component expanded lending seven fold in current terms, but almost 45% of total lending has not been recovered. This was not a local but a national trend. However, the fact that despite the Mzuzu investments, extension performance does not appear to be sustainable after very promising initial improvement, must cause concern. The Food Security and Nutrition surveillance scheme introduced into the project after appraisal, has become a vital tool for the Government to monitor national food supplies and nutrition status and design remedial actions if and when needed. Additional contributions of the project include: (i) a major strengthening of Malawi's smallholder extension delivery system throughout the country; (ii) contribution of the extension service to the spread of hybrid maize from about 70,000 ha before the project (average 1984/85- 1985/86) to about 330,000 ha in 1992/93. The hybrid expansion had increasing impact on production both in the 1991/92 drought, when the 16% of the area sown to it accounted for 40% of the harvest, and in the 1992/93 "boom" year, when 25% of the area under hybrids accounted for 50% of the iv  record two million tons maize crop. However, the hybrid maize area dropped significantly in 1993/94 (to about 175,000 has), primarily due to the collapse of the rural credit system; and (iii) contribution of the extension service to the conversion by farmers from using low nutrient content fertilizers to those with high content. The conversion to high nutrient content fertilizers resulted in considerable foreign exchange savings. Limited availability of empirical evaluation makes it difficult to provide conclusive evidence of production impact at this time, although the accrual of project benefits may be forthcoming as the improvements in the extension system are fully implemented. Sustainability The project, in line with its objectives, tried to raise the effectiveness of the extension service. However, with operating fund shortages that have handicapped the extension service during the project period and with worsening cuts in MOA's revenue budget, the sustainability of the project's main activities is in doubt without greater and sustained Government commitment and continued donor support (during the 1990s) for operating costs, including vehicle and equipment replacement. Cost economies, in addition to improvements to be gained from better planning and budgeting, must therefore be seriously and constantly considered and implemented, including detailed review and streamlining of existing staffing and services to improve staff working conditions. Ways of doing so without affecting extension performance materially should be explored. It is important to note that the Agricultural Services Project (ASP) is supporting various actions to address these issues and develop and implement workable solutions. Lessons Learned The major lessons to be derived from the implementation of the project are the following: a comprehensive pre-project/pre-program analysis within country of farming systems and technology needs of the main farmer types, the available stock of technology, and the technology generation capacity are needed to ensure cost-effective and sustainable interventions; - an improved program budget system for recurrent cost funding should be developed and implemented within MOA, and this should be coordinated with an ongoing review of staff levels and job description; - more thorough assessment of the borrower's financial and procurement capabilities is needed during project preparation/appraisal, as well as operational requirements for their strengthening; - it is important to establish effective project management and coordination arrangements before project implementation; - field supervision of financial, technical and management aspects of projects with IDA donor assistance is extremely important; - a detailed review of the various NRDP is needed to ascertain whether there are sustainable and effective alternatives to cost-intensive staff and operating requirements of the extension system; v  greater attention needs to be devoted to ensuring that improvements in extension methods are complemented by ensuring strong research-extension-farmer linkages; the importance of separating extension and credit activities to ensure extension workers devote their time to extension activities; and - monitoring and evaluation of project activities are essential in order to communicate the positive results of extension activities to Government and obtain Govermnent commitment/support (particularly financial). The design features of the follow-up project ASP, have taken into account all of the above lessons. It is important that ASP implement effectively the various action plans prepared to address the key lessons from MAESP. Regarding improved and sustainable extension services, ASP is giving emphasis to: improved mechanisms to ensure a constant flow of improved technologies to a broader range of smallholders through greater emphasis on low-cost technical messages and more effective researcher-extension-farmer linkages; improved MOA budgeting and planning procedures and capacities, including development and application of improved 'yardsticks" for determining the level of non-salary operating funds for ensuring effective extension services; and improved monitoring and evaluation of the effectiveness and impact of improved extension services. vi  PROJECT COMPLETION REPORT MALAWI AGRICULTURAL EXTENSION AND PLANNING SUPPORT PROJECT (CREDIT 1626-MAI) PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE 1. Project Identity Project Name Agricultural Extension and Planning Support Project Credit No. 1626-MAI RVP Unit AF6AE Country Malawi Sector Agriculture 2. Background Introduction 2.1 The project was identified by IDA in early 1983, prepared by the Ministry of Agriculture (MOA) Planning Division assisted by Bank staff, and appraised by the Bank in September/October 1984. The Credit Agreement was signed on December 20, 1985 and the project became effective on March 20, 1986. The project formed part of the fifth phase of the National Rural Development Programme (NRDP) which also included the National Agricultural Research Project (NARP, Credit 1549-MAI) and the USAID-financed Malawi Agricultural Research and Extension Project (MARE), which provided parallel co-financing to the two IDA Credits. The Agricultural Sector 2.2 Agriculture is the backbone of Malawi's economy. Up to 1990, it accounted for about 40% of GDP, and 85% of exports and employment. Maize is by far the most important crop in terms of cultivated area, number of growers, and food security. Other major food crops include groundnuts, pulses, cassava and rice. Exports are dominated by three crops: tobacco, tea and sugar. Production is predominantly under rainfed conditions. Malawi's agriculture is characterized by smallholders and estates, differentiated by regulations concerning production, marketing, pricing and land tenure. While estate production focuses mainly on export crops, the smallholder sector accounts for about 80% of food production and only 10% of exports. It involves about 1.8 million families operating under customary tenure on about 1.75 million hectares, producing primarily for subsistence. Land and labor productivity among the majority of smallholders is very low, and nearly half of the population lives below the level of absolute poverty. High population growth, low productivity of land and labor in agriculture, limited employment opportunities, low quality of human capital due to low nutritional levels and limited access to education and health services, are the main causes of poverty. Ibis 1  situation was, at project preparation time, further aggravated by the failure of research to find high- yielding maize varieties acceptable to smallholder farmers, inadequate supply and use of fertilizers and a narrow range of high-value cash crops available to smallholders. The Agricultural Development Policy 2.3 As outlined in its "Development Policy (DEVPOL) 1987", the Government accorded priority to promoting agricultural growth. The primary objectives for the agricultural sector were to ensure food security at the household and national levels, improve rural incomes and increase export earnings, while protecting the environment. In addition to increasing agricultural productivity, the strategy envisaged policies and investment measures to expand off-farm employment opportunities. The donor community, including the Bank, supported this Government strategy. 3. Project Objectives and Description 3.1 Project objectives were described in the Development Credit Agreement (DCA) as follows: (i) improve MOA's analytical and long-term planning capability; (ii) strengthen the national extension system, including the establishment of a mechanism for better defining extension strategies, priorities and programs and rationalizing existing structures; and (iii) develop the human resources needed to put in place a more efficient extension programme. 3.2 The following is a description of project components: (i) strengthening the Ministry of Agriculture's (MOA's) Planning Division (PD) through training and consultant services; (ii) improving the national agricultural extension system, including developing an extension planning and budgeting system; (iii) upgrading the Department of Agriculture's (DOA) technical capacity by establishing a professional career stream and promotion system for extension staff, and an in-service training programme; (iv) production and distribution of teaching and training materials; (v) improving extension services in the Mzuzu Agricultural Development Division (ADD)' by constructing offices, staff houses and marketing sheds, and purchasing vehicles and motorcycles for extension managers and supervisors and equipment and materials needed by extension staff; (vi) providing incremental credit for farmers in the Mzuzu ADD; and (vii) improving research-extension linkages through the participation of extension staff in Adaptive Research Teams (ARTs), research-extension workshops and through the development and dissemination of locally-specific technical recommendations. The DCA allowed for component modifications mutually acceptable to the Government of Malawi (GOM) and IDA. 4. Project Design and Organization 4.1 Initial project preparation followed NRDP lines, to strengthen staff and infrastructure in areas not covered by earlier projects. However, during preparation it was believed that, given Government budgetary constraints, priority should be given to consolidating and improving maintenance of existing NRDP investments. Both GOM and the Bank considered that there was already a basic framework for satisfactory extension in Malawi, which needed considerable strengthening by improving: cost effectiveness, extension planning and evaluation, the staff's technical capacity and linkages with research. Consequently, project design focused on consolidation and institutional improvements of the agricultural planning and extension systems. However, planning aspects went beyond extension and were designed to have the Planning Division provide sector strategy and advice I The Mzuzu ADD, as Lhe second largest ADD and representative of most of the country's agricultural area, was selected to pilot the first phase of a national programme to improve extension service (SAR). 2  for policy formulation and determine investment priorities. PD was also expected to improve project preparation, and monitoring and evaluation services. Orthodox NRDP investments (paras 3.2(iv) and (v)), accounting for 30% of project cost, were designed for Mzuzu Agricultural Development Division only, which was to serve as a pilot area for improved extension planning and budgeting and resultant modifications (improvement) of the extension service; and for a "mass communication system" that was not described in project documents. Technical assistance and training, totalling 55% of project cost, was envisaged to cover the entire extension service. 4.2 Overall responsibility for project implementation was vested in the MOA Principal Secretary (PS), assisted by the Project Steering Committee, comprising the two Controllers of Agricultural Services (CAS), one for the NRDP and one for institutions; the Chief Agricultural Officer; the Chief Veterinary Officer; the Chief Agricultural Research Officer, the Chief Projects Officer, the Chief Accountant; and a Ministry of Finance (MOF) representative. Procurement services were to be provided by the Consultant funded by NARP, and the same firm of architects, already appointed under NARP, was to be entrusted with supervision of the civil works. The USAID contribution to the project under the MARE project, which had been approved in July 1985, was managed by the Consortium for International Development, a consortium of four universities, with Oregon State University (OSU) serving as the lead institution. Design Weaknesses 4.3 In retrospect, there were design shortcomings that led to implementation difficulties or changes, notably: (i) a weak project management structure (para 5.31); (ii) inadequate provision for procurement of goods (para 5.29); (iii) inadequate analysis of extension methods in use in the field (paras 5.10-5.15); (iv) inadequate analysis of and arrangements for ensuring adequate operating fund shortages (paras 5.13 and 7.1); and (v) inadequate attention to ensuring effective research-extension- farmer linkages. 5. Project Implementation General 5.1 Implementation of IDA-financed components started slowly, despite two launch workshops, which were held a year apart, and labelled successful by participants. Action plans for implementing each component were drawn up, but these seemed to have little impact on subsequent implementation (Annex 1, Table 1). The main reason for the slow start was a change of virtually every senior MOA manager, including the PS, the two CASs, and heads of DAR, DOA and the Planning Division. By contrast, USAID-funded components, managed by OSU (para 4.2), made good progress. 5.2 Three key developments emerged from the Project. The first of these was initiation of a pilot operation in Mzuzu ADD to improve field extension, which spread to the rest of Malawi and became the centerpiece of the project (para 5.10). The second was the establishment of a Food Security and Nutrition Unit (FSNU) in the Department of Economic Planning and Development (EP&D) in the office of the President and Cabinet, and agreement by the Bank to finance TA to support it (para 5.26). Finally, at GOM's request, the Bank agreed to some changes of TA to be financed by the Project (para 5.4). 5.3 Implementation was constrained by procurement (para 5.29) and financing problems (para 5.30) and by lack of project direction/coordination/monitoring (para 5.31). Nevertheless, most of the project components were completed as expected at appraisal. The closing date was extended twice, by 3  a total of 22 months, mainly to ensure continuation and replication to other ADDs of the extension pilot operation until the work could be funded by the follow-up project Agricultural Services Project (ASP). Implementation of Individual Components 5.4 Strengthening the MOA Planning Division. PD has five sections, respectively responsible for: (i) price policy and marketing; (ii) project preparation; (iii) surveys, statistics and evaluation; (iv) planning and budgeting; and (v) data processing. The project was to finance about 84 months of long-term TA for a Planning Adviser, Price Policy Economist and Systems Analyst, and 12 months of short-term TA. Other investments included six Masters fellowships and six vehicles plus office equipment for the Division. Project activities were delayed because the Chief Planning Officer (CPO) post was vacant for the first year (para 5.1). Two further changes of this post and staff vacancies, some due to staff absences on training, impeded the Division's work, but ultimately the components were implemented more or less as planned with only a few exceptions. EEC provided the planner on a grant basis and MOA requested and IDA agreed to finance a monitoring and evaluation specialist instead. At the end of the project, 90 person-months of long-term TA were delivered. Figures on short-term TA were not available. The overseas graduate training programme was doubled, and there were 12 Masters fellowships instead of the six projected in the SAR. Furthermore, some 156 person- weeks were provided to the Planning Division under the short-term (off-shore) training component against the 100 foreseen at appraisal (Part III, Table 4). 5.5 The quality of the TA provided during project implementation has varied. Monitoring and Evaluation TA was poor and is reflected by the lack of extension or project impact studies. The first national survey of extension impact, conducted in early 1993, had not yet been collated and analyzed at the time of PCR field work. By contrast, the systems analyst is credited with designing the MOA computerization programme, conducting effective training, and designing the systems that remain in use today. Of the two pricing policy advisers, the first was so difficult to manage that he had to be replaced. The second performed his pricing policy evaluation duties competently, but it is difficult to judge whether there was a lasting impact of his work on the Division. 5.6 The enlarged masters training programme (para 5.20), supplemented by short-term training for 26 PD staff (156 person-weeks) disrupted the work of the Division during the project period, and the performance of the Division remains constrained by staff vacancies, particularly at Principal Economist level. The Division was to prepare a five-year Agricultural Strategy and a draft was provided to IDA for review over two years after the target date (Annex 1, Table 1). Improving National Extension System 5.7 General. The SAR agreed with the 1982 NRDP review that a satisfactory institutional framework to deliver extension services was in place, but that effectiveness of the extension service was increasingly hampered by inadequate numbers of properly trained technical and management staff; absence of structured career development opportunities or incentives for high achievement; paucity of technical packages for smallholder crops; inadequate extension planning and budgeting systems; and unsatisfactory and scarce extension materials. The project was designed to address these shortcomings as well as to address the need for special work to help the many women farmers. Specific investments for these purposes (Part III, Table 4) were to be mainly: (i) 276 months of TA for two Training Specialists, a Women's Programme Adviser and a Communications Specialist, both financed by USAID; and an Extension Planning Adviser, Financial Controller and Seeds Specialist, financed by IDA, plus 74 months of short-term TA; (ii) the US$ 3.5 million staff training 4  programme, comprising 10 Masters fellowships, 34 diplomas, 120 weeks of short-term training abroad, plus some 13.000 weeks of in-country training; and (iii) 14 vehicles for DOA Headquarters and 4 for the Agricultural Communications Branch (ACB), office equipment for both and printing, binding and audio-visual equipment for ACB. In general, USAID-financed vehicles, equipment and operating costs were to support its TA, in this case the ACB and the MOA Women's programme. At project completion, long-term TA amounted to 244 person-months (178 financed by USAID). Data on actual delivery of short-term TA were not available. The training component was successfully implemented and included 1 Ph.D., 23 MSc. and 20 diplomas. With regard to short-term training, actual delivery included 808 weeks and 850 weeks for off-shore and in-country courses respectively. Details on SAR targets and actual implementation are given in Part III, Table 4. 5.8 Extension Planning and Budgeting. The main purpose of the project was to strengthen the extension service by eliminating wasteful expenditures identified by improved planning and budgeting. The long-term TA for this purpose proved ineffective, as were proposals to strengthen impact by moving TA from the Accounts Department to either PD or DOA. Although DOA did not accept proposals resulting from TA, efforts to improve planning and budgeting continued. Short-term courses in the later stages of the project in the use of a logical framework approach to designing a 'Five-Year Extension Strategy", led to its adoption by all DOA units and all ADDs. However, adoption is so recent that the impact is not yet apparent. 5.9 Management Planning. The main output from the Extension Management Planning Adviser was the Five-Year Extension Strategy, published in January 1990, and amended in June 1992. The document became the policy guideline for all branches and programs of the extension service. Aninal husbandry work was not included because the Department of Animal Health and Industries was primarily responsible for related advisory work. The Strategy details resources required, but does not take into account the Government's budget constraints. 5.10 Improving Extension Methods, Procedures and Practices. Under NRDP, the country was divided into eight ADDs, each directed by a Programme Manager (PM); subdivided into 30 Rural Development Projects (RDPs), each under a Project Officer (PO), further subdivided into 147 Extension Planning Areas (EPAs), each supervised by a Development Officer (DO) and each of which was divided into sections. These represent the activity sphere of Field Assistants (FA) who are the frontline extension workers interacting with farmers. In 1991 there were 1,898 sections staffed by about 1,777 FAs2. In 1981, MOA introduced the Block Extension System (BES) to be standard for Malawi. This was a modified Training and Visit (T&V) system, entailing FA visits on a regular schedule to farmer contact points and bi-weekly FA training. For this purpose, FAs were to divide each section into 8 Blocks, each one to be visited on the same day, once every two weeks. The main extension tool was to be a demonstration plot, called the Block Garden, to be provided and tilled by farmers of the area. Recommended practices were to be demonstrated there and discussed with farmers by the FA on his visits, which it was hoped would attract most farmers. Farmers were encouraged to form "Block Committees" to help guide and organize extension work within their Blocks. FAs were supported by a range of Subject Matter Specialists, both at ADD and RDP level, and primary responsibility for their guidance and training lay with their immediate supervisors, the DOs. 5.11 The March 1987 Project Supervision Mission found that attendance at Block Gardens was low, which a Mzuzu ADD review ascribed to "monotony of subjects and failure of BES to cater for 2 Source: Five Year Extension Strategy Plan - MOA - June, 1992. 5  farmers' demands and aspirations". It was hard to tell what extension work other than visiting Block Gardens extension workers were doing. To address this situation, the mission recommended either replacing Block Gardens or supplementing them with demonstrations on farmers' land, to cater to their needs and interests. These findings led to subsequent regular visits of a Bank extension specialist. In October 1987, in line with appraisal objectives to experiment with improving extension in Mzuzu ADD, the extension specialist initiated a small pilot operation to strengthen BES in two of its EPAs. It included the establishment of demonstration plots on farmers' fields, allied to more structured FA work programs and intensified bi-weekly training. Because of the hostile reception to the March 1987 supervision findings by Mzuzu extension staff, the pilot operation was designed to minimize changes and reinforce existing practices rather than replace them. Although the new system was viewed with suspicion by managers and by DOA HQ staff, the field staff found it to be helpful and in 1988, it was expanded to other Mzuzu EPAs and to small parts of two other ADDs. By 1992 it had been adopted country-wide. 5.12 The rapid spread of the above methodology to the national scene had not been initially planned or anticipated, but the new system was popular in the field because it addressed the difficulty of establishing Block Gardens. In 1991, 10 years after BES introduction, only 32% of blocks had a Block Garden and it was difficult to tell what extension work FAs were doing in blocks without gardens. These gardens were especially difficult to establish in the densely populated South, where land is scarce, and hard to obtain for communal purposes. There and elsewhere, FAs found that farmers were interested in "On Farm Demonstrations (OFDs)" on their own land, reflecting their own interests. The Block Garden/OFD combination remained the most desirable option, with a range of recommendations shown on Block Gardens and specific ones, selected by specific farmers, on OFDs. However, OFDs enabled FAs to demonstrate improved practices, as foreseen in BES, even in areas without Block Gardens. By 1990/91, about 5,000 Block Gardens had already been reinforced by about 22,000 OFDs (Annex 1, Table 2). The Pilot Operation seemed to benefit even the establishment of Block Gardens which, by 1992/93 had increased to 8,000 and that of OFDs to 30,000. The latter is only about 30% of the minimum target of establishing five OFDs per block, but quite a commendable effort, considering that the project had not been designed to remodel extension procedures in the field. OFDs allied to block gardens, could become a powerful extension tool if the national target of operating 75,000 OFDs is attained. 5.13 Successful BES needs mobility for close FA supervision. A Chronic shortage of vehicles was exacerbated by a two-year delay in arrival of new motorcycles for the Mzuzu ADD. Special efforts to overcome breakdowns and lack of spares, which were the cause of most of the national MOA fleet being inoperative most of the time, brought only temporary relief. Recommendations for managing MOA transport (presently valued at MK 35 million), by an expert study, are to be dealt with by the successor project (ASP). Operating fund shortages, equally crippling to extension work, were somewhat alleviated by a DCA amendment that reallocated SDR 1 million of unallocated funds for extension operating costs (Part m, Table 5-B). However, it took two years from the DCA amendment for any funds to reach the field, and only SDR 678,000 of the allocation (which could have been replenished) was spent, despite dire fund shortage in the field, which seriously impeded extension work. 5.14 NRDP had established monitoring and evaluation (M&E) teams in each ADD. However, extension impact studies were conducted only in the early stages of the pilot operation, mainly by interested ADD-based Evaluation Officers. With staff changes, vacancies and, allegedly, also operating fund shortages, the studies stopped and there seemed to be little that the MOA Central M&E Unit could do to restart them. This defect received considerable attention during ASP preparation activities. The early studies showed that only credit club members, usually commercial 6  smallholders with larger holdings, attended block demonstrations, and that the same farmers were conducting OFDs. The majority of the farmers who are subsistence smallholders farming on less than I ha of land did not benefit from block demonstrations. It seems that despite the efforts of supervision and encouragement of Bank staff, there has been little change. As a result, in the 1993/94 crop year when, due to particularly poor loan recovery, the number of credit clubs and farmers entitled to loans was drastically reduced, the number of farmers interested in extension work and conducting OFDs also fell. One reason for farmers' loss of interest in FAs was the FAs' responsibility for collecting loan repayments. Since 1987, supervision missions have tried unsuccessfully to separate credit servicing from extension functions, but this delinkage was finally initiated on a pilot basis only in July 1992. However, it is expected that the credit operation will be transferred from ADDs to a new rural financial institution to be established under the recently approved Rural Financial Services project. Such delinkage would enable FAs to devote more time to providing improved extension services. Up until now many FAs spent more time on supervising credit and loan repayment than on demonstrating agricultural technology, an estimated more than 80 percent of their time was spent on credit-related activities. 5.15 By the time of project closure, the new extension methodology had been adopted as the National Extension Policy and had spread to all parts of the country. However, the faults identified during implementation, have not yet been eliminated, such as poor financing to meet incremental operating costs, lack of transport, inadequate supervision, weak training, and lack of success in expanding the extension clientele or evolve messages of value to poor farmers, unable to afford purchased inputs. Adequate funding has never been available for equipping and operating the whole niational extension service, but it remains disappointing that even ADDs that benefited from support by this and other projects were unable to overcome the above-listed problems. The recently approved ASP has been designed to address these problems. 5.16 Women's Programs and Communications. Both USAID funded consultants did good work that left considerable impact. The TA for Women's programs helped to formulate policy and procedures including the change of emphasis from home economics to equal attention to agriculture by Farm Home Assistants (FHA). Helped by some well focused short-term assistants, the TA helped to accelerate implementation of Income Earning Activities Programs. These policies and programs have continued to the present. The associated fellowships programs have strengthened the management and execution of the Women's programs. The result of intensified MOA Women's Programs has been the provision of extension and credit to many women farmers previously deprived of these facilities, as exemplified in Mzuzu ADD (para 6.5). 5.17 The Communications TA, training and equipment assistance gave ACB a much strengthened desk top publishing capacity. The sharply focused short-term training programs in various printing and publishing skills was very effective and could serve as an example for further training. ACB now conducts regular radio programs that have a wide audience, and on demand, supplies well designed and produced advisory leaflets and other printed materials. The consultant was also responsible for introducing a Pilot Mass Communication System for Mzuzu, which became known as the Pilot Communication Programme (PCP). In line with the SAR outline, this Programme focussed on 'visual presentation", which taught ADD staff how to formulate technical advisory leaflets. In addition, video cameras were provided to most ADDs. The Communications Programme provided important support to extension staff, it also helped the Pilot Extension Improvement Operation materially, but fell short in identifying a more cost effective alternative to BES. 5.18 Professional Career Stream. At appraisal time a major study of the entire Malawi civil service was expected to bring about comprehensive changes of terms of service, that could include 7 'U  improved career streams for MOA staff. This did not happen. Subsequently the Government insisted that isolated career improvements for MOA staff alone, were not possible because of the budget implications, and eventually the Bank waived the relevant DCA requirements. GOM is addressing these broader issues of career structure and incentives under a proposed institutional development project and a recently completed Bank subsector study on civil service pay and employment. 5.19 Training, together with TA, were the major investments expected to have most impact on improving PD and the Extension Service. The project established the Training Unit in MOA and consultants (USAID funded) did excellent work together with Malawi staff, to organize the massive MOA training programs arising from MARE, NARP and MAEPS. However, the Unit remained without a Chief Training Officer after the consultants' departure and became less well staffed and therefore with less capacity to serve MOA's training needs. There is now a recognition that the full Unit performed a vital service and it is likely to be strengthened. 5.20 Details of MAEPS training are shown in Part III, Table 4. In MAEPS, academic training abroad was entirely financed by IDA and reportedly resulted in 12 Masters degrees for planning staff (rather than 6 as planned) and one Ph.D., 10 Master degrees and 20 diplomas for DOA and extension staff instead of targets of one, eight and 34 respectively. This was supplemented by about 800 person- weeks of short-term off-shore training for about 100 staff (40% USAID financed). However, the main programme focus was on intensive in-country training (50% USAID financed), especially the Training of Trainers (TOT) programs. Training concentrated on senior and middle level management and supervision staff, resulting in roughly SDR I million higher IDA disbursement for training than scheduled at negotiations. Only a relatively small percentage of the 7,800 person-weeks training projected for FA staff in the SAR could be conducted, because of the high cost of training. 5.21 Improving the Mzuzu ADD Extension Service. The main investment was the civil works programme, which consisted of constructing two offices, 117 staff houses, 16 marketing sheds and five diptanks, and was completed as planned. Several problems were encountered. One contractor was slow and completion was delayed. More seriously, water supplies could not be provided at seven sites. The Ministry of Works Water Supplies Department failed to suggest alternative supply systems when trial boreholes failed. The 16 marketing sheds could not be used for inputs sales initially, because housing for store-keepers was overlooked. This did not come to light until the 1989 inputs supplies crisis (para 5.24), and was not corrected until 1993. However, the most serious complaint came from the staff, about project funded houses, arising from IDA insistence that costs based on standard Malawi designs, should be reduced by 20%. Complainants pointed out that rooms were too small, finish was unsatisfactory, kitchens and stoves were unusable or causing smoke to pollute the house; and that temperate designs with inadequate ventilation were built in the hot lakeshore areas. While a considerable amount of housing was provided, the users were unhappy. 5.22 Provision of vehicles. Altogether, 28 motor vehicles, 107 motorcycles and 100 bicycles were bought for Mzuzu ADD. Their arrival was seriously delayed by inept procurement by MOA. In particular, motorcycles ordered in early 1987 arrived later than expected and some of the replacement transport, ordered in 1991, had still not arrived at project closure. Since motorcycles are vital for field staff supervision, the delays undermined the effort to improve extension services. 5.23 Seed production at Mzuzu ADD was to be supported by a TA seeds specialist. Early during the Project, GOM felt and IDA agreed that there was no need for this TA. The modest bean and groundnut seed production programme did not expand, mainly because the difference between ware and seed prices was too narrow. The unsuccessful seed production attempt at Mzuzu, confirmed that 8  MOA supervised farmers cannot produce seed at significantly lower cost than seed companies, and that effective seed bulking is best done by such companies. 5.24 Incremental Credit. Incremental credit for seasonal loans Mzuzu ADD was channelled through the smallholder (SACA). Disbursement and repayment performance are shown in Annex 1, Table 3. Lending started slowly, began to expand gradually in 1987/88 and jumped dramatically in 1989/90, from MK 4.2 million to MK 10 million respectively. This came about because the Agriculture Development and Marketing Corporation (ADMARC) closed its retailing outlets, and with the 16 inputs sheds constructed by the project inoperative (para 5.21), distributed inputs only from a few major markets. As a result, most farmers found that they could obtain all their fertilizer requirements easier and much quicker through credit systems and changed from cash to credit inputs purchases. Until then, recovery was satisfactory but the ADD staff could not handle the sudden large increase in credit clubs and borrowers and this caused the first serious repayment default. Repayments improved only marginally during the next two years and then deteriorated drastically due to the 1991/92 drought. The recovery rate, slid even further in 1992/93 to as low as 11.4% due to a combination of overgenerous repayment moratoria after the drought, even for those who could have paid; another doubling of lending, which, like in 1989 could not be properly administered; the emergent politics giving rise to suggestions that loans need not be repaid, and the relaxation of lending criteria requiring 100% repayment by a club, to qualify for new loans. This led to about 43% overdues in 1992/93, amounting to MK 23.5 million. Total lending during the project period was MI 54.5 million, of which about MK 13 million was incremental (MK 5.8 million of this amount came from 1992/93 lending). Only MK 5.2 million were claimed from project funds, which translated into SDR 1.4 million or only 50% of the SDR 2.8 million allocation to this category. Additional funds were channelled to Mzuzu ADD from other resources including emergency funds released by the National Agricultural Research Project (Credit 1549-MAI) and from the Smallholders Agricultural Credit Project (Credit 1851-MAI and IFAD Loan 212-MW). 5.25 Initially, the credit component brought credit rationing to an end, but in the last project year no project resources were used to expand lending. There was an eightfold increase, in nominal terms, of seasonal lending, mainly used for hybrid maize seed and fertilizer. Part of this increase paid for higher inputs prices and much of it replaced cash purchases. To what degree the higher lending helped to expand production is difficult to gauge, but it helped enable farmers to continue using inputs. 5.26 Food Security and Nutrition Unit. Following food security studies by the World Bank and a 1986 seminar on the subject by Principal Secretaries, the Government decided to form a new Food Security and Nutrition (FSNU) in the Economic Planning and Development Department (EP&D) and requested Bank help to finance technical assistance to guide the new unit. FSNU was to design, evaluate and assist with implementation of food security and nutrition policies and programs; monitor food security and nutrition conditions; coordinate food security and nutrition activities; and coordinate Government's relations with donors with regard to food security and nutrition. Bank funding was used for a contract with the Harvard Institute for International Development to provide senior advisors for several three-week visits per year, and to provide training for local staff. Emphasis was on building local capacity and this was successful up to the point where, due to promotion opportunities, some key staff left the Unit. Consequently, after five years of Bank assistance, TA has continued, but on grant terms from other donors. 5.27 One reason for requesting further TA was the key role being played by FSNU in the country's food security system. Early on, FSNU assessments helped to mobilize donor support to avert disaster during the 1987 food crisis. Subsequently, a continuous food security surveillance and 9  assessment system has been installed, a strategy for managing strategic grain reserves established, options for meeting grain crises prepared and a food and nutrition data base established. The TA also collaborated with UNDP financed experts in similar nutrition oriented work. FSNU has become a key GOM unit in the Food Security and Nutrition field, but apparently cannot yet dispense entirely with foreign expertise. The importance of the FSNU is confirmed by its useful role in the droughts of 1991/92 and 1993/94. 5.28 Improving Extension Linkage was funded entirely from USAID, MARE funds. The design seemed ideal, with an Adaptive Research Team (ART) in each ADD, guided technically by the Research Department, and accountable for its work programme to the ADD Manager. Each ART was composed of an economist and agronomist and would conduct on-farm trials to generate location specific production recommendations and identify constraints affecting small farm productivity. Their level of performance, however, was below expectation. This was partly because of the excessive work load, but also because the economists who were needed for diagnostic work on farmers' economic constraints found more lucrative employment and left. Research-extension links were not improved nor were location specific recommendations generated. Nevertheless, the most important research findings, notably on flint maize hybrids, reached a large proportion of smallholders. It may well be that extension workers obtained the most important research findings at the excellent annual Research Station field days. With regard to local recommendations, quite a lot of field trials were conducted, but were not evaluated by Research Headquarters. Evaluation should be worthwhile and may result in useful production recommendations. Special Implementation Problems 5.29 Procurement became a problem even before the departure of the NARP funded consultant. His counterpart was not qualified for the work and this led to some contracts either not being awarded at all or not being delivered because of letter of credit or down payment problems. There were two changes of MOA procurement staff; in each case with staff without procurement qualifications and experience, and in each case claiming that the predecessor had left no hand-over notes or documentation on past and ongoing transactions. Typing and communication facilities provided were inadequate. Tender documents took months to prepare and required many amendments, even after short-term TA provided model bid documents; and skilled guidance for evaluations was lacking. The result was long delays in vehicles and motorcycle deliveries (para 5.22). Towards the end of the project there was some improvement with a more senior officer in place and provision of a computer and telephone, several training courses, and most importantly, the benefit from field visits by a knowledgeable Bank procurement specialist. As identified at a Project Implementation Review, there are very few procurement specialists in Malawi, pointing to the need for major training programs. 5.30 Finance and accounting problems became apparent early when credit withdrawal difficulties were encountered. Assistance from visiting Bank disbursement staff proved invaluable in overcoming the problems, but recommendations for considerable strengthening of MOA financial and accounting services were not followed. Then, the DCA amendment, establishing a new disbursement category for "Operating Funds", caused major problems. Since this category did not exist initially, there was no provision for it in the GOM Development Estimates and its introduction seemed to baffle MOA and Treasury staff. The eventual solution, two years after the DCA amendment, required an input from a Bank financial specialist. Even then, the MOA Accounts Department could not provide a satisfactory cash flow to the field (para 5.13). Almost concurrently, it became apparent that GOM Development Estimates were inaccurate. There was no mechanism for adjusting balances to exchange rate changes and the estimates did not reflect project off-shore spending (direct donor disbursements). The net effect was that the GOM Estimates often showed low or no balances of project funds while the Donor 10  showed considerable balances. As a result some project activities requiring local expenditures were 'cut' from the estimates due to the apparent lack of resources. The delay in constructing storekeepers' houses for Mzuzu marketing sheds (para 5.21) is a typical example. Workshops to improve the GPM estimates were suggested by Treasury staff and were to be funded by the project, they were arranged, postponed and ultimately abandoned. The present status of Development Estimates accuracy is not known. Under the ASP special emphasis is being given to addressing the above issues, especially through development of a yardstick as a basis for determining required non- salary operating funds, improved financial management procedures and associated in-service training. 5.31 Project Management. The Principal Secretary (PS) for Agriculture was given overall responsibility for project management, assisted by a Steering Committee. The PS and his Senior Manager, who had helped design the project, were moved soon after project approval (para 5.1) and the replacements were not familiar with the project and its management concepts. The new PS had little feeling of 'Project Ownership" and was probably not informed that he was the Project Manager and, among other things, responsible for convening the Steering Committee. In any case, he was too busy to undertake day-to-day project management tasks. As a result, the Steering Committee rarely met, and there was no overall project management or coordination. Consequently, PD, DOA, Mzuzu ADD and later also FSNU handled their components almost as free-standing projects. There was virtually no project reporting and only occasional submission of work programs. Above all, there was no central manager prepared to deal with financial, procurement and other problems. 5.32 With regard to extension management aspects, ADD managers reported directly to the Controller of Agricultural Services (CAS) of (NRDP) and were on the same seniority level as the Chief Agricultural Officer (CAO), who therefore had relatively little impact on field activities. That may be the reason why there was little push from MOA headquarters to introduce extension planning and budgeting reforms and why headquarters initially had little interest in the pilot operation to improve extension. This changed in 1992 when a new PS made ADD Managers responsible to CAO who was retitled Chief Agricultural Extension and Training Officer (CAETO), after which Head Office became an active participant in the field operations. While extension work benefitted from the line of control change, it left the direction of animal husbandry and health services undefined. Project Costs and Disbursement 5.33 The final cost of the project is estimated at US$ 23.0 million, or 14% above SAR estimates (Part III, Table 5-A). Credit disbursement lagged considerably behind SAR estimates averaging 55% of the amount expected to be disbursed throughout the project life. The last disbursement was made on October 1, 1993, when the total amount had reached SDR 11.2 million (US$ 15.0 million) representing 96% (if expressed in SDR) or 129% (if expressed in US$) of SAR estimates (Part m, Tables 3 and 5). This increased US$ amount reflected the appreciation in the value of the SDR vis-a- vis the US$ during the project period. 6. Project Results 6.1 Overall, project performance was generally satisfactory, and the outcomes were mixed. Observable results and impact suggest that the project made significant progress toward achieving many of the original objectives (training, production and distribution of teaching and training materials, infrastructure development, improvement of the extension system in Mzuzu ADD, and provision of incremental credit to farmers in Mzuzu ADD) and was disappointing in other aspects (planning and budgeting system, professional career stream, research-extension linkages). 11  6.2 The agricultural sector underwent two major changes during the course of the project. The first was the conversion from predominantly low nutrient content fertilizers to using mainly those with high nutrient content, resulting in about US$ 14 million savings per year on the import bill (Annex I, Table 6). The second and even more significant change was the rapid increase of the hybrid maize area from 80,000 ha during 1983184-1985/86, to 330,000 ha in 1992/933. The impact of hybrid maize (para 6.11), was high in the favorable 1992/93 season when the 330,000 ha (25% of the maize area) produced 50% of the record 2 million tons of maize crop. It was even more significant during the serious 1991/92 drought, when the 217,000 ha hybrids (only 16% of the maize area) produced 283,000 tons, over 40% of the 660,000 tons maize crop. Extension staff taught farmers to convert to new fertilizers, and similarly extension played an important role in the expansion of hybrid maize, but how much and what role the project played has not been measured, and is discussed further below. 6.3 The national pattern is mirrored in Mzuzu. In the 1991/92 drought season, hybrids, sown on 19% of the maize area, accounted for 50% of production, while in the "boom year" 1992/93, the hybrid maize area, which had expanded to 30% of the total, produced 60% of the record 180,000 tons crop. An indication of the benefits of hybrids is given in the recently published Draft Agricultural Sector Memorandum, which records that according to field surveys, converting one hectare of unfertilized local maize to hybrid, results in a net benefit of about US$ 120 equivalent, or about US$ 36 million for the entire hybrid maize area in Malawi, in 1992/93. 6.4 Two major MOA departments were to be strengthened by the project, the Planning Department and Extension Service. The graduate training programme strengthened PD's operational capacity, while the TA impact was mixed with the most visible and lasting impact being MOA's computerization. Continuing problems in filling and retaining key senior posts, to some extent, continue to constrain PD's performance. 6.5 The emphasis of the extension improvement component changed during the project. It proved difficult in the early project period to introduce a better planning and budgeting system. Such a system was adopted late in the project period and has not yet made an impact (para 5.8). 6.6 Review of the extension system in the field showed critical problems at the FA/farmer level (para 5.11). This led to the Pilot Operation to address the problems in Mzuzu, which aroused so much interest in MOA and extension staff in other ADDs, that it spread country-wide (para 5.11 and 5.12), and became the project's centerpiece. Such a development had not been anticipated or funded at project appraisal, and it is therefore laudable that such a system could be installed nationally, but not surprising that imperfections remain to be addressed. For instance, in Mzuzu ADD, where the Pilot Operation started, the new system was accepted enthusiastically and spread rapidly until 1990/91. Since then, a drop in OFD numbers (Annex 1, Table 2), indicates problems that have not yet been diagnosed. Consolidating the new system and dealing with the problems that were identified during the course of intensive field work (para 5.15), formed the core of ASP preparation. However, a measure of the difficulties encountered by MAEPS, indicates that several project cycles may be needed to solve all the problems. 6.7 The extension service has been censured for dealing mainly with credit club members of the "commercial smallholder sector". This sector, which comprises 45% of Malawi's farmers, and 3 Information received after field work for this report was completed shows that the hybrid maize area has dropped steeply during the current season, due to the breakdown of the smallholder credit system. As the credit constraints are resolved and the smallholders are able to gain access to cash crops and world market prices for their crops, the use of maize hybrids will expand significantly during the 1990s. 12  operates two-thirds of the arable land, has been largely responsible for the hybrid maize expansion. Extension has been working with up to 30% of Malawi's farmers, or almost 75% of the commercial smallholder sector, and has played a part in their agricultural improvements. However what part, and how much of that was achieved through credit-related visits to farmers and how much by the improved demonstration system (para 5.12), has not been measured. ASP preparation efforts will be looking into this issues. 6.8 It is disappointing that the new system has so far failed to help the majority group of farmers, who farm less than one hectare of land. They constitute 55% of smallholders and operate about one- third of the arable land. Improved farming would have enormous impact on the welfare of this group, but less on national production. This subject has assumed more urgency this season, when many former borrower clients are ignoring extension workers, but it would be counterproductive to discount the needs of the commercial smallholder sector in future. 6.9 The Women's Programme has made important contributions towards bringing women farmers into the credit and extension network. National figures were not obtained, but Mzuzu ADD is a good example replicated in most other ADDs; where the number of women's extension groups bringing extension to women not previously in contact with extension workers increased from 175 to 530 with membership increasing from 2,200 to 13,400, and the number of women who received seasonal credit rose from 3,500 to 14,500 an increase of 11,000 borrowers compared to the increase of about 8,500 male borrowers. 6.10 As indicated (para 5.17) improving ACB's publishing capabilities and capacity was an important project success and benefit. By contrast, the success and benefit of the Pilot Communication Program (PCP) is. less apparent. It may be that support for this programme ceased too early, before it had firmly taken root, and follow-up should be considered. 6.11 Although establishing the Food Security and Nutrition Unit (paras 5.26 and 5.27) was not planned at appraisal, the unit has developed into an indispensable food security and nutrition monitoring agency. In 1987, its work helped to avert a serious food shortage. Since then, a food security strategy, including guidance of strategic grain reserves operations, has been developed and food security and nutrition monitoring systems have been established. It is expected that further TA will be needed to maintain effectiveness. Economic Re-evaluation 6.12 At appraisal an economic rate of return (ERR) was calculated for the directly productive components of the project, with incremental production in the Mzuzu extension area of maize, groundnuts, tobacco, cotton, wheat and rice being the major quantifiable economic benefits. The ERR was estimated at 22%. In trying to reassess the economic viability of the project it is important to consider some relevant aspects. While average yields of all crops grown by smallholders have fluctuated or remained fairly constant over the 1980s and early 1990s, those of maize increased from about 1 ton/ha in 1986/87 to about 1.5 ton/ha in 1992/93. However, this increase is not part of a sustained trend in view of the fact that total cereal production in Malawi in 1993/94 is estimated .t some 900,000 tons, only 40% of the record harvest of last year. The decline reflects lower plantdr area (13% less) and poor yields, as well as the collapse of the rural credit system. The fact remains that whether in a good or bad year and whether fertilized or unfertilized, hybrid maize varieties always out-yield local varieties. This has been demonstrated through a series of experiments conducted by the Maize Agronomy Team, FAO, MOA and the Maize Commodity Team, and indicated that hybrid maize grown without fertilizer performed significantly better than unfertilized 13  local maize in 3 out of 4 years. It was also demonstrated that local maize which received 40 kg/ha N and 10 kg/ha P205 did not yield significantly more than unfertilized hybrid maize. The above clearly indicates that hybrid maize is more efficient than local maize in utilizing and partitioning nutrients to the grain, and therefore production of hybrid maize should be encouraged. To the extent farmers have resources to purchase fertilizers, the hybrid/fertilizer package (with appropriate site-specific recommendations and application practices, will provide attractive financial and economic returns. 6.13 At the national level, the increase in average maize yields in 1992/93 (an exceptional year), mentioned above, is the result of an increase in the proportion of hybrid maize planted from 3% of total maize area in 1986/87 to nearly 25% in 1992/93 (Part 111, Annex 1, Table 8), as well as increased use of inorganic fertilizer throughout the 1980s and early 1990s, and the one-off extra availability of nitrogen in the soil as a result of the drought the previous year. The first two factors are the results of: (i) technological changes; (ii) increased supplies of fertilizers and credit to the smallholder subsector; (iii) Government extension efforts, and (iv) changes in smallholder produce pricing (i.e. increase in maize prices) and marketing policies (i.e. liberalization of markets). 6.14 It is also important to recognize that the average yields of hybrid maize have been relatively stagnant e.g., averaging about 2.5 tons/ha, well below attainable yields of about 4 tons/ha under smallholder conditions. This yield gap is partly due to deficient fertilizer recommendations and farmer crop husbandry practices. However, the increased use of hybrid maize and fertilizer is encouraging, and if smallholders are able to obtain access to world market prices for other crops and to seed, fertilizer and credit supplies, the increased use of hybrid maize should lead to enhanced household food security and a diversification of production in this sub-sector. This is supported by the fact that since the liberalization of agricultural produce marketing in 1987, there has been an increase in the production of a variety of high value crops,, including tobacco, pulses, cashewnuts, sunflower seeds and chilies (ASM-1994). Bearing in mind the above considerations, the economic re-evaluation of the project, relative to Mzuzu ADD only, has been carried out on the basis of an indicative breakeven analysis. The analysis, the details of which are presented in Part III, Annex 1, Tables 7 to 12, shows that in order to generate an ERR of 13%, which would define the project as economically viable, the hybrid maize area, presently (average 1990/91-1992/93) representing about 20% of the total maize planted area in Mzuzu, should increase to about 40% in 15 years from now. This assumption is considered conservative in view of the fact that in 1984/85-1985/86, the without-project situation, hybrid maize already accounted for about 14% of the total maize area in Mzuzu. A similar approach taking into consideration not only Mzuzu ADD, but the country as a whole, would have produced a much higher ERR and well above the opportunity cost of capital. Economic viability would be improved if the extension system is able to provide better site-specific fertilizer recommendations. Under the above possible scenario, it makes economic sense for GOM to invest in sustaining the extension system. Further improvements in the cost-effectiveness and ensuring dissemination of relevant extension messages to a broader scope of smallholders would further enhance sustainable economic benefits. Some considerations in trying to estimate other possible benefits to be attributed to the project are given in paras 6.1 and 6.2. 7. Project Sustainability 7.1 The project, in line with its objectives, tried to raise the effectiveness of the extension service. However, with operating fund shortages that have handicapped the extension service during the project and with the revenue budget cuts worse, sustainability of the project's activities is in doubt without greater and sustained Government commitment and donor support (during the 1990s) for operating costs, including vehicle and equipment replacement. Cost economies, in addition to improvements to be gained from better planning and budgeting, must therefore be seriously and 14  constantly considered and implemented, including detailed review of existing staffing and services, benefits from them and whether they are all needed. For instance, the ADD structure seems excessive and should be reviewed to ascertain the value and contribution that all the staff and sections there are making, and whether the vehicle fleet noticeable at some ADD offices is all needed. In addition, the value of the large SMS complement, particularly at ADD and RDP levels is questionable. Most of the time SMSs are office-bound through lack of travel funds. With most farmers practicing a simple rotation, the well qualified RDP and EPA managers should have the technical capability to deal with technical problems and train the staff (particularly since many SMS are no more specialized or qualified than the managers), so that all but a very few SMS may be redundant. Equally, some EPAs have as much specialist FA level staff, usually serving few beneficiaries, as the section staff. A reduction or virtual elimination of these, by attrition, early retirement or transfer to FA vacancies, may generate considerable economies without affecting extension quality. Another issue, mainly concerning revenue expenditures, is the high allowances paid to trainees. This factor led to curtailment of project training for FAs and is reportedly straining training resources of ASP. These costs rule out sizeable training programs on revenue budgets, and need to be reviewed. 7.2 Without really stringent economies, the extension system remains vulnerable to the vagaries of donor assistance. Whether the service needs to be so large and costly as at present needs to be re- examined. In the meantime, the improved extension system emerging from the project, to be adequately equipped and funded by ASP, is well poised to make important contributions to Malawi's economy, with challenging tasks such as spreading the new flint hybrid maize technology widely, especially to the small farm sector; and teaching farmers the recent promising findings of involving site-specific fertilization recommendations and associated husbandry practices to enhance the economic efficiency of fertilizer use. Also, with the delinking of credit and extension, greater effort should focus on low-cost technical messages for a broader range of food and cash crops. Such assistance continues to be justified to steer effective extension to a sustainable stage, which should continue to be treated as an important objective. 8. Bank's Performance 8.1 The Bank's main strength was the close supervision of the project. Details of staff inputs by the Bank during appraisal, negotiations and supervision, and data on supervision missions are provided in Part III, Table 8. Staff time on supervision totalled 421 weeks. The project was supervised 15 times, mostly as one of a group of inter-related agricultural projects supported by IDA, and benefiting from the interaction between several of these. The Bank provided considerable implementation assistance through the visiting extension specialist, and this was reinforced from 1989, by a resident extension specialist. This appointment enhanced even further the key role of the Bank's Field Office in trouble shooting and assisting all project implementation aspects. The valuable assistance to the project of Bank disbursement, computer, procurement and financial specialists must be especially stressed. In addition, the Bank played a particularly active role in supporting the hybrid maize program. 8.2 Even though the Bank showed flexibility in amending the DCA to alleviate operating fund constraints, and agreed to extend the project by a total of 22 months to continue to support the extension improvement programme, the change in supervision emphasis in 1987 from planning and budgeting aspects to replication of the extension system to other ADDs resulted in the poor implementation of planning and budgeting aspects. In addition, there was lack of specificity on the part of the appraisal team. For example, programming and budgeting was to eliminate waste in the extension service, but there was no guideline on what waste the appraisal mission had identified. 15  Similarly, the project was to implement a "Mass Communication' component, but there is no hint as to what project designers had in mind. 9. Borrower's Performance 9.1 The Borrower showed committment to the ideals, objectives and substantive goals of the institutional development of an extension system, and GOM's committment to a better performing extension system is reflected by the willingness to extend the Mzuzu ADD pilot extension scheme nationwide. Performance was enthusiastic in the field, but there were shortcomings in project direction and services, summarized as follows: (i) lack of project management and coordination, exacerbated by several changes of MOA senior management staff, in addition the Project Steering Committee met rarely and only after repeated supervision mission reminders (para 5.31); (ii) weak procurement and financial services (paras 5.29 and 5.30); and (iii) only "lukewarm" support of field activities from headquarters except for the last two years of the project (para 5.32). 9.2 Compliance with DCA covenants is shown in Part III, Table 7. In general, covenants were fulfilled, albeit late in the case of time-bound obligations. Two covenants, requiring credit studies and a modified lending structure, became unnecessary with implementation of the Smallholder Agricultural Credit Project. There were two covenants that GOM was unable to fulfil the establishment of a career stream and the provision of incremental operating expenses. Of these, the Bank waived the requirement to establish a career stream for extension staff (para 5.18). The Government failed to meet 100% of incremental operating costs, and its stated intention to do so was overtaken by the deteriorating national economy, requiring major budget cuts in all ministries. The extension improvement pilot scheme needed higher operating expenses than foreseen at appraisal, which led to project financing of some incremental operating costs. 9.3 Project accounting and auditing was complex, because there were separate accounts for Mzuzu, PD and DOA. They were completed but invariably late. Reporting and submission of budgets and work plans varied from late to no submission at all and reflected weak project management. 10. Lessons Learned 10.1 The major lessons to be derived from the implementation of the project are the following: - a comprehensive pre-project/pre-program analysis within country of farming systems and technology needs of the main farmer types, the available stock of technology, and the technology generation capacity are needed to ensure cost-effective and sustainable interventions; an improved program budget system for recurrent cost funding should be developed and implemented within MOA, and this should be coordinated with an ongoing review of staff levels and job description; more thorough assessment of the borrower's financial and procurement capabilities is needed during project preparation/appraisal, as well as operational requirements for their strengthening; it is important to establish effective project management and coordination arrangements before project implementation; 16  field supervision of financial, technical and management aspects of projects with IDA donor assistance is extremely important; a detailed review of the various NRDP is needed to ascertain whether there are sustainable and effective alternatives to cost-intensive staff and operating requirements of the extension system; greater attention needs to be devoted to ensuring that improvements in extension methods are complemented by ensuring strong research-extension-farmer linkages; the importance of separating extension and credit activities to ensure extension workers devote their time to extension activities; and - monitoring and evaluation of project activities are essential in order to communicate the positive results of extension activities to Government and obtain Government commitment/support (particularly financial). 10.2 Other lessons include: the following problems, identified during the course of field work (para 5.15) require special attention: field work financing, transport provision and maintenance, improving field staff supervision and technical skills, and expanding extension staff clientele to the "small farm' group; the lack of discernible impact from high training expenditures is troubling and warrants a review of the benefits from the type of training financed by the project and the high cost of local training; and MOA has become the owner and custodian of very considerable assets, by way of buildings and vehicles, and care needs to be taken to ensure they are properly maintained. 10.3 The design features of the follow-up project ASP, have taken into account all of the above lessons. It is important that ASP implement effectively the various action plans prepared to address the key lessons from MAESP. Regarding improved and sustainable extension services, ASP is giving emphasis to: improved mechanisms to ensure a constant flow of improved technologies to a broader range of smallholders through greater emphasis on low-cost technical messages and more effective researcher-extension-farmer linkages; improved MOA budgeting and planning procedures and capacities, including development and application of improved "yardsticks' for determining the level of non-salary operating funds for ensuring effective extension services; and improved monitoring and evaluation of the effectiveness and impact of improved extension services. 17  11. Project Relationship 11.1 There was good Borrower/Bank collaboration, due largely to the constant dialogue between MOA and the Bank's Field Office. This was reinforced by the appointment of the Malawian Agricultural Specialist, who devoted much of his time to the project, in particular to fostering progress of extension improvements at the field level. In addition, as pointed out elsewhere, supervision missions provided considerable implementation assistance, which was effective, and appreciated by the Government. 12. Consultancy Services 12.1 The quality of the consultancy services provided by the project varied. The two USAID funded consultants who collaborated with the MAEPS project - on communications and women's programs - did excellent work. Two of the IDA financed six multi-year consultancies failed. Interestingly, short teaching courses succeeded in introducing log frame planning and budgeting systems, after the two year TA had failed to do so. On the whole, the TA for this project was needed and well-conceived, but success depended on the quality of the consultants and the adequacy of the supervision by MOA staff. 13. Project Documentation 13.1 The SAR, in general, provided a useful framework for project implementation. However, certain specific aspects were missing that would have made implementation, supervision and PCR preparation easier, for example, the omission of detailed cost tables, which are the only record of proposed investment details. On the Borrower's side, lack or irregularity of reporting would have made supervision very difficult, but for the close contact maintained by the Bank's resident staff with MOA. Lack of recent impact evaluation makes it impossible for the Borrower and the Bank to judge the merit of ongoing extension work, and this aspect of the project deserves urgent attention. 18  PROJECT COMPLETION REPORT MALAWI AGRICULTURAL EXTENSION AND PLANNING SUPPORT PROJECT (CREDIT 1626-MAI) PART II: PROJECT REVIEW FROM THE BORROWER'S PERSPECTIVE4 1. COMMENTS ON PART I AND II. Project Design 1.1 The project is consistent with Government efforts to improve analytical capability within the Ministry of Agriculture to provide for a conducive policy environment to facilitate delivery of agricultural services. It therefore focussed on provision of Technical Assistance and training of staff in both the Planning Division and the Department of Extension. 1.2 Project Management and Organization: The Management of the project rested in the Principal Secretary of the Ministry of Agriculture and coordination to be done through the Project Steering Committee. This Management Structure proved ineffective in coordinating activities of the project which were in three separate units, namely, the Planning Division under the Chief Planning Officer, the Department of Agriculture and Mzuzu ADD. However, it is not true that ADD Managers were at the same seniority level with the Chief Agricultural Officer. The CAO was a grade higher than the ADD Managers. The lack of direction in the Department of Agriculture (DOA) was largely due to staff turnover. The CAO's office saw five different officers, during the life of the project. This made follow-up on issues difficult. Implementation of Individual Components 1.3 The implementation of most project components was largely satisfactory however the project faced several major constraints which slowed or prevented the implementation of some components of the project. Although most of these were resolved during the course of implementation of the project, some still persist. There is need therefore to take cognisance of these and address them in future projects. Strengthening of the Planning Division 1.4 To strengthen the analytical capabilities of the Planning Division, the project focussed on provision of three Technical Assistants totaling 84 man months, post-graduate Master's training, vehicles and equipment. The original estimates of the training programme for short-term offshore courses was revised upward from 99 foreseen at appraisal to 585 person weeks on 19.2.1988 with the Bank's concurrence, although only 156 person weeks were achieved. The doubling of the Master's programme to 12 M.Scs was made possible because some took one year instead of the two years planned. In addition there was a transfer of resources from the short-term offshore training to the 4 Prepared by the Ministry of Agriculture and Livestock Development 19  Master's programme to meet the expressed need of the Division and the revision that was made to include one Masters for an Economist in the Food and Nutrition Unit of the Department of Economic Planning and Development. 1.5 The performance of the Technical Assistance Inputs is mixed. As observed in Part I of the PCR, while others did develop lasting systems, others either left little or no imprint. This calls for need for careful selection and supervision of TAs to ensure that they perform up to expectation. The fact that most of the TAs served up to the end of their contracts without being checked leaves one wondering as to whether they were supervised or not. Improving the National Extension System 1.6 The assessment of the extension component in Part I and III has been reviewed and the borrower has the following comments to make: (a) Part III Table 4 has been revised in line with the revision that was done on 19.2.1988. The actual delivery for in- country extension courses was 8,479 person weeks out of the planned 13,109 person weeks. (b) Section 5.20. The low coverage of field level technical staff was due to the reallocation of funds to the Training of Trainers (TOT) programme which was felt to be a prerequisite for the field one. In addition, the programme used hotels for this training. This proved very costly and was later changed to utilize Residential Training Centers (RTCs). (c) Section 5.23: The failure of the Seed Production Programme was due to poor pricing structure which did not give -any incentive to seed producers. This however does not automatically mean that smallholder farmers cannot multiply seed at competitive prices to commercial seed producers. (d) Section 5.25: It is not true that the injection of incremental credit into Mzuzu ADD stopped credit rationing. An examination of the smailholder fertilizer sales on both cash and credit shows that cash sales have been on the increase throughout the projected life except in 1989/90 and 1992/93. (Part II Table 2 and Fig. 1). (e) Section 5.28: While the design of the linkage between Research and Extension seemed ideal, experience shows that there was no full establishment of the ARTs. This was due to the requirement for training of the staff and the unforeseen staff turnover of scientists for better oportunities. In addition the ARTs did not receive adequate support from Commodity Research Teams. There is need for future programme to address this. (f) Section 5.31: It should be noted that while the Principal Secretary is held responsible for the overall responsibility of the Project, much of the weakness in project implementation was due to its poor design. The project did not have a definite Project Management Unit despite its dispersed nature. (g) Section 5.32: The CAO is a grade higher than Programme managers. (h) Section 6.5: The PCP activities did not expand outside the pilot areas due to staff turnover and lack of funds. Although all ADDs did benefit from equipment purchased under the project. 20  (i) Section 6.6: It is not entirely true that extension messages went to credit farmers only. There have been messages on low cost technologies such as use of organic manure, soil conservation and nutrition. It should also be noted that the SAR does not make definite recommendations for reaching the poorest smallholders. There is therefore need for future programmes to be more explicit as to how these can be reached. (j) Section 6.8: National figures for women credit groups increased from 34,273 to 104,598 while Women IGA groups increased six-fold from 157 to 1,032. (k) Section 7 - Project Sustainability: Whilst it is recognized that the present extension system is resource intensive and expensive, there does not seem to be any viable alternatives. The suggestion to retrench some staff at the ADD level in favour of lower level staff may not reduce costs as there will be need to strengthen this level and may compromise quality. The high costs of training programmes are not due to high allowances paid to trainees. Trainers are given normal government rates which are the subsistence allowances paid to officers. Cost cutting measures have been adopted including holding of courses at Residential Training Centres (RTCs). 2. ADEQUACY AND ACCURACY OF FACTUAL INFORMATION IN PART I 2.1 The country data used in the definition of the project profile for the PCR was in a large part provided by the participating institutions including the Planning Division, the Agricultural Training Unit, Departmnent of Extension & Training and the Chief Accountant. This has been reviewed and a few amendments made. The following are the major corrections that were made: (i) Table 4 page 3 of Part HI: The Appraisal Estimates were revised as indicated in Table 1 below. Consequently the PCR estimates ought to be assessed in light of the revised targets. 3. EVALUATION OF BANK'S PERFORMANCE AND LESSONS LEARNED 3.1 Bank personnel played a significant role in both the appraisal and execution of the project. The Supervision Missions during the implementation phase were judged to be professionally qualified and did assist in guiding the project, including effecting the amendments made to the DCA and the Extension System. However, it is surprising to note that they were unable to resolve the management problems which were recognized early at the beginning of the project. For example, the lack of a definite project manager, despite early recognition of the need for one; nothing was suggested by the mission to correct the situation by either suggesting the positioning of a project manager or facilitator or setting up a Project Management Unit (PMU). 3.2 The Bank's insistence on forcing the borrower to accept its recommendations without careful consideration of the implications and justification did contribute adversely to the implementation of the project. The instance on having the borrower to accept modification of standard housing designs in the interest of cost cutting without considering the effects and availability of funds from the totality of the project is one example which resulted in building of inappropriate house designs which will require further financial resources to correct. 3.3 The introduction of the Training and Visit (T&V) Extension System and the associated OFDs to the rest of the ADDs was done without due appraisal of the availability of resources to implement 21  the system although this was later recognized and reallocation of SDR 1.0 million was made to the operation costs category. Even this proved very insignificant to effectively assist eight ADDs which had serious shortages of operating funds. 3.4 These shortcomings have been recognized and attention has been given to provision of operating funds, transport and a project facilitator in the ASP. 3.5 The performance of other co-financiers was satisfactory as can be seen from the success made in the Women IGA Component and the External Training Programme. 4. EVALUATION OF THE BORROWER'S PERFORMANCE AND LESSONS LEARNED 4.1 Overall performance of the borrower was satisfactory given the environment within which the project was implemented. Most of the actions indicated in the project action plan were implemented. These included the establishment of the Project Steering Committee, the recruitment of the Technical Assistance. 4.2 There were however, a number of implementation weaknesses as evidenced by the ineffectiveness of the Project Steering Committee and failure of Government to comply with the one important covenant to provide adequate financial resources to cover incremental recurrent operating costs. This had an important bearing on the ability of field staff to implement the extension system. However, with hindsight it would be appreciated as indicated by the Bank evaluation report that this commitment was frustrated by the deteriorating situation of the national fiscal budget. 4.3 Frequent staff turnover of senior staff both in the Department of Agriculture (DoA) and Planning Division mitigated against follow-up and continuity of crucial project activities for example, the supervision and control of Technical Assistant staff. The Project saw five different Chief Agricultural Officers (CAOs) heading the Department of Extension while the Planning Division saw three changes of the Chief Planning Officer. In the absence of elaborate handover notes, follow-up of issues was seriously compromised. Compliance with project progress reporting requirements and preparation of annual workplans was partial. This was largely due to inadequate staffing of the Central Monitoring and Evaluation Unit which was to coordinate the activity. This weakness has been addressed in the preparation of the follow-up project (ASP). 5. MAJOR LESSONS LEARNED 5.1 The major lessons that were learnt from the experiences with the Project are: (a) The need to ensure an effective management unit. It should be noted that the office of the Principle Secretary cannot be made a coordinating unit for a particular project. As has been pointed out in Part I, this has been taken into account by the provision of the post of a Project Facilitator in case of the ASP and the post of Programme Manager for the GOMIUNDP Fifth Country Programme. (b) The need for a thorough mid-term project review would avoid focusing on a few selected components to the exclusion of other equally important components of the project. Experience with the project clearly shows that the Bank supervision missions remained partial in the treatment of the management problems. A thorough Mid-term review would give a more balanced review of the project implementation. 22  (c) The need for the borrower to keep staff transfers to the minimum. Although some staff movements may be inescapable it would be advisable for Government to keep staff movements to a minimum particularly for those working in key areas of projects. (d) The insistence on forcing new introductions as evidenced by the forced modification of staff houses in Mzuzu and the wholesale expansion of the T&V Extension system with its associated OFD underscores the need for donors and implementers to appreciate the existing or client specifications and availability of local resources. Future projects should ensure that any modifications to standard designs in buildings and systems are carefully evaluated before adoption. (e) The need to restore credit discipline. The relaxation of the 100% repayment condition for eligibility for credit, coupled with wholesale moratorium extended to all farmers during the 1991/92 drought season led to the credit crisis which affected the whole country including Mzuzu ADD. Given the important role credit plays in the uptake of improved inputs and production, it is important that such moratorium be selective. A study on the Credit Crisis was carried out on a national scale to determine the causes of the credit crisis and made recommendations to ensure the restoration of the smallholder farmer credit system. It is important that these recommendations be adopted if agricultural productivity is not to be compromised. 23  Table 1 (a) EXTENSION APPRAISAL REVISED ACTUAL PESTMAIE ESTMAIM (PCR ESTIMATE) Ph.D 2 1 MSc 16 24 23 Diploma 34 34 20 TOTAL -PhD 2 1 - Diploma 34 34 20 (b) SHORT TERM (OFF SHQRE1 - Planning Division 99 585 156 - Extension 122 103 808 TOTAL 221 688 964 (c) SHORT-TERM IN-COLNThY __ - Training of Trainers 1,328 1,328 p. weeks 2,380 p.weeks - Technical and others 13,109 13,109 6,009 24  Table 2: Smallholder fertilizer cash and credit sales, Mzuzu ADD. 1984/85 - 1993/94 seasons (tons of fertilizer) Season Cash Credit Total 1984/85 6615 5781 12396 1985/86 5530 6474 12008 1986/87 6639 4290 10929 1987/88 6914 5047 11961 1988/89 7298 5398 12696 1989/90 4570 10745 15315 1990/91 6195 8126 14321 1991/92 8859 5654 14513 1992/93 7624 10461 18085 1993/94. 9411 . 3622 13033 25  Figure 1: Cash and Credit Sales Mzuzu ADD (Tons of Fertilizer) 15 10 0L l . 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 -C CASH CREDIT W TOTALS 26  PART III: STATISTICAL INFORMATION Table 1. Related Bank Loans and/or Credits Loan/Credit Title Purpose Calendar Year of Status Comments Approval (1) National Rural The project was designed to November 1978 Closed The project included co- Development assist GOM to implement the June 30, 1986. financing from CIDA, EDF, Programme first five-year phase of the PCR dated UK and Fed. Rep. of NRDP Phase I programme. Completion of June 1987. Germany. (Credit 857-MAI) Lilongwe Land Development Programme (LLDP). (2) National Rural Very similar components to October 1981 Closing date Third IDA project to support Development NRDP IV. Main objectives to September 1986 rural development in Programme increase agricultural production, extended to Karonga Agricultural NRDP Phase III5 income and welfare of the September 1988. Division (ADD) in North (Credit 1183-MAI) smallholder sector. It also PCR dated Malawi. included forestry. May 1990. (3) National Rural The project included components April 1983 Closed Fourth IDA project to Development to strengthen and support April 30, 1991. support rural development in Programme extension services, training and PCR dated Lilongwe ADD. Extension NRDP Phase IV adaptive research and made April 1991. services satisfactory and an (Credit 1343-MAI) provision for improved credit effective mechanism for the and marketing facilities. issue and recovery of seasonal credit was established. (4) National Provide improvements in the February 1985 Closed In terms of its general Agricultural infrastructure and planning and October 31, 1993. objectives, the project has Research Project management procedures of DAR PCR dated been successful, particularly (Credit 1549-MAI) and research-extension linkages. February 1994. on its breeding programme to produce flint type grain of high yield maize. Linked with USAID financed lMARE. (5) Smallholder To transform Malawi's 1988 Under IFAD co-financier. Agricultural Credit smallholder credit system into a implementation. Project better managed, more efficient, Closing date (Credit 1851-MAI) and financially viable national June 30, 1995. credit delivery system. (6) Agricultural Designed to help formulate and 1993 Under Total project cost: US$ 79.2 Services Project implement Institutional and implementation. million. Co-financing (Credit 2514-MAI) Management reforms in Launched in provided by IFAD and Research, Extension and Input December 1993. African Development Bank. Supply Systems. NRDP II was a Wood Energy Project. 27  Table 2. Credit Timetable Item Date Planned Revised Date Actual Date - Identification6 January 1983 - January 1983 - Preparation' June 1983 - June 1983 - Appraisal' September/October 1984 - September/October 1984 - Credit Negotiations9 July 25-30, 1985 - July 25-30, 1985 - Board Approval September 19, 1985 - September 19, 1985 - Credit Agreement December 20, 1985 - December 20, 1985 - Credit Effectiveness February 1, 1986 March 20, 1986 March 20, 1986 - Credit Closing September 1, 1991 February 28, 1993 June 30, 1993 6 Carried out by IDA. 7 MOA/APD was responsible for preparing the project proposal. Bank staff assisted the GOM in the preparation work and periodically reviewed progress. s The project proposals were submitted to IDA and USAID in June 1984 for co-financing. The proposals were reviewed and discussed with GOMIMOA by an IDA mission which visited Malawi in June/July 1984. 9 Held in Lilongwe, Malawi at Ministry of Finance. 28  Table 3. Credit Disbursements Cumulative Estimated and Actual Disbursements (IDA) IDA Fiscal Year F1985/86 1986/87 19M/88 | 1988/89 |-1989/90 | 1990/91 1991/92 1992/93 1993/94 ...................................... million)..................................... A. In US$ Appraisal Estimates 0.7 3.3 7.2 9.3 10.8 11.6 11.6 11.6 11.6 Actual 0.5 0.7 1.8 4.0 6.9 9.2 11.8 13.5 15.0 Actual as % of Estimates 71 21 25 43 64 79 102 116 129 ..................................... .(SDR' million) . w B. In SDR Appraisal Estimates'0 0.7 3.3 7.3 9.4 10.9 11.7 11.7 11.7 11.7 Actual 0.5 0.6 1.4 3.1 5.3 7.0 8.9 10.1 11.2 Actual as % of Estimates 71 18 19 33 49 60 76 86 96 Date of Final Disbursement: November 16, 1993: MK 67,650 (SDR 10,958). The undisbursed balance of SDR 509,750 was cancelled as follows: SDR 500,000 as of March 19, 1993 and the remainder as of November 16, 1993. 10 Mission's Estimates: Figures in SDR were not given in SAR.  Table 4. Project Implementaton (Page 1) Indicators Appraisal Estimates Actual (or PCR Estimates) A. Civil Works For Mzuzu ADD: . Houses 118 136 . Offices 2 2" Marketing Sheds 16 17 Dip Tanks 5 5 B. Vehicles Planning Division 3 4 Department of Agriculture 14 14 Mzuzu ADD 21 28 Communications/Adaptive Res."2 _11 _ Total 49 57 C. Other Transport Mzuzu ADD: | Motorcycles 99 107'3 | Bicycles 25 100 Other ADDs: l Motorcycles - 224"' . Bicycles 200 Total: . Motorcycles 99 331 l Bicycles 25 300 Also several large consignments of motorcycle spares D. Equipment Office Equipment For PD Believe more than projected DOA was supplied Mzuzu Extension Equipment Mzuzu Supplied Audio Visual Equipmentv For ACB Supplied Computers 2 20 With Software l There was additional construction for renovating and expanding residential training centers. 12 USAID financed. 3 Of these, 26 had been paid and reportedly skipped but had not yet arrived in Mzuzu. 14 Paid and reportedly skipped but not yet arrived in Malawi. 30  Table 4. Project Implementation (Page 2) Indicators Appraisal Estimates Actual (or PCR Estimates) E. Technical Assistance (Long-termn) (i) Planning Division Planning Management Advisor 36 months Financed by EEC Senior Systems Analyst 24 months 30 months Pricing Policy Specialist 24 months 30 months M&E Specialist 30 months Sub-total 84 months 90 months (ii) Department of Agriculture Extension Planning Advisor 30 months 36 months Financial Advisor 30 months 30 months Seeds Specialist 36 months Training Advisor"5 60 months 60 months'6 Training Specialist" 60 months 48 months" Mass-Media Specialist" 30 months 34 months Women's Programs Advisor" 30 months 36 months Sub-total 276 months 244 months'" Total 360 months 334 months F. Technical Assistance (Short-term) Planning Division 12 n.a. Training (Extension) 35 n.a. Ag. Communications (Extension) X Believed mostly used Total 83'5 n.a. G. Training I. Long-term (i) Planning Division MSc 6 12 I5 USAID financed. 16 Mission Estimate. 17 Of which 178 months financed by USAID. 18 Of which 74 for extension. 31  Table 4. Project Implementation (Page 3) Indicators Appraisal Estimates Actual (or PCR Estimates) (ii) Extension PhD MSc 1019 23 Diploma 34 Total: . PhD - 1 . MSc 16 3520 . Diploma 34 20 H. Short-term (Off-shore) (i) Planning Division 100 person-weeks 156 person-weeks (ii) Extension 120 person-weeks2' 808 person-weeks22 Total 220 person-weeks 964 person-weeks4' M. Short-term (In-country) Training of Trainers 1,328 person-weeks 2,380 person-weeks Technical and Other Courses 11,800 person-weeks' 6,099 person-weeks24. Period during which training took place was 143 weeks 19 SAR figures of which 6 to be financed by USAID, however training unit data shows a revised plan for 24 MScs with 16 for USAID finance. 20 Includes 13 financed by USAID. 21 SAR projection of which 72 weeks were to be financed by USAID training unit records show revised plan of 210 weeks half for USAID finance. 22 334 person-weeks financed by USAID. 23 Of which 9,900 to be funded by USAID. Total includes 7,800 person-weeks training for FAs and FHAs. 24 Mission estimate. 32  Table 4. Project Implementation (Page 4) Activity Expected Implementation' Actual Implementation H. Key Implementation Indicators I. Ministry of Agriculture (MOA) (i) Establishment of the MOA By March 1, 1986 Committee established, but met Programme Steering Committee infrequency and irregularly. (ii) Establishment of Project Special By credit effectiveness Implemented. Account II. Plannino Division (PD) (i) Recruitment of Planning and By April 30, 1986 Financed by EEG. Management Advisor (ii) Employment of Systems Analyst By March 31, 1986 On September 8, 1987. (iii) Review of PD's proposed By September 30, 1986 Done, but date not recorded. functions and reorganization- completed (iv) Staff Recruitment Programnme By September 30, 1986 Done, but recruitment is EP&D developed and submitted to IDA responsibility, and due to staff shortage for comment there remained vacancies, somewhat (v) Recruitment of additional senior By March 31, 1987 reduced in later project years. staff-completed (vi) Formulation of Five-Year Plan- By July 1. 1987 1st draft October 1988 never Completed and sent to IDA for satisfactorily completed. comment (vii) Review of Planning and By February 28, 1988 Not applicable. Management Advisor's contract for purpose of possible extension (viii) Annual Work Plans and By September 30 of each year Submitted spasmodically, sometimes Budgets Due incomplete, usually late. HII. National Extension System (DOA) Reorganization (i) DOA - Internal reorganization By March 1, 1986 Not clear from SAR what was merit. If completed appointments to DOA HQ, a Deputy CAO was appointed, but no economist, due to staff shortages. (ii) Recruitment of Extension By March 1, 1986 On August 28, 1987. Management and Planning Advisor 25 Expected Project Effectiveness February 1986. Actual Project Effectiveness March 1986. 33  Table 4. Project Implementation (Page 5) Activity Expected Implementation Actual Implementation Management and Systems Improvement (i) Recruitment of Financial By July 1, 1986 In February 1987. Advisor/Controller (ii) Establishment of DCAO post By March 1, 1986 Done, but date not recorded. and appointment (iii) Design and Implementation of Long-term TA for this purpose was ineffective Extension Planning and but short-term training brought results and log Budgeting Systems frame system adopted in 1992/93 by most - Mzuzu ADD By April 30, 1987 ADDs and DOA units. - Other ADDs Thereafter (iv) Implementation of Mass By October 31, 1986 Pilot communication project to design and Communication Pilot Scheme produce advisory leaflets introduced March 1989. (v) Annual Workshops May/June of each year Implementation not recorded, but believed that they took place. (vi) Annual Extension Work Plans By September 30 of each year Unsatisfactory most of the time, but improved for DOA and Mzuzu ADD in 1992/93. (vii) Five-Year National Extension By July 1, 1987 Ist version published January 1990. Revised Strategy Plan Completed and incorporating Bank's suggestions, July 1992. sent to IDA for comments Strengthen DOA's Technical Capacity (i) Recruitment of USAID financed By September 30, 1985 October 1986. Training Advisor (ii) Training Advisor's responsi- By September 30, 1985 October 1986. bility to be broadened to include all professional and technical training within MOA (iii) Establishment and staffing of By credit effectiveness Implemented. the Training Unit (iv) Recruitment of Agricultural By July 1, 1986 On October 10, 1986. Communications Specialist (financed by USAID) (v) Proposed establishment of March 31, 1987 MOA attempted to implement this requirement, Career Stream and Promotion but it was turned down by the Government on System for DOA the grounds that a new career stream cannot be introduced for one department or ministry in isolation from the rest of the public service. Rural Credit Study (i) Formulation of TOR By September 30, 1985 Studies were carried out as part of project (ii) Study completed and sent to By March 31, 1986 preparation of the Smallholders Agricultural IDA for comment Credit Project, which took over the obligations for credit system improvements from MAEPS. (iii) Implementation of agreed By January, 1987 recommendations 34  Table 5. Project Costs and Financing A. Project Costs | SAR Estimates26 Revised" Actual28 Actual as a % I I I ~~~~~~~~~~~~~of SAR Estimates ................. (US$000) . A. IDA Financing 1. Civil Works 1.8 1.9 2.2 122 2. Vehicles, Equipment and Furniture 1.0 1.2 2.3 230 3. Consultants' Services 2.8 2.9 4.9 175 4. Training 2.9 3.3 2.7 93 5. Credit 3.1 3.6 1.9 61 6. Incremental Operating Costs - 1.3 1.0 ... 7. Unallocated - 1.0 ... u' Total IDA Financing 11.6 15.2 15.0 129 B. USAID Financing2e 6.2 6.2 5.6 90 C. Government Contribution 2.4 2.4 2.4 100 Total Project Cost 20.2 23.8 23.0 114 26 As given in SAR; IDA Financing as per DCA amounted to SDR 11.7 million equivalent as indicated in part B below. 27 Mission estimates. For IDA financing. figures are derived from the amended DCA expressed in SDR and converted into USS using the exchange rate of I SDR = 1.3 (Average 1989). 25 IDA financing derived from actual disbursement equivalent to SDR 11.2 million (US$ 15.0 million) as indicated in part B below. 29 Mission estimates. USAID contribution amounted to USS 14.8 million for both NARP and MAEPS. Separate contributions for each project were not available and it is assumed that USS 9.2 million and USS 5.6 million were disbursed for NARP and MAEPS respectively in line with SAR estimates.  Table 5. Project Costs and Financing B. IDA Financing Planned Revised Actual Actual as a % Categories Value T % Value % Value f %Pn (SDR'000) (SDR'000) (SDR'000) 1. Civil Works 1.5 12.8 1.5 12.8 1.7 15.2 2. Vehicles, Equipment and Furniture 0.9 7.7 0.9 7.7 1.7 15.2 3. Consultants' Services 2.2 18.8 2.2 18.8 2.1 18.8 4. Training 2.5 21.4 2.5 21.4 3.6 32.1 5. Provision for Incremental Credit 2.8 23.9 2.8 23.9 1.4 12.5 o 6. Operating Costs - - 1.0 8.6 0.7 6.3 7. Fund - - - ... ... 8. Unallocated 1.8 15.4 0.8 6.8 - - Total IDA Financing 11.7 100.0 11.7 100.0 11.23