Docunent of The World Bank FOR OFFICIAL USE ONLY Report No: 24291 PROJECT PERFORMANCE ASSESSMENT REPORT SOCIALIST REPUBLIC OF VIETNAM AGRICULTURAL REHABILITATION PROJECT (CREDIT 2561-VN) June 11, 2002 &ctor and hematic Evaluation Group Operations Evaluation Deartment Currency Equivalents (amul averages) Currency Name: Vietnamese Dong, Currency Unit: VND 1994 US$1.00 VNDI0,966 1997 US$1.00 VNDI1,683 1995 US$1.00 VNDll,038 1998 US$1.00 VND13,268 1996 US$1.00 VND11,033 1999 US$1.00 VND13,943 Abbreviations and Acronyms ARP Agricultural Rehabilitation Project CAE Country Assistance Evaluation CIMP Community integrated pest management CMP Computerized Project Management EA Earning Assets FAO FFS Farmer Field school GDP GRC General Rubber Corporation LAS International Accounting Standards ICR Implementation Completion Report HDA International Development Association IPM Integrated pest management MAFI Ministry of Agricultwe and Food industries MAR} Ministry of Agriculture and Rural Development MOF Ministry of Finance OED Operations Evaluation Deatment PCU Project Coordination Unit PPAR Project Performance Assessment Report mTI program of targeted interventions SAR Staff Appraisal Report SOEs State-owned enterprises SRC State Rubber Company TA VAS Vietnam Accounting Standards VBARD Vietnam Bank for Agriculture and Rural Development Fiscal Year Govemment: Januauy 1 to December 31 Director-Gnral, Operations Evahlation : Mr. Robert Picciotto Director, Operations Evahlation Departmnt: : Mr. Gregory Inp8 Manager, Sector and Thematic Evaluation : Mr. Alain Barbu Task Manager : Mr. Ridley Nelson The World Bank Washington, D.C. 20433 U.S.A. Office of the Director-General Operatons Evaluation June 11, 2002 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Assessment Report on Vietnam Agricultural Rehabilitation Project (Cr. 2561-VN) Attached is the Project Performance Assessment Report (PPAR) for the Vietnam Agricultural Rehabilitation Project for which a credit in the amount of SDR 69.4 million (US$96 million equivalent) was approved on January 25, 1994. The project closed on June 30, 1999, 18 months behind schedule. The final total disbursed was US$107.5 million equivalent, 100% of the original amount. There was no cofinanc;ng. The main objective of the project, classified as a program of targeted interventions (PTI), was to support government efforts to increase rural incomes and to assist Vietnam's transition to a market economy. This was to be done through assistance to key agriculture institutions serving the newly emerged smallholder sector and through support for the estate rubber sub-sector. The project, the first World Bank-funded project in the sector since 1978, was expected to assist the ongoing adjustment from collectivized agriculture to smallholder family farmiing through four main components: (i) rehabilitating agricultural support services by providing assistance to decentralized extension in 23 out of the 61 provinces, improving plant protection services including integrated pest management, and strengthening research in three key institutes (16% of project costs); (ii) providing supervised credit to about 50,000 private smallholder farm households in 10 provinces to purchase fertilizer, improved seeds, and other inputs and on-farm development works (51% of project costs); (iii) rehabilitating about 160,000 hectares of established rubber of high potential productivity, providing additional processing capacity, and upgrading the quality of rubber being produced through construction of five new latex processing facilities and rehabilitating four old facilities (34% of project costs); (iv) providing training and technical assistance for policy analysis and preparation of follow-on projects to give more specialized assistance to the sector and to develop a smallholder rubber program (about 10% of project costs but included within the above components) The project was generally well-designed and well implemented, although as a Poverty Targeted Intervention there was limited explicit poverty design rationale. It would have benefited from a logical framework matrix built around the poverty objective. OED rates outcome as satisfactory, sustainability likely, institutional development as modest, Bank performance as satisfactory, and Borrower performance as satisfactory. The findings are consistent with the recent OED Country Assistance Evaluation (CAE) that the performance in the rural sector was satisfactory. The main objective of the project-to support govemment efforts to increase rural incomes and to assist Vietnam's transition to a market economy through support for key agnculture institutions-was largely met, although attributable impact data on incomes is not available except, to some extent, from FAO data in the integrated pest management component. The rural finance component more than achieved its targets at 95% repayment rates while strengthening the institution and giving the institution its first experience of rural lending to small farmers on a substantial scale. Rubber rehabilitation achieved the improved estate productivity planned and largely achieved the improved processing planned although at higher costs than anticipated. The livestock breeding component performed well. Substantial skills development in the sector was achieved. The OED assessment focused particularly on one sub-component of the project, the Integrated Pest Management sub-component, because, although it was only 3% of the project costs, it is the first time that OED has looked at impact from this new and expanding area of Bank mrual lending. In this component, efficiency and fiscal sustainability are questionable. The Farmer Field School (FFS) approach to extension and integrated pest management has so far reached only 4 percent of Vietnam's farmers over about an eight-year penod. Public expenditure would have to increase dramatically to achieve substantial FFS coverage. Moreover, the impact data available raise some questions about the extent to which FFSs have really achieved incremental benefits and about the extent to which FFS farmers have sustained some of the 1PM practices in later years. However, with further adaptation there may be the germs of a promising extension system. Experinentation at lower cost levels is needed as well as further impact studies including better understanding of the level of farmer to farmer transmission and the extent of beneficial environmental externalities. The project findings tempered by other OED expenence offer five main lessons. First, that developing a poverty-related logical framework helps to clarify the poverty rationale. Second, that as decentralization progresses mechanisms to ensure sustainability become more a function of reforms that deliver self-sustaining institutions and cost recovery at the local level than public expenditure commitment from the center. Third, that alongside other OED evidence, this project raises efficiency and sustainability questions about the use of separate PCUs - which proliferate partly to meet donor procedurl demands. Fourth, key elements in scaling up extension include maximizing farmer-to- farmer diffusion, enhancing the capacity of farmers to develop their own solutions, cost recovery or privatization, minimizing the costs of trainers through the use of farmer-trainers, and lower cost mass media approaches. Pilots of alternative models to test the impact of different extension intensities may be important for fornulating extension strategies. Fifth, Projects with a number of components and PCUs, and complex links to provincial and district management units, and with substantial challenges in procurement, could benefit from the use of Computerized Project Management (CPM) systems to track and manage expenditure. Attachment i OED Mission: Ening development effectiveness through excellence atd Independence In evaluation. About this Report The Operatios EvakUtO Depatmnent assesses the progaras and actites of Fe World Bank for two purposes: first to ensure the irnegrity of the Bank's self-evaluation pmcess and to verify that the Bank's work is producirg the expeded results, and second, to help develop mproved directns, policies, and pmcedures trugh the dissemination of lessor. drawn ftrm experience. As part of this work, OED annually assesses about 25 percent of the Bank's lending operations. In selectirg operatorns for assessment pference is given to those that are innovative, large, or oomplex. those that are relevant to upcoming studies or counbry evaluations; those for which Executive Directors or Bank management have requested assessents; and those that are likely to gnerate important lessons. The projects, topics, and analytical appoaches selcted for assessment support arer evaluation studies. A Project Perfonnance Assessment Report (PPAR) Is based on a review of the Implementation Completion Report (a self-evaluation by the responsible Bank department) and fieldwork conducted by OED. To prepare PPARs. OED staff examine project files and other documents, Interview operational staff, and in most cases visit the borrowing country for onsite discussions with project staff and beneficaries. The PPAR thereby seeks to validate and augmrent the information provided in the ICR, as well as examine issues of special interest to broader OED studies. Each PPAR is subject to a peer review process and OED management approval. Once cleared internally, the PPAR Is reviewed by the responsible Bank department and amrended as necessary. The completed PPAR is then sent to the borrower for review; the bonowers' omments are attached to the document that is sent to the Bank's Board of Executive Directors. After an assessment report has been sent to the Board, it is disclosed to the public. About the OED Raing System The time-tested evaluation methods used by OED are suited to the broad range of the World Bank's work. The methods offer both rigor and a necessary level of flexibility to adapt to lending instrument, projeci design, or sdcoral approach. OED evaluators all apply the samne basic method to anrve at their project ratings. Following is the definition and rating scale used for each evaluation criterion (more information Is available on the OED website: htIp:Worldbank.orgloeeta-mainpage.htrl). ReIevance of 04ecit The extent to which the projects objectves are consistent with the country's current development prioritles and with current Bank country and sectoral assistance strategies and corporate goals (expressed in Poverty Reduction Strategy Papers, Country Assistance Strategies, Sector Strategy Papers, Operational Policies). Possible rtings: High, Substantial, Modest, Negligible. Efficacy The extent to which the project's objectives were achieved, or expected to be achleved, taking Into account thedr relative Importance. Possibfe ra:qns: High, Substantial, Modest, Negligible. Ef1deny: The extent to which the project achieved, or Is expected to achieve, a return higher ftan the opportunity cost of capital and benefts at least cost compared to alternatives. Possfibe raings: High, Substantial, Modest. Negigble. Th rating Is not generaly applied to adjusbment operaons. Sustuinablty The resience to risk of net benefits flows over time. Possible ratings: Highly Ukely, Ukely, Unlikely, Highly Unlikely. Not Evaluable. nsIbtona Development bnp.ct The extent to which a project improves the ability of a country or region to mnake more efficient equitable and sustainable use of ts human, financial, and natual resources through: (a) better definition, stability, transparency. enreabilty, and predictability of institutional arrangements and/or (b) better alignment of the mission and capacty of an organization with its nandate, which derives from these Institutional arrangements. Institutional Development Impact incudes both Intended and unintended effects of a projecL Possible rating High, Substantial, Modest Negligible. Outcome: The extent to which the prjetsr major relevant objectives were achieved, or are expected to be achieved, efficiently. Possible ratings: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory. Highly Unsatisfactory. Bank Pertbinanew The extent to which services provided by the Bank ensured quality at entry and supported Impementation through appropriate supervision (including ensuring adequate transifton arrangements for regular operation of the prect). Possbbe ratWs: Highly Satisfactory, Satisfactory, Unsatisfactory, Highly Unsatisfactory. Benuwer Pefrbnnce: The extent to which the borrower assumed ownership and responsibility to ensure quality of preparaton and Implementation, and momplied with covenants and agreemenxt, towards the achievement of deveopmet obetives and sustainakbty. Possible ratings: Highly Satsfactory, Satisfactory, Unsatisfactory, Highly Unsasfactry. Contents Principal Ratings ........... ..... . _. . , . ...... .. . v Key Staff Responsible ......... ..... - v Preface ... _ .. ... .. . _. ..... . vii 1. Background._. . .. - .. . ........n _ I 2. EFindings .. ................... . ... . ..... . . . ....._ . . . ........ . .2 Lessons .3 Future Directions .............4 3. Analysis. .... ...... ............ . _._ . . .___.... . ..... S Relevance .5 Efficacy .S Efficiency .8 Institutional Development .10 Bank Performance .10 Borrower Performance. 11 &SUianabify11 A nnex A. Basic D.ata S h e e t ........................................................ 15 Annex B. Comments from the Borrower .. .. .17 Annex C. Inplementafion of the Plant Protection Sub-component Supporting IPM Throug Farmer Field Shools... _ . _ ._ . 20 bi repor wa prepare by Rdley Nelon (Task Maager). Wifliam Hurlbut edited the report Helen Pifllip providd ad=unslatJive suppot v Vietnam Agricultural Rehabilitation Project (Credit 2561-VN) Principal Ratings ICR ES Audft Outcome Satsactoy Stdsatory Satiactory Insehtonal Desownent Partiad Modest Modest Impacl Sustinablity Ukely Ukely Likely Borrower Prfokmance Satsactoy Satisfactory Satisfactory Bank Performanco SatsfaCtDry sactory Satactory * The Implementation Completion Report (ICR) is a self-evaluation by the responsible operational division of the Bank. The Evaluation Sunmairy (ES) is an intermediate OED product that seeks to independently verify the findings of the ICR. Key Staff Responsible Tak Mnr e Dvision Chie County" Dcto Aprasal W. Cuddlhy 0. Fox C. Madavo Midtbem Phlflppe Boyer G. Fox Andrew Steer Competion PhMllppe Boyer Mark Wison Andrew Ster vii Preface This is a Project Performance Assessment Report (PPAR) for the Vietnam Agricultural Rehabilitation Project (Cr. 2561-VN) for which a credit in the amount of SDR 69.4 million (US$96 million equivalent) was approved on January 25, 1994. The project closed on June 30, 1999, 18 months behind schedule. The project was extended to enable completion of extension, plant protection, research, rubber rehabilitation, and institutional development activities. The final total disbursed was SDR 68 million (US$98.8 million equivalent, 98 percent of the original SDR amount). An Implementation Completion Report (ICR) was submitted on June 28, 2000 (Report no. 20385)*. The PPAR was prepared by the Operations Evaluation Department (OED) based on the Implementation Completion Report, the Staff Appraisal Report (Report no. 12065, December 22, 1993), the Development Credit Agreement, and review of Bank files. The project was also discussed with Bank staff, beneficiaries, commune officials, govermment staff at the central, provincial, and district levels, and local staff of the Food and Agriculture Organization. The cooperation and assistance of all stakeholders and govenmment officials is gratefully acknowledged, as is the support of the staff of the World Bank Country Office in Vietnam. The ICR is clear, informative, and well presented, although it leaves room for further drawing of lessons. The main rationale for selecting this project was to assess the performance of the Farmer Field School (FFS) approach to extension and the associated Integrated Pest Management (EPM) elemnent. While a financially small component, this is the first time OED has carried out an assessment of an FFS/IPM case - an approach which has evoked some controversy recently with respect to extension efficiency and sustainability. Given the current dearth of successful extension techniques and questions about the data this may warrant further impact study. Following standard OED procedures, the draft PPAR was sent to the borrower for conunents before being finalized. Borrower comments have been taken into account, and are included as Annex B. The bomrwa's conmnts in Annex B sggest that these cost figures need clarfication. However, the differences in percentages noted in Annex B are due to tie fact that the figure of USS107.52 quoted by the borrower is the total project cost figue not the credit disbursed figure and the percntages quoted here are SDR percentages not USS paecentges. The SDRJUSS rate has changed. The borrower is correct that in USS temis the actual credit amount was 102.9% of the appmisal estimate. 1. Background The economy of Vietnam is predomimantly agraian and based on rice production. Farm sizes are small, and holdings, especially in fte nort, are fragmented. The average area per fann worker is about a quarter of a hectare m the nordh and about half a hectare m the south. In 1998, the average amnual income of a rural household was estimated at about US$750 with income from agriculture proding about 50 percent of total income. With the dismantling of coumimes and the hberalition of markets, a surplus of rice has become available and Vietnam is now one of thtree top exporters ofrice in the world. Exports of rice increased fiom about 1.4 million tons in 1989 to about 4.5 nillion tons in 1999. Cropping intensity for paddy is now about 150 percent The average annual growth rate of the agriculture sector since 1995 has been about 4.5 percent, fueled mainly by diversification into high-value crops, increases in rice productivity, and expson of cultivated area including iTigated rice. The most important industrial crop is estate rubber. Rubber yields, although improving, are still low compared to other produces in the region. Other important industrial crops include peanuts, jute, sugarcane, soybeans, coconut, coffee, tea, and mulberry for sericulture. The increase m agricultural incomes resulting from growth in the sector has been a main source of poverty reduction in rual areas. But uban icomes have nsen faster, and the gap betweem nual and urban incomes has widened. About 90 percent of Vietnam's poor live in nual areas and in 1995 the incidence of poverty in rual areas was 57 percent, more than twice that of urban areas. Under- employment in r areas is high despite a high level of litercy at over 90 percent It has been estimated that some 6 nillion people could leave farmng vithout affecting agricultral output At about 6 pacent of the total govemment budget and about 5 percent of total agnculual oulput, public expendtue in agricldure is low compared to other South and East Asian countries. Expenditure on research iS particulary low at about 1.7 percent of public expditu in agriculture and 0.08 percent of agicdul GDP. rigation absorbs a little over 50 percent of the agricultural state budget with about 85 percent dedicated to capital expndiur. Cost recovery in irrigation, at about 50 percent of recrrent costs, compares favomably with Thailand and Malaysia but is weD below the Philipl. The are still more than 5,000 state,owed enterprises (SOEs) in sugar, rubber, coffee, tea, seeds, livestock brdin and fertlizers. Thes we estima to be responsble for about 45 percent of agricultural invesmnt About 40 percent of all SOEs are clasified as profit-making but hiis percentage is considerably lower for food enterprises. The main objective of the project, classified as a poverty tgeted internian (F),' was to support govement efforts to mcrease rural incomes and to assist Vietnam's tntion to a maret economy. This was to be done tough assistance to key agriculture institution sring the newly emerged sullholder sectr and through support for the estate rubber subseetor. The project, the first World Bank-funded project m the sector since 1978, was pected to assist the ongomg adjustent fimn coilectivized agiculture to smalUholder family fanmng through four main components aied at 1. This clasficaiof inTroduced in 1992, is given to a project if it has a specific meaam for bargting the poor (suh as a foo sceuity component or a public wvorks adiem) an&dor if the proporton of poor people amog its beneficiaries is signffi&mdy lae han dte pmportion of the poor in the lotal populaon 2 * Rehabilitating agricultural support services by providing assistance to decentralized extension in 23 out of the 61 provinces, to improve plant protection services including integrated pest mnanagement, and to strengthen research in three key institutes. (16% of project costs) * Providing supervised credit to about 50,000 private smallholder farm households in 10 provinces to purchase fertihzer, improved seeds, and other inputs and to support on-farm development works. (51% of project costs) * Rehabilitating about 160,000 hectares of established rubber of high potential productivity, provide additional processing capacity, and upgrade the quality of rubber being produced through construction of five new latex processing facilities and rehabilitation of four old facilities.(34% of project costs) * Providing traming and technical assistance for policy analysis and preparation of follow-on projects to give more specialized assistance to the sector and to develop a smallholder rubber programn (about 10% of project costs but included within the above sub-sectoral components) The project had some level of operation in 51 of the 61 provinces. It was seen as the first in a long series of IDA projects to develop agriculture in Vietnam 2. Findings Tne project was relevant to the needs of the rural sector of Vietnam at the time of appraisal and, with the help of some mid-course conrection, retaned its relevance. While substantially greater shifts in the direction of liberalzation, poverty focus, and sustainability may have been beneficial, for a first project after 15 years of Bank absence, and considering the substantial need at the time for quick action and rehabilitation in some sectors, it represented a commendable start to a longer-term programzIt pushed reforms modestly, but probably as far as a concern for borrower ownership would permit at the time. Perhaps most important, it provided a relevant investment focus around which to develop a relationship with government in the sector. Efficay is rated satisfctory-in agreement with the ICR. The main objective of the project-to support govemment efforts to increase rural incomes and to assist Vietnam's transition to a market economy through support for key agriculture institutions-was probably largely met, although impact data on incomes is limited. The nual finance component more than achieved its targets at good 95% repayment rates (sitce then up to 98%/e) while stengtheing the institution and giving the institution its first experience of large-scale rual lending to small farmers. Rubber rehabilitation achieved the improved estate productivity planned and largely achieved the improved processing planned although at higher costs than anticipated. The hvestock breeding component performed well. Substantial skills development was achieved. The efficency of the project is rated satisfactory overall though this is more difficult to rate because it varies by component With respect to rural finance, it is now time to move forward on liberalization, but again, over the period of the project, going with a state bank was prudent and has proved a better experience than most public financial institution interventions. With respect to nrbber, it is doubtfiu whether now, 8 years after the project started and 2 years after closing, continued publc sector involvement is efficient However, the argument that, before considering privatization in the rubber sector, the plantations and factories needed to be brought back to a reasonable level of productivity and efficiency under the project, was not unsound. With respect to pest management, as discussed latbr, and more fully in Annex C, there are questions about the efficiency of the small integrated pest manarent (PM) component. 3 Institutional development is rated modest in agreement with the ICR. The failure to carry out the study to reorient the organiztion of the Ministry of Agriculture and Rural Development (MR ), which was highly relevant at the ime, was among the biggest weaknesses of project implementation. There remain some broader institutional questions vwth respect to rubber and rural finance. Also the effectiveness, efficiency, and particularly the sustainability of the Project Coordination Unit (PCU) approach to project management is doubtful. Sustainability-the resilience to risk of net benefits flows over time-is difficult to assess. On balance it is rated likely but this is a marginal call. In this respect, an evaluation advantage in this PPAR is that the mission was undertaken more than two and half years after the project closing date, enabling reasonably good observation of the substantial proportion of activities that bas actually been sustained up to this point. Nevertheless, in some components, resilience may now be threatened. In the rural finance component and the rubber component, in the absence of lieralization of interest rates, accelerated movement on privatization, and improved rubber prices, sustning resilience will be difficult although, VBARD is still financially profitable. Govenmmets strong overall commitment and budgetary constraints suggests the needed adjustments probably wiU be forticommg. They are likely to be carefilly sequenced since this has been a particular concern of both the Bank and the borrower. However, the pace is still slow. With respect to the FFSIJPM component, a special focus in this PAR, efficiency and fiscal sustainability are questionable. The Farmer Field School (FFS) approach has so far reached only 4 percent of Vietnam's farmers over about an eight-year period. Expenditure would have to increase dramaticaDly to reach a reasonable coverage-sy 50 percent over ten years. Moreover, the impact data raise questions about the extent to which FFSs have really achieved incmmental benefits and about the extent to which FFS farmers have sustained some of the 1PM practices m later years. However, there may still be the germs of some promise in the approach-which, for extension, with few susbinable options, is important. Experimentation at lower cost levels is needed as well as frther impact studies including better understanding of firmer to fiamer transmision. This reporating project outcome satisfactory-is consistent with the findgs of the recnt OED Country Assistance Evaluation (CAE) that the performance in the nrral sector was satisfiatory. LESSONS I. Developing a poverty-related logical fiamework helps to clarify the poverty rationale. Poverty focus is more than simply selecting areas where there is a high percentage of poor. (Tlhis project had no logical fiamework matrix and suffered from that The aetent of explicit poverty focus varied across components and the poverty design logic was not entirely clear or consistent. Pardy as a result there was no monitoring of project income impacts. The PT label appeared to be more an afterthought dtan a design premise.) 2. As decenUalization progresses, mechanisms to ensure sustmablty become m a function of reforms that deliver self-sustang instutions and cost recovery at the local level than public expenditure commitment from the center which results m less predictable funding with increasmng decentrliztion. 3. Alongside other OED evidence, this project raises efficiency and susainability questions about the use of separte PCUs-prdy a function of donor prooedural demads. It may be more efficient and susainable to have permanent project management departments, with a career 4 structure, embedded in existing ministries and sharing project management skills, including procurement skills, across all projects. 4. Key elements in scaling up extension include maximizing farmer-to-fanner diffusion, enhancing the capacity of farmers to develop their own solutions, cost recovery or privatization, minimizing the costs of trainers through the use of farmer-trainers, and mass media approaches. Pilot programs in extension systems should test alternative models with different extension intensities to give information that can lead to lower costs and thus greater efficiency. 5. Projects with a number of components and PCUs, and complex links to provincial and district PCUs, and with substantial challenges in procurement, could benefit from the use of Computerized Project Management (CPM) systems to track and manage expenditure. FuTURE DIRECTIONS - For both the Bank and the borrower a number of issues call for attention in Vietnam in the future, some of which are already being addressed under the ongoing Agricultural Diversification Project: * Following some progress on policy reforms and a period of Support for research into yield gains (which were, in economic terms, in the nature technologies for intensification of catching up to the production frontiers), agricultural and diversification is urgently technology increasingly will be a constraint to further needed development in Vietnam. While the mission did not have the opportunity to focus on agricultural research, technology needs are evident at village level, while farmer receptivity appears exceptionally high.2 Given the 7- to 10-year lag in technology flow from research and the need for diversification, research system reform to improve efficiency and increased investment has already been left dangerously late. * Both the borrower and donors should consider revisiting the PCU arrangements, seeking options for consolidation. Separate PCUs, even when coordinated under one body, may not be the most efficient management arrangement. * Following the shift from centrally planned commune governance systems toward increasingly participatory Government staff need more approaches at the village level, more priority should be training in participatory given to training government field staff and leading farm techniques. households in participatory processes. * There are privatization options in the rural sector (in rubber, rural finance, and input supply, for example) that warrant increasingly urgent exploration. However, since such decisions should not be made on principle but on careful analysis of the efficiency of alternative options, and with appropriate phasing and understanding of detailed implementation steps (often neglected), studies of this issue in the different subsectors will be important prerequisites. Some relevant studies are underway. 2. This is a classic example of the Ester Boserup thesis that changing land/labor ratios drive technology demand, in this case the very small and shrinking farn size is driving an exceptional thirst for intensification technologies. 5 3. Analysis RELEVANCE Relevance is assessed as substantial, although, as with many Bank projects of the early 1990s bearig the PTI label, the extent of explicit poverty focus in the design was limited and no commensurate poverty monitoring was included. With the benefit of more recent understanding of the nature of rural livelihoods issues, it is Poverty impact is likely but the being increasingly appreciated that, rather than assuming that any extent of explicit poverty focus rural investment is pro-poor, more explicit poverty analysis and was 1nifted--1onnlstent with focus may be necessary. In this case, the extent of poverty focus a mI label-a logical of different components was variable. The Staff Appraisal Report framework would have helped. (SAR) predated the requirement for a logical framework that might have shown up this strategic weakness, and it was not retrofitted. Mostly the various components do not appear to have explicitly picked up the poorest provinces. Although OED assesses relevance on the extent to which project objectives are consistent with the current Bank and borrower priorities and goals (that is, at the time of evaluation), in this case, where lending was being restrted after a 15-year bresk, weight must also be given to the relevance of the design to the prevailing economic transition circumstances, in effect giving weight to the fact that these were early tentative steps on a relatively unpredictable path of transition for the rural sector. Thus, while significant further progress in the privatization of SOEs, including the Vietnam Bank for Agriculture and Rural Development (VBARD), and in cost recovery and privatization might have been desirable, it is probable that such substantial strides in transition would have been unrealistic at the time and counterproductive to borrower ownership. In other words, the Bank probably pushed as far as it was prudent to push. The relevance of the major switch from support for general crop extension to livestock, which occurred after effectiveness at the request of the borrower at the highest level, is assessed as substantial. It was argued that crop extension was, by then, adequately covered by government and donors but that livestock support services, particularly artificial isemiation, needed substantial and would be particularly relevant to the poor, including those with very little land. This was wisely and rapidly sported by the Bank, an example of rapid supervision response to client needs, although arguably evidence of some weakness in initial project identification and prparation. While seemingly not an explicit decision criteria at the time, this late shift also happened to lead in the direction of cost recovery in one of the easier entry points for cost recovery-artificial insemination and veterinary services-useful early experience for later liberalization. EFCACY Overall, efficacy-the extent to which the project objectives were achieved taking into accoumt their relative importance- The objectdves - albeit was substantld, although this is difficult to assess because predominantly Input oriented indicators were not established at the outset and objectives were - were largely achieved nore input than output focused. To recapitulate, the main although Impact data Is objectives, were to support efforts to increase rural mcomes and lmited. Some targets were to assist the transition to a market economy. These were to be substantially exceeded. achieved by rehabilitating agricultural support services, providing credit for smallholders, rehabilitating state-owned nibber plantations and prwssg facilities, and providing trammg and 6 technical assistance for policy analysis and preparation of follow-on projects, including developing a smallholder rubber program. By component, the following were the significant achievements of objectives: Extension for Livestock In 27 provinces, where the majority of the national herd is located, 5,200 extension agents, 2,365 cattle inseminators, and 25,000 farmers were trained. An estimated 470,000 households benefited over the five and a half year project period from the component as a whole. These achievements were all above the original targets. The 360,000 improved calves produced was more than double the target, and growth rates were 30-40 percent higher than the local breeds. Extension advice was also provided on feeding and other husbandry aspects. A system of private inserninators was set up. The cost-benefit ratios suggest substantial income gains to participating households but no household surveys are available to confirm that. Box 1. The Impact of Farmer Field Schools Farmer Field Schools generally consist of a half-day of training every week covering a 12 week crop season for about thirty farmers. A central element has been 1PM but they cover a wide range of aspects of rice husbandry. There are questions about efficacy and efficiency. While insecticide use in many areas has been reduced and yields have been increasing, this does not appear to be substantially attibutable to the FFSs. FAO surveys indicate that farmers who attended FFS did not increase yields by much more than farmers who did not attend nor did they reduce insecticide use by more. The mission found what appeared to be genuine enthusiasm for FFS at the village level. This may be evidence of perceived benefits not picked up by the surveys. It may, however, simply reflect high demand for a free good. There is also enthusiasm for FFS at the Provicial and District govemment level. Tlhis may reflect a perception that FFS is an effective extension system However, it is difficult to conceive of Provinces or Districts not welcoming this substantial donor support There are also questions about sustaitabilfty. Over 10 years the national F!7S program has reached about 4% of farmers. It is currently over 80% donor-funded. To reach a reasonable coverage, say 50% of farmes over 10 years, would require about 10 times the past level of total expenditure, and, in the absence of donors, about 40 times govermnents past expenditure - alnost certainly unralistic. If famer to farmer transmisson were high, a low farmer coverage might be adequate, but findings elsewhere, and mission field observations, do not suggest high farmer to farmer transmission, although certainly some. Given the questions on susainability, it would have been valuable to have had more experimetation over a range of depth and coverage, including, in particular, a high share of mass media Reducing costs without excessive loss of knowledge transfer and raising cost recovery appear to be the main future options. Plant Protection. Some 68,500 farmer-leaders were reached through about 2,300 FFSs against the target of about 1,600 FFSs. However, as discussed in Annex C and in Box 1, the FAO survey data indicates that firners without FFS training but within targeted Provinces also reduced insecticide use and increased yields. Box I addresses this more fully. Loan volume in the credit Agricultwral Research. While this was a small, "bits and pieces" component was well beyond component floating largely without an overall research strategy, it target and high repayment did lay some groundwork by training 2,000 scientists, equipping rates enabled substantial three food crop research institutes, and expanding the tial recycUng. The public finance program on rice. For 10 provinces, higher-yielding rice vuieties, ins votuteon was largety mcluding some new high-yielding hybrids, were released, partly effective over the project arising from project-supported work. period but policy and Institutional reform are now needed. 7 Rural Credit Component. This component, which provided short-term loans for crop and livestock inputs (average loan size about US$280) and medium-term credit for improved livestock and machinery to small-scale farmers, more than achieved its targets of lending with about 95% repayment rates although, the household income impact is not clear. This was achieved through a public rural finance institution-effective for lending in the early years but, with interest rate control and imposition of additional tusks by government less promisig for sustainable rural finance in the longer-term. There was more focus on rural credit than on rural savings. Most of the savings fumds used as the source of credit were fumds mobilized from urban areas although savings in rural areas are now increasing. Perhaps the most imnportant achievement of this component was that it gave VBARD and Vietnam its first experience in small farm rural lending. It benefited an estimated 750,000 fann households and as many as 2 million if recycled loan funds are included-an extraordinary scale of lending achieved very quickly-partly due to the discipline of the organized commune system already in place. Repayznent rates were consistently above 95 percent. (Under the follow-on Rural Finance Project repayment rates as of May 2001 were up to 98 percent). Rubber Rehabilitation. There was no intent to directly impact on short-term poverty with this component. Estate and factory rehabilitation was expected to provide growth through improved production and indirectly sustain longer-term employment but in the short-term excess labor actually needed to be reduced. Improved productivity was achieved. In rehabilitating plantations, the period of immaturity of rubber was reduced from 8 years to 6.5-7 years, meeting the targets. Average yield increased by about 30 percent. Incremental rubber production from 1994 to 1998 was estimated at around 50,000 tons. This has increased since then as more trees have reached maturity. Five new factories were constructed and completed after some delays. Only one old factory was rehabilitated since the cost was found to be high and anticipated quality improvement benefits modest. Factory capacity increased by 48,000 tons, significantly below the SAR target of 74,000 tons. This was mainly due to the lower number of rehabilitated factories, the smaller size of new factories, and the higher- than-projected factory unit cost. Quality of processed latex improved, which has been important in trying to hold market export share against strong competition. However, marketing remains a problem with a recent slump in world prices. Eight rubber effluent treatment plants reduced negative enviromnental impacts. The mission found the quality of construction of the factory and effluent treatment plant satisfactory. Rubber research programs were supported in a few key areas. A smallholder rubber production study was completed and provided a basis for the follow-on IDA- financed Agricultural Diversification Project currently under implementation with generally satisfactory performance. Sector Management. A comprehensive training program for the MARD was implemented. It trained 1500 officers, largely The Bank should have taken a achieving the targets. A study aimed at reorienting the mmistry stronger position on the need for organization was never undertaken, however, because of the institutional study. structural changes already underway. This was a serious weakness in achieving the objective of supporting the program of liberalization in the sector. Although some adjustment of the study objectives and timing may have been appropriate to accommodate the changing circumstances, the Bank should have taken stronger action to sell and initiate a modified study. Efficacy of Consultants. The borrower/implementing agencies were happy with the quality of the technical assistance consultants. Consultant recommendations reviewed by the mission appeared generally sound - in one case excellent. 8 EFFICINCY Economic Rate of Return. The ICR, while not aggregating the components in an overall project economic analysis, found satisfactory ERRs or C:B ratios for all the main components, although the analysis was limited in depth . This PPAR did not have the resources to develop a more complete analysis but broadly agrees with the coefficients used. Based on the evidence presented it would be unlikely that the ERR would be below the opportumity cost of capital. However, this PPAR questions the economic efficiency of the FFS portion of the Plant Protection component mainly on the basis of the FAQ study findings. As noted in Box 1, the study found that non-FFS farmers also increased yields and reduced pesticide expenditure, although they did so to a lesser extent than FFS farmers. Second, it found that after Efficency of the smal FFS support several years FFS-trained farmers shifted back to higher compoient o thest uncert levels of expenditure on insecticides, although still lower component is at best uncertah. a n than before the training. As noted in the IPM Annex C, the survesi yields were higer and .. A . . . h~~~~~~Isecticlde use lower for FFS farmers but findings of the FAQ study are at vaiance with the non-FFS farmers in area without FFSs widespread enthusiasm for FFS and IPM among fanners also Increased elds and reduced interviewed and by the widespread demand for FFS inse hdeuse. Aldso threde d training-but one would expect enthusiasm for the only inSectigIde use. Also, three years after extension option available. This training is now party trainig, nseticide use had risen agahi. financed by farmers and communes, but farm There is idespread farmer enthusiasm contributions remain very small. They should be rised for FFS but ths does not test efficiency. substantially. Two further aspects of the FFS efficiency issue, on which there is insufficient data to pass judgment, are: (i) it could be argued that a share of the benefits associated with FFS tamiing come with a significant lag through the enhanced capacity of famners in developing their own solutions, for example by experimentation on their own plots; and (ii) there appear to be positive environmental externalities but there is no data to place a value on them. Rural Credt. The relatively low operating costs at about 4% and good recovey-now 98% suggest adequate levels of operational efficiency. But profitability is still elusive. There is some subsidy to beneficiaries because government controls interest rates, bas a scheme to forgive and refund to VBARD loans in default under certain circumstances, and some administative activities are done by govemment staff. There are therefore some longer-term financial sustainability questions. However, over the period of the project efficiency improved and the financial strength of VBARD, initially very weak, improved. Currently, VBARD makes a profit if accounting is done following the officially approved Vietnam Accounting Standards (VAS) but not if it is done following International Accounting Standards (IAS). Provisioning for bad debts was, and still is, too low. There were, and still are, State Bank of Vietnam ceilings on on-lending rates, although cost of funds and operation is part of the consideration in setting this rate. Briefly, the 2000 situation using IAS was approximately as shown below: Interest Expense as % of Earning Assets (EA) 4.9% Cost of Operation as % of EA 4.1% Loan Loss Provision as % of EA 4.00/O Total Earnings as % of EA 10.9% Profit/Loss as % of EA -2.5% (1999 was -0.4%) As is evident from the above, there was a deficit in 2000, although this is reported to be somewhat improved in 2001. (At the time of the project interest rates were, for most of the time, just positive in 9 real terms reaching a high of around 2.5 percent per month due to the higher inflation rate compared with about 1 percent now.) . The table below indicates some key elements of the Profit and Loss Account and Balance Sheet at the end of the project in 1999 and 2000 based on Intemational Accounting Standards. Us$ mOllon 1999 2000 Total Earnings 267.4 283.0 Total Operating Cost 43A 103.6 Provson fr LOSS 400,223 204,019 Net profit after tax (taxed since profit (9.0) (1996 - (38.3)) (64.9) made by Vietnam account*ig standards) Total Assets 2,621 3560 Total Uabiltle 2,540 3545 Funds Empbyed (EquIty) 81.4 15.0 Despite improving operational efficiency, there has been a decline in capital and reserves and continued losses. Government policy is to move toward liberalization of interest rates, but there has been limited progress to date.3 Discussions are curently proceeding on banking restructuring in Vietnam, including VBARD, under the forthcoming Poverty Reduction Credit and associated with the Rural Finance H Project. Ways to address the inherited bad debts are expected to be part of this restructuring. Bank staff consider it unmlikely that VBARD could be sold to a private buyer as a single entity due to its very large size, however, there may be possibilities for gradually selling off local groups of branches. Obviously, there would be no buyer until interest rates have been freed off. Cunently, a high level of about 85 percent of funds are mobilized from savings with only 10 percent from donor assistance. Commendably, early in 1994, partly associated with systems and technical support from the project, VBARD introduced International Audit Standards- a significant step forward-which enhanced reputation and financial oversight. The next anticipated shift is to keep the accounts in LAS fomat. The economics of the rubber component was satisfactory. However, one efficiency issue is whether the use of herbicide funded under the project for plantation rehabilitation was really warranted in a situation of over-employment where SOEs were finding it difficult to reduce labor (although they gradually achieved it). Hand-weeding of lalang grass rather than using herbicides funded through the project may have been more financially efficient, at least until such time as estate labor levels had been furter reduced. Based on a discussion at the estate level, this may explain why herbicide use was lower than projected. It may have been an appropriate reaction by some estate managers to the excess labor issue.' 3. The Bank's 1995 Financial Sector Review did not recommrtd, at that stage, allowing the market to detemine interest tes because: (a) maccnomic stability was not felt to be yet strong enougb; (b) perforrance of the bankig sysun was not sufficiatly sound; and (c) tder was concern that liberaliwtion might encourage a shift into more risky lendig ae Some of those concerns should now be substantially dissipated, but ticy probably were valid over much of the life of the projea 4. However, this issue is not clear-cut because unless hand-weeding is done thoroughly to get at deep roots it is likely to be less effective than herbicide. 10 INSTITUTIONAL DEVELOPMENT Institutional development impact was modest, a rating also given by the ICR. The mission has doubts about the efficiency of a management structure using PCUs. There are longer- Efficiency and sustainability of separate term sustainability issues since PCUs come and go project PCUs is questionable as it often with projects and since staff tend to feel outside the results in proliferation and duplication mainstream of ministry function. While the use of which it has probably done in this case- PCUs may have been practical for a first project such the solution lies partly in standardized as this after a long Bank absence and with weak lead donor procedural requirements. ministry skills, donors continue to encourage separate PCUs within MARD even today, often with part-time staff with other responsibilities. This leaves potential for duplication of skills in management, procurement, and technical support. The alternative of a department within MARD to manage mrual projects was floated several years ago by the project director of ARP. This seems to offer a more institutionally embedded, efficient and sustainable structure than the cunrent arrangement with an umbrella coordination committee5. VBARD. Generally satisfactory institutional development was achieved within the limitations of a public finance institution still operating under a number of government imposed constraints. Significant improvements were achieved within the organization itself. At the time of appraisal VBARD had been making losses for the years 1991 and 1992 (about US$4 million in 1992). Profitability has now improved, partly due to increased volume and efficiency improvements-project investments in institutional strengthening being partly responsible for turning this around. Privatization and freeing of interest rates at the time of the project was not a realistic option for rapidly achieving a substantial scale of small farmer rural lending. Therefore institutional strengthening of VBARD as a public sector financial institution was entirely appropriate as a first stage. However, as indicated above, now that rumral lending has become well-established there lie ahead some important broader institutional development and policy decisions within the framework of the current financial sector restructuring dialogue. Training. The training advisors and training components supported by the project appear to have been particularly relevant and effective. Outlines of training courses reviewed by the mission were relevant and practical for the needs of subsectors. The human resource development components for both General Rubber Corporation (GRC) and VBARD appear to have had a sigmficant impact on the institutions and are paying off in the follow-on projects. Documentation of the training courses on file is above average and may be partly attributable to the Bank skills applied during supervision. The Bank was flexible and BANK PERFORMANCE provided strong technical skills, but the amount of Bank performance was satisfactory. Preparation was strong and procQrement carried out by included sector analysis in collaboration with other donors. An non-specialists raises appropriate linkage was formed with the FAO IPM initiative, and questions about the there was a suitable degree of pragmatism by the Bank regarding efflciency of the Bank's what was realistic in this emerging economy. Supervision appears skills mix. 5. The boower comms in Annex B that three seprate PCUs were needed for the three separate components in diferent fimctional entities. Howeve, the PPAR here is suggesting that different manget PCUs wim MARD for different donor projects may not be the most efficient and sustanable armmgement, although pragmatic at the time. 11 to have been flexible and quite creative. As noted above, the task manager reacted appropriately to the request at the start of implementation for a substantial change in the agricultural services component from the crop orientation to the livestock orientation, although the need for this may indicate some weakness in appraisal of the extension component. In 1996, the files indicate that Bank technical staff were conscientiously questioning the fit between training being requested under the rubber component and the project human resource objectives. There were 14 supervision missions over five and a half years, averaging more than two a year, a relatively high intensity. Skill mix was generally adequate- particularly strong in the important training area and, in the early years, in areas of technical specialization such as rubber. Some highly experienced Bank technical specialists from outside the region were brought into the teams-technical Bank skills that are now in decline. The large volume of correspondence on procurement in the project files, while not necessarily reflecting the relative importance attributed to this subject by Bank staff, shows, as OED so often finds, a huge routine procurement workload. One set of files over a fixed period taken at random had well over 90 percent of the documentation on routine procurement issues, mostly handled by the task manager at HQ. It is doubtful whether a task manager skilled at the design and supervision of rural projects has a comparative advantage to perform work that could be more efficiently handled by procurement specialists with procurement support staff. However, Bank performance on procurement is not in doubt.6 BORROWER PERFORMANCE Borrower performance was satisfactory. Indeed, allowing for the absence of recent project implementation experience, in some respects it was more than satisfactory. This was a borrower starting to take a new direction with many challenging reform changes to be accommodated and new skills to be learned. While better performance would have been possible in some areas-such as the release of counterpart funds, the institutional study for MARD, and some areas of procurement- ownership was strong from the start and skills were quite rapidly built although the legacy of that remains liMrited given staff movement and retirements. The request at the outset for the shift in the extension component toward livestock was evidence of proactive ownership that did not simply acquiesce to a blueprint. On the negative side, in 1996, two years into the project, there were still problems of funds release evident in Bank correspondence due to 'complex budgetary, financial control, and fiund flow procedures."7 Procedures at that time were under review and they were later modified. SusrAINADILrfT Sustainability is rated likely, but this is a marginal call. With two and a half years since project closing the mission was able to observe some significant period of post-closing acual sustainability and resilience. But ther are some sustainability risks. For the overall rating, given the relative scale of components, the Rural Credit and Rubber components are the most significant and although, on balance, both seem reasonably resilient, they both present stiability questions. With respect to components the sustainability situation is as follows. 6. For example, the files indicate that prompt action was take on a procurment conmphint wher a poential bidder demonstatd that technical specifications for an item of equipment limited responsive bidders to one equipn model 7. For example, the project files indicate that in 1996 funds for the FFS in four centmal provinces arrived some two months afler the FFSs had begun and departments had to borrow finds from elsewhere or in some cases delay payments to trane Some FFSs had to be cancelled. 12 Rural Credit. The component is rated as, on balance, likely for sustainability but this is a marginal call. This positive assesmet is on the grounds that VBARD, within its Vietnam institutional setting of strong government comrmitment, is arguably quite resilient, although still not profitable by International Accounting Standards. It has been more successful at delivernmg flmds relatively inexpensively to large numbers of farmers than most other agricultural banks in the region. However, it exhibits a recently inproved but then agam, in 2000, weakened equity position. As aheady noted, it is squeeed by goverment interest rate ceilings and other forced obligations which, while manageable in the short-tenm, are not sustainable in the longer-term unless strong action is taken fairly soon. It is the mission's assessment that, given Government's strong overall commitment and the current budgetary constraints, the overdue action probably will be forthcoming over the next couple of years in association with the new rural finance project. But there remains a risk. The project intention is to ensure improved interest rate spreads, agreed procedures for setting interest rates, and elements of an overall financial sector restructuring program. Anticipating lower taxation in future based on IAS rather than VAS standards, to achieve an adequate spread is projected as requiring about another 4% on the interest earnings - an increase that should be achievable. It has been assessed that the market for small rural loans could bear significantly higher interest rates against the much higher informal rates. Also of relevance to the sustainability question is that, in 1995, the goverment established a "Bank for the Poor" outside the project to lend at subsidized rates to cormme-generated lists of those below a subsistence level using some VBARD staff capacity. It is a very small operation comnpaed to VBARD. World Bank concerns about demand-related impacts on interest rates and more generally about the impact on institutional sustainability have been expressed. Although small, the "Bank for the Poor" clearly represents some additional sustainhility risk to VBARD which needs to be managed. VBARD have not always been reimbursed in a timely manner for this service. The Bank for the Poor results in competition at village level on interest rates. While, its existence as a separately accounted entity may at least help to keep separate the books of accounts to identify the explicit subsidy of this particular operation, it cannot entirely enclave it. Training is now sustained by own funds through a VBARD training center and training appears to be continuing at a substantial level. The new Rural Finance Project, the successor to this credit component, is progressing generally satisfactorily-supervision ratings are satisfactory across all criteria. Rubber Component. On balance, sustmabiity is rated as likey, but, again, this is a margin call. Technical sustainability is not in doubt but there are uncertamties about financial susaibiity. Both plantations and factories have been rehabilitated and/or built and are generally well maintained. Producivity and quality are improving. However, there has recently been a period of very low rubber prices and some SOE's are stnugglg financially. At least three SOEs are currently in arears on loans from VBARD, pardy because, with current low prices, they have held back substantial unsold stocks and have serious cash flow problems. However, in aggregate it is reported that the 10 companies supported under the project made a profit in 2001 and the aggregate profit level has been improving since 1998/99. The current rubber price situation would be serious for industry resilience if it holds for long at the current low level. However, with low wages relative to competitors and yields and quality now imPrOVing due to new planting and improved processing facilities, the Vietnam rubber industry appears to hav a significant comparative advantage over the commg decade and a better chance of being sustained than most of the higher labor cost competitors. Privatization is the next step now that improved productivity has been achieved. With respect to environmental sustanailty, there appears to have boen a violation of envionental safeguard procedures but satisfactory performance. With the rubber procesing component, this project should have been reported as a Category B project at the time of appraisal and Board presentation. 13 However, the effluent problem in rubber was given a full assessment and the mission found fully satisfactory implementation in the field. Livestock Component. In this component, sustainability is variable by province and district The mission observed that some districts are continuing public fumding for artificial isemation services and some doing much less. However, those districts where sufficient private insemmators have established a fincially viable practice (generally charging US$1 per isemination for labor and tavel plus US$1 for materials) have a reasonable chance of sustaiabiity since demand appears to be growing. Even if Al or exotic bulls' natzal breeding were not sustined the genetic injection itself from the project will be susabned with increasing breed dilution but compensated for by increasing spread. In other words, genetic infusions are not, in fiact, lost, only diluted with each generation after the infusion. In AL a fallback position being adopted in some Districts under reduced public investment after project closing is slower genetic improvement but with lower overhead costs using natural breeding with improved bulls. Plant Poteciion. The susainability story for Integrated Pest Management is complex. The current Community IPM program (CIPM) is, in fact, being sustained quite well but with a high share of funding from donors. The program, fimded by FAO and Norway but no longer by the Bank, covers 83 percent of costs; the remainder is covered by provincial and district governments, cooperatives, and farmers. Most of the direct farmer funding is, appropriately, for the lower cost Farmer-to-Farmer Field Schools. As of August 2000, 19 provinces had significant CIPM programs. But m the longer term, allowing for inevitable donor Only 4% of farmers have attended an withdrawal, is this program sustainable? Over eight FFS. To reach a significant number In a years the Vietnam IPM National Program has trained decade will caU for public expenditure 515,900 farmers. At a cost of about US$30 per farmer beyond govermnent capacity unless there including management, training and TA overheads , this are unusually high levels of farmer-to- represents a little under USS2 million per year'. This fanner diffusion, ways to reduce costs still would be about I pecent of the total budget for further; greater cost recovery; and, agriculture and about 4 percent of the recurrent modest needs for "booster shots." expenditures.9 Yet, so far the coverage of the program Currently, the best may be the enemy of has been limited. To date, only 4 pement of Vietnamese the good. Reduced costs and greater cost farming households have had a member trained in a rice recovery are by no means unattainable. IPM FFS although about 70 percent of communes have had at least one FFS . To achieve say 25 perct to 50 percent farmer coverage over 10 years would require about a five to tenfold increase in the level of expenditre at past unit costs, somenwhat less if technical assistance and overhead costs per farmer could be reduced as skills and efficiency improve. But this is not allowing for any 'booster shot" FFSs that would probably be needed to maintain skills. It appeas therefore to be not sustainable in its present form at present sectoral resource levels without a large increase in cost recovery. But, outcomes and sustinability cannot be easily assessed without a clear understading of the indirect multiplier effect of FFS traiing. It is therefore extremely important for the fiture of technology dissemination in Vietnam to test out lower cost altematives and "cocktails" such as the IRRI experiment multi-media approach quoted in Annex C and to monitor and 8. The borower notes in Annex B that the costs of the IPM component of the project wee S1 8 per fimner. However, this does not cover al dhe thaig, management, resec and TA overbeads to apply a FPS apprah. Estmates of only the duect costs obtamed il the field came to earlly that fiure. The Mmission behes th the true fill econcmc cost with al overheads is closer to about S30 but the fully costed figue does not appear to be known. 9. Irigaton absorbs about 50 percent of the agricultue budget, reforestation and land reclanalion about 1S percen4t and foresty services about 10 percent About 60 percent is spent at the central level and the rest by local goveanment in the provDne 14 understand how well the farmer-to-farmer transfer is occurring from FFS trained farmers and what factors enhance it With respect to envirnmental sustainability, the project reportedly had significant positive environmental nmpacts in several areas: reduced insecticide use and some improved balance in fertilizer use for adopting farners, and improved effluent management in rbber. However, there is no environmental data on the 1PM impact. 15 Annex A Annex A. Basic Data Sheet VIETNAM AGRICULTURAL REHABILITATION PROJECT (CREDIT 2561-VN) Key Project Data (Amounts in US$ milHon) Appraisal Actual or Actual as percent of Edimate cunent estimate Apprail esimatb Total project oosts 106.7 107.52 100.8 Credit amount 96.0 98.8 102.9 Cumulative Estimated and Actual Disbursements FY34 FY95 FY96 FY97 FY98 FY99 FY00 Appraisal estimate 5.0 35.0 75.0 96.0 Actua (US$M) 0 65.4 72.6 88.3 96.5 98.4 98.8 Actual as percentofappraLs 187 97 92 101 103 103 Date of final disbursement December 2, 1999 Project Dates Orginal Actual IniHating memorandum 7/90 Negotiatlons 12193 Board Approval 1125194 1/25194 SignIng 313/94 3/3194 EffecIeness 6/1/94 6N194 Cosing date 12131/97 6/30/99 Staff Inputs (staff weeks) Actual Weeks Actual US$000 Prappralsal 124.2 274.0 ApprIsal 146.4 371.2 Negotiations 19.9 57.7 Supervbison 168.1 475.5 Completbn 14.0 60.0 Total 472.6 1238.4 * Superision began In FY1994 " AS of June 1999. no furter report was avaiable under SAP-Estimates Indude pardcpation of FAO/CP 16 Annex A Mission Data Dahi No. of 3f f S.ilzf. Pwmmop ,ufuig 7Amp of Pmb4sms' (nnVWw ,_ dyts h ,eesend Appaiaal Febnuary 1993 6 16 EAF.S PostApprsal July1993 2 18 E,F Appnrisal NA NA NA through Board AppWoval Supervisn March 1994 5 24 EAFAP.S S HS L I iJuty1994 6 22 A,T(2),S(2),Ex S HS Novenber 2 20 AS S HS A 1994 2 8 A,F S HS FPA Febuary 1995 2 14 A,S S S FA May 1995 1 7 A S S T.P August 1995 7 16 A.T(2), L. S(3) S S F Novenber 1 8 A S S A 1995 2 15 AT S S F,T Februa 1996 3 19 AL.S S S A May 1996 2 14 AT S S FAT November 3 9 A,T,L S S F 1996 2 7 A S S F may 1997 1 15 A S S F Januay 1998 Sept 1998 March 1999 Completion Odobor 1999 2 14 E,A Completion Febary 2000 1 14 E A-Agronornist,E-B-Eomit, Ex -Extensid., P-Foacial, L-Ivuok P-Procrament SSpeciaHt (e.& nabbf) T= Tlribg/Inhimntzan * FA0/CP Mission Other Project Data _oofMc A som. FoLLou.OW_ON_ __n_ opealown CQ Ft no. Ama Bow dteb flS$ amOon) AgiwAe Dlveaflon June 1996 Ifmgatin Rehabitafion Aprl 1995 Mekong Delta Water may 1999 17 Anmex B Annex B. Comments from the Borrower MMISTRY OF AGRICUJ TURE SOCIALIST REPUBLIC OF VIETNAM AND RURAL DEVELOPMENT Independence - Freedom - Happiness Department of Planning Hanoi, June 5. 2002 and Projection To: Mr. Alain Barbu, Manager Sector and Thematic Evaluation Group Operations Evaluation Department The World Bank Dear Mr. Barbu: Re: Vietnam- Agricultural rehabilitation Project (Credit 2561-VN) Comments on the Draft Project Performance Assessment Report Thank you very much for forwarding us the Draft Project Performance Assessment Report (PPAR) on the Agricultural Rehabilitation Project (Credit 2561-VN), including the Performance Audit Report, the Memorandum to the Executive Directors and the President, and the Annex A, Annex C. We realize that, it is very helpful for us to withdraw interesting lessons from this very comprehensive and detailed PPAR, even two years after the Project Implementation Completion. In general, we would have no objection regarding all parts of the PPAR, including Principal ratings, Findings and Analysis. However, as being one among the three Project Coordination Units as well as Project Implementing Agencies, we would have some few detailed comments presented in the attachment hereafter. We expect that these comments would be taken into consideration and will be reflected in the final report that we will receive later. Thank you for you kind cooperation and assistance. Yours sincerely, rTrang Hieu Dung, Director Department of Planning and Projection Ministry of Agriculture & Rural development cc: Mr. Abdul Aziz Mohd. Yaacob, Mr. Andrew Steer, Mr. Mark Wilson, Mr.Philippe Boyer 18 Annex B Attachment CO6MMENTS FROM MARD ON PPAR 1- Project final total disbursed The Project final total disbursed was US$107.52 million equivalent, 100.8% of the original amount (US$106.7 million at Appraisal estimate) as presented in the Memorandum to the Executive Directors and the President and in the Annex A of the PPAR. This figure is coincident with those of stated in the Bank ICR and the Aide Memoire of the ICR Mission in February 2000. Whereas, in the Preface of the PPAR it was stated that the final total disbursed was SDR 68 million (US$98.8 million equivalent, 98 percent of the original SDR amount. We learned that the amount of US$98.8 million is the IDA Credit amount and the percent of this is 102.9% comparing with Appraisal estimate. Therefore, a clarification of this would be required. 2- PCUs efficiency and sustainability The use of three separate PCUs for three project components can -profiferule partly to meet donor procedural demands" . But in case of the ARP, the first IDA funded project for Vietnam Agricultural Sector, with highly diverse activities. benefited by different functional entities (MARD, GRC, VBARD), following different financial mechanisms for three different project components, the project worked best with a separate PCU for each component. Obviously, it would be more efficient and less difficult in implementation if a Joint Steering Committee would have been established. 3- Institutional development The postponement of the study for restructuring MARD during the late of project implementation should not be assessed as among the biggest weaknesses or serious weakness of project irnplementation. By the end of 1995, the three former Ministries-Ministry of Agriculture and Food Industries, Ministry of Forestry and Ministry of Water Resources were merged into a new and biggest one- Ministry of Agriculture and Rural Development. This structural change caused some difficulties for the Sector Management sub-component to proceed the implementation this study. Moreover, this complex and difficult study could not be completed in a short time remained and with a very modest fund allocated for the Sub-component. At the same time, by that time the on-going internationally funded project of Forestry Sector Reform was already underway and the Public Administration Reform was in the course of preparation. 2 19 Annex B 4- IPM through Farmer Field Schools under Plant protection Sub-component * The integrated pest management (IPM) is a main activity among the others (general plant protection, plart quarantine strengthening) under the Plant Protection Sub-component. The final total disbursed of this Sub-component. was 2.7% (US$2,667,500 equivalent), in which IPM was only 1.2% (US$1,250,200 equivalent) of the total project costs. IPM is an approach of multi-purpose, mainly for insect pests control but not plant diseases and weed control. So, in most cases in order to avoid losses caused by diseases and weeds, even IPM trained farmers have to spray fungicide when certain disease was notified, and pray herbicide in case of high labor cost of manual weeding (mainly in the South of Vietnam). It does not mean that the IPM trained farmers are still spraying or increasing insecticide use. * Under the ARP, 68,500 farmers were reached through 2,300 FFSs at a total cost of US$1,256,200 (not US$2.61 million presented in Annex C), giving the cost of US$18 per farmer (not US$38 in Annex C). The lower cost per IPM trained farmer through FFS is because of the regulated low unit cost in Vietnam, but has never been contributed partly by the trainees or the commune (as mentioned in Farmer Field School costs and Extension Alternatives, Annex C, page 21). 20 Annex C Annex C. Implementation of the Plant Protection Sub-component Supporting IPM Through Farmer Field Schools The integrated pest management (1PM) component, though only about 3 percent of total project costs, is of particular interest because it is the first time OED has evaluated a new-generation of IPM projectJcomponent. It is important for future Bank learning to draw lessons from new approaches as early as possible. The extension method used-the Famner Field School (EFS)-is of interest in itself as a technology dissemination method. This analysis focuses on empirical evidence to the extent possible and atterpts to separate questions related to IPM from questions related to FFS because the forner is an approach to integrated crop maPnagement with an emphasis on pest management whereas the latter is a knowledge transfer technique that could be relevant even in an entirely pest-free environment. It should be noted at the outset that nearly all the survey evidence emanates from the various levels of the IPM support program itself. Bol 1. Integrated Pest Management IPM is an approach to crop pet management, initiated in Vietnam in 1992 with FAO support, which follows ecological prnciples. It is predominantly used on paddy but is expadig into other crops, especially vegetables. So far, it is only used for insect pests not plant diseases. The aim is to exloit natural enemies of damaging insect pests. It relies party on early season plant compesation for vegetative damage by pests. It does not avoid the use of insecticides where needed but aims for reduced application consstent with maig yields and for icreased pest specificity. Thus, it is intended to have environmental benefits off-site. It incorporates such elenmes as pest-resisbnt variety selection, planting times consisent wth reduced pest damage, optimal and pest- reducing planting densities and fertilizer application, and appropriate husbandry, including water management aimed at pest reductian and plant vigor, and appropiate land preparation. It is generally supported by removal of subsidies. Pest Damage In Plots Using lPM Compared with Those Not Using 1PM No evidence was found that the level of paddy pest damage is lower in those plots using an IPM approach compared with those not using an 1PM approach. Generally, farmers not using IPM appear to be spraying as fiequently as needed to maintain a substantially pest-free environent. The benefits are therefore largely the pesticide cost savmgs. ield Comparison Between IPM Plots axd Non-lPM Plot The analyses available do not allow differentiation between IPM and non-IPM plots holding all other input factors constant However, no one interviewed claimed that, from IPM alone, yields could be expected to be higher than a fiul inscticde spray regime. The main differences noted were with respect to reduced insecticide costs and (observed but not yet quantified) externality benefits. Yields 1. Bnefly, the methodology adopted mvolved surveys Seven provinces whch surveyed sanple fmers the season befor IPM aig and die sane fams the same season afler PM training i.e. two seasons latr. For a smaller control group in two ofthe provnces, sveys were undertak of frmsm who were not scheduled to take part in IPM traing and did not ltve h arew where an FFS wasplanned. This was carried out in the same season as the EPM group. The same control group sample ws then resuveyed two seasons later as for bie EPM group. 21 Annex C remaining the same following IPM reduced-spray regimes is consistent with IRRI research on expei}mntal plots. Costs Comparison Between IPM Plots and Non-IPM Plots Using the data from the 2000 FAO study,2 and calculating the arithmetic average of actual insecticide cost saving and percentage insecticide cost reduction reported in that document across the 7 provinces surveyed, gives a mean insecticide cost saving after FFS training of about US$5 per hectare and a pesticide cost reduction of 80 percent. However, in the two control groups of farmers-Quang Ngai and Binh Thuan provinces-in areas where no FFSs were undertaken, farmers who did not attend FFS showed a higher absolute cost reduction, although it was lower in percentage terms (66 percent over the same period). Thus, it appears that farners not trained in EPM were also lowering insecticide expenditure. The expenditure levels on insecticides are not a large part of the input costs and, therefore, savings in that area alone cannot make large impacts on costs. Typically, insecticide before 1PM (see Table 5-12 in the FAO study) was about 3 percent of input costs including labor, or about 5 percent excluding labor. As a comparison, a 5 percent reduction in fertilizer expenditure (excluding labor differences) approximates the cost reduction from reducing insecticide expenditure to zero. There is a possible explanation for reductions in pesticide use beyond FFS training that may account for the decline in non-FFS fanmer insecticide use. Especially in the north, Vietnam has been shifting away from centmlly planmed control reaching down to the commiue level. Previously, spraying was often determined by commune technicians who, responding to local district edicts, largely determined when and how much to apply and probably had incentives to apply often and at high levels to avoid any risk of being accused by superiors of letting an infestation get out of hand. Indeed, commune teams often carried out the spraying for farmers. Somne decline in pesticide use right be expected, therefore, as management decision-rnaking gradually returned to farmers. Communes are now evolving gradually-but with wide variation between Provinces-into participatory community groups (possibly benefiting from the history of discipline compared to community groups that have to build from scratch). However, a counteracting trend in the incentives for pesticide application levels has been the real decline in pesticide prices. With respect to costs, there are also labor reductions, since less spraying takes less labor. However, there are also costs in checking on levels of pest damage and generally managing the IPM technique that may offset this. Unfortunately, there is no data in the surveys that allows the net labor change to be calculated. However, the difference in total crop management labor excluding land preparation transplanting and harvest between IPM and non-lPM farmers found in Chung and Dung 1996,3 could not be more than approximately $4 (the total difference in hired labor for crop management and total household labor for all purposes) and probably considerably less. 2. The npact oflPMFaner Felu Schools on Farmers'uQdton Practi in Their Own FiL. A report submitted to the FAO In_county Progamne for Commnity 1PM in Asia. Jonathn Pincus, Economics Pepartm School of Oriental and Afican Studies, Univesity of Londbn. August 2000. 3. Pest Management in Rice Praduction in Vlenam: A Socioeconomic Assesment. Do Kim Chung and Kim Ibi Dung, Faculty of Economics and Rural Development, Hanoi Agriculunal University. Hanoi Sqptember 1996. (Dr. Chung was a consulant for the PAO study.) 22 Annex C IPM Coverage Across Different Crops Under the Agricultural Rehabilitation Project the majority of activity was on paddy. However, there is an increasing shift evident toward applications on vegetables, tea, maize, fruit, and sugar cane. The relative scale of this shift is not yet clear from the data. The Level of Interest of Different Stakeholders in IPM Considerable enthusiasm for both the FFSs and EPM was evident during the village visits by the mission'. There was also widespread enthusiasm among the provincial and district technical staff. Observation suggests that farmer enthusiasm is greater for the FFS than for the narrower 1PM technology itself, but both are seen as important. As discussed later, the FFS training has broad coverage of all main rice husbandry topics over the 12 half day sessions through the season. The interest at the farmer level may also be reflected in the growing number of FFS covering subjects beyond simply rice, for example, on rat management, soybeans, vegetables, tea, fruit, maize, and sugarcane. However, since for farmers, in most areas, there are no other alternative technology transfer mechanisms to choose and since the FFS is usually free, it would be surprising to find disinterest. It would also be surprising not to find enthusiasm among staff since the FFS program is a source of substantial funding providing allowances, vehicles, etc. Relationship with Commercial Pesticide Suppliers Farmers at the village level had no major concerns about pesticide suppliers, neither did staff at provincial or district level. However, FAO staff expressed concerns, and have done so in the past, about suppliers wanting to sell more pesticide encouraging excessive spraying and working against IPM. It is reported that some pesticide suppliers market insecticides as "safe for natural enemies" when, in fact, they are not. Huan and Abubalar 2000 in an FAO supported study report that pesticide sellers have been negatively affected by IPM. They found some who had had to seek alternative sources of income by shifting to selling fertilizer, coal, and animals feeds to supplement reduced income from pesticides following FFS coverage in their area. However, there is a substantial interplay of incentives on this issue for the public sector, noted by Chung and Dung 1996, since the Plant Protection Departments still play some role, although a decreasing one, in pesticide supply. On the other hand, a main source of Provincial and District funding comes from the IPM donor-supported program. Environmental Externalities Discussions with farmers, government staff, and FAQ, indicate that in IPM areas, after two or three years of 1PM, improvements in frogs, snakes, fish, and snails has been observed. However, given findings discussed below that non-FFS farmers also reduced pesticide use it is difficult to attribute this to the FFS program although it might be attributed to reduced insecticide use. However, since other pesticide use, mainly fungicides, increased in many of the surveys, and since some of these other pesticides also have negative environmental impacts, it is difficult to assess net impact of IPM. In any case, there appears to have been no systematic survey of these positive environmental changes. This is surprising since there remain questions about financial sustainability and economics and any 4. Village visits, while not random, wee mostly spread across two different areas in the 4 provinces visited and the nission was fiee to talk to any frners. The nission spoke to finners who had done FFS training and frmars who had not The mison also spoke to trainers and commune officials as well as District stafl 23 Annex C external benefits not picked up in plot-level net benefit estimates would be very important in justifying any subsidy and/or continued high levels of public expenditure. Farmer Field School Costs and Extension Alternatives What does an FFS cost? How does that cost compare to alternative extension methods in Vietnam? Under the ARP, 68,500 farmers were reached through FFS at a reported cost of US$1.26 million, giving a cost of US$18 per farmer. However, this excludes some costs including a share of the technical spin-off from the on-going FAO program, a share of the government program management overheads at several levels, training costs, and other Technical Assistance. The full project supported cost of the FFS alone, therefore, is unclear in the absence of a full cost accounting but was probably closer to about US$25 to US$30 per farmer. But this has to remain a guess. It may be able to be reduced over time. The direct cost of an FFS, including only salary, travel, equipment, free farmer hand-outs and participant subsistence, which is often the cost quoted by staff in the field, comes to about US$16 per farmer for a course of 12 half days, again suggesting that there are a number of overhead costs missing from the US$18 derived from the project costs figures above. Given different cost allocations it is difficult to compare the costs with other countries such as Indonesia or the Philippines, where costs per farmer, including training of staff, MIS, and technical assistance, were US$62 and US$48, respectively. Possibly Vietnam is a bit lower than either, but this is still speculation. There is some evidence that costs in Vietnam are falling because -trained farmers, who are paid lower wages, sometimes paid partly by the trainees or the commune, are increasingly undertaking FFSs. About 20 percent of FFS expenditure over the recent 1998 to 2000 period was on FFSs where training was farmer-to-farmer. (It is worth noting that the lower level extension staff incomes are below the often-used global poverty standard of US$1 per day per capita.) Assuming that insecticide reduction was attributable to the FFS (which the surveys suggest may not be a valid assumption) and assuming that the longer term full costs, fall between $25 and $30 per farmer and assuming the need for a "booster shot" raised this to about $40, the saving of about $9 of insecticide including labor saving per season would give an adequate rate of return. But attribution of benefits remains the problem if conclusions are drawn from the available survey data. It is difficult to estimate the costs per farmer contact of alternative methods because there are no alternative methods under wide application in Vietnam other than individual visits on request, such as those undertaken by artificial insemination specialists and a lower cost audio-visual media approach tested once in the south which showed some promise. While AI is an entirely different type of service, the costs may give some indication of the cost of getting an individual extension officer to a farmer and spending a short period transferring knowledge. Al services are now privately operated and have a labor and travel cost of about $1 per visit (unsubsidized, direct charge to the farmer, semen straw cost is additional). Since Al service visits are usually not more than an hour, including travel, and assuming a farmer would need the equivalent of 12 timnes 4 hours of contact to equate to the FFS, and adjusting the travel costs, the total cost per farmer of an individual contact private extension operation would probably be around $35, but this is only the direct costs, excluding training and management overheads. As noted above, the equivalent direct cost for an FFS is about US$16. So individual extension is probably about two to three times more costly. . It is unfortunate that, after nearly ten years there appears to have been no significant program of experimentation with different levels of extension/training intensity. The FFS has evolved towards a relatively standardized (nationally) 12 session program provided by trainers. Given sustainability questions, better decisions on the difficult policy issue of depth versus coverage would have been enabled by a range of pilots with more and fewer visits, longer and shorter sessions, broader and narrower technology elements, larger and smaller groups, and larger and smaller proportions of mass 24 Annex C media. In the absence of such experimentation, decision making is now constrained by limited comparative cost/effectiveness data. One of the few alternative models with some data on outcomes, but still little on costs, is a Mekong Delta program aimed at reducing insecticide use (Huan, Mai, Escalada, and Heong 1999M). This program was presumably very much lower cost, but how much lower is unclear. Farmers reached by the mass-media campaign as well as a FFS sprayed 0.5 times, farmers reached only by the campaign sprayed 1.2 times, and those reached by neither sprayed 2.1 times. The FFS, over a period of about 5 years, had reached 4.3% of farmers and the mass media campaign over a shorter period of about 3 years had reached 92%. The coverage of technologies is substantially different between the two techniques with FFSs covering essentially all practices and the mass media campaign, in this case, covering just insecticides, the yield differences were not statistically significant (all around 6 tons). While these findings do not indicate whether, for a given extension investment, one or other or some combination would be optimal, unless the mass media campaign is extraordinarily costly by global standards, they do suggest substantial opportunities for further exploration of optimality of outcome and sustainability, particularly in the direction of mass media knowledge transfer. (Note that, in this study, with similar yields across knowledge transfer treatments, the benefit was only in the cost saving on insecticide net of changed labor costs and in any net positive externalities related to the number of sprays.) FFS Beneflts Unfortunately, the 2000 FAO study did not estimate net income changes for sample farmers in the sample provinces from before to after FFS. However, it estimated a number of input and output quantity and cost changes that can be used as approximations with some caveats regarding the averaging of averages. The average increase in paddy yield across the 7 provinces surveyed was 258 kg per hectare after farmers attended an FFS. However, the long-term paddy yield increase in Vietnam has been running at about 2.5 percent, which would give an expected seasonal increase on about a 5,500 kg per hectare base yield of 137 kg, leaving about 120 kg per hectare seasonal increment attributable to FFS-valued at about US$22. Over the same period, expenditure on pesticides plus fertilizer rose marginally by about US$7. This leaves a net gain for the FFS farmer of about US$15 per hectare per season. Ignoringfor the moment that non-FFSfarmers have also increased production, at an average farm size across the sampled provinces of 0.42 hectares (influenced considerably by the high 1.1 hectares of Dong Thap Province), this would give a seasonal net benefit per trained FFS farmer of US$6.30 and annual of about US$12. Set against an initial FFS cost of say US$30 to $40 per farmer (including "booster shots')calculated above, this would give an acceptable financial and economic retum. However, the same surveys showed that non-FFS farmers in areas where no FFSs were undertaken had also increased yields and lowered pesticide costs. Yields in the two control-group provinces (Quang Ngai and Binh Thuan) increased by an average of 115 kg per hectare for the non-FFS group, about 140kg less than the FFS group. Costs of insecticide use went down by an average $1 I per hectare, which was more than the US$7 reduction for the FFS group. Moreover, in the FAO study, a group of FFS farmers in Dong Thap Province, who had reduced their insecticide expenditure from US$4.5 per hectare before the FFS to US$0.15 the season after FFS, three years later had increased it back to US$3.0. Thus, there seems to have been some reversion which, if it correctly reflects what is happening, would reduce the benefits of the initial FFS investment. The FAO report argues that there are problems with the selection of the control groups because of the great variation in local agro- ecological conditions, because the control group may have been indirectly influenced by the FFS 5. Huan, N.H., Mai, V., Escalada M.M., and KL Heong. ( Changes in rice firme's pest ma tin the Mekong Delta, Vienm." Crop Protcion 18, p557-563, 1999 25 Annex C interventions, and because of differences in other investment influences including pesticide campaigns by commnercial suppliers. The impact evidence is therefore, at best, not conclusive. The FAO study could be interpreted as showing that impact has been small. As noted, field observation suggests quite widespread enthusiasm for both FFSs and IPM at the village level. However, as also noted, FFS is almost free given relatively low opportunity costs for labor and low or zero cost recovery. Moreover, no alternative technology transfer option is available on any scale. Unfortunately, since the FAO study, there does not appear to have been further work to improve the sampling and coverage of control groups or to test different intensity alternatives. Breadth of Coverage in FFS The mission had anticipated a relatively narrow focus on insecticide use but was impressed with the breadth of coverage of technology issues in the FFS training program. Since the training takes place every week over 12 weeks of the growing'season it covers the different stages of crop growth and the problems associated with that stage. It encourages an appropriately holistic approach to managing the paddy crop. A main aim is to give farmers a process of learning and experimenting for themselves rather than deliver sets of messages. But the mission's observation was that the reality is a mix of both. The FFS normally covers the following aspects: * land preparation and timing of land treatments * variety selection, variety pest resistance, and plant density (reduction of seed rates has been a particular area of attention) * crop establishment * fertilizer selection and application and the association between nutrient management and pests and diseases * crop husbandry, including water-level management * the use of pesticides generally and pest control techniques including reducing the serious problem of rat damage * insect identification including experimentation with "insect zoos" to understand pest/predator balance and the field ecosystem with a view to reduced insecticide application and observation techniques to assess pest and disease levels, for example, insect counting, bird behavior observation, etc. * simple experimental techniques such as comparing fertilizer rates through on-farm plot trials. Financial SustaiabNty of FFSs The current Community IPM (CIPM) program is being sustained through a mix of donor, provincial and district government, cooperative, and farmer contributions. The World Bank is no longer involved. In the CIPM program funded by FAO and Norway the share of funding by the two donors is 83 percent and the share by provincial and district government's, cooperatives, and farmers is 17 percent. Most of the direct farmer funding is for the lower cost Farmr-to-farmer Field Schools. Thus, the program remains substantially donor-fimded. As of August 2000,19 provinces had significant CIPM programs. Over eight years, the Vietnam IPM National Program has trained 515,900 farmers. If the full cost is about US$30 each this represents a little under US$2 million per year or US$ 1 million at the lower operating cost figure. The.US$2 million is about 1 percent of the total budget for agriculture and 26 Annex C about 4 percent of the recurrent expenditures6. However, the coverage of the FFS program has been limited. To date only about 4 percent of Vietnamese farming households have had a member trained in a rice IPM FFS. While 4% over 8 years seems way too little for significant impact, what would be a reasonable coverage? Would, say, directly reaching 50 percent of farmers over a 10-year period be appropriate or excessive? In the absence of more data on farmer to farmer knowledge transfer one cannot answer that question. But scenarios can be estimated. To achieve, for example, the 50% over 10 years directly through FSSs would require a tenfold increase in the full cost level of expenditure, or fivefold at the lower operating cost figure. This would call for a program that would take app.oximately 40 percent of the recurrent expenditure in the sector or 20 percent at the lower figure. This appears neither realistic nor sustainable. However, as suggested above, the issue of efficiency and sustainability cannot be resolved without a clear understanding of the knowledge transfer multiplier effect. Extent of Farmer-to-Farmer Transmission For every FFS farmer trained, how many non-FFS farmers benefit either fully or partially? This is an important question with implications for financial sustainability. The data on diffusion in Vietnam is insufficient to answer this question and it warrants study. In the Philippines, a still to be published study (by Rola et al 2002-personal communication) appears to have found no differences between neighbours of FFS graduates and farmers in villages where no one was trained in the FFS. Mission observation at the village level suggests that diffusion of knowledge related to IPM may be occurring but probably not beyond a ratio of I to I or 1 to 2 and even then probably not without a significant transmission loss of quality. Indeed, there is some indication in the field that farmers who have had FFS training may need what might be termed a "booster shot" after about three years. This would further increase the cost. However, it could be argued that, as technologies are advancing, 'booster shots" could be expected to not simply slow the degeneration of technical knowledge but top up with new, more recent, knowledge, and thus be attributable at least partly to future benefit streams. Moreover, they may be less necessary where farmers are being taught to do their own experimentation and technology development as opposed to simply receive technical messages. Level of Cost Recovery Data is limited on the level of cost recovery from farmers, but it appears to be extremely low. It is probably well below 1 percent, including both direct collection and commune funding. Over 1998 to 2000 approximately 17 percent of community IPM activities were funded by provincial and district governments, cooperatives, and farmers together, but a small share of that appears to have been contnbuted by communes or farmers. The mission did find that in some villages farmers are collecting small contnbutions for EPM activities. This has been also found in surveys. A third of a kilogram of paddy seems to be a conmon community-agreed collection level (see Hoan and Momales-Abubakar 2000), about 5 US cents per farmer. This would cover only a small fraction of the costs of FFS training, really just a token. Conclusion Based on the FAO analyzed survey data available, including the control groups, farmers who attended FSS do not appear to have increased yields by much more than fanners who did not attend nor did 6. ITigation absorbs about 50 percent of the agriculture budget, reforestation and land reclamation about 15 percent, and forty services about 10 percent About 60 percet is spent at the central level and the rest by local goveranent in the provinces. (Lcves of exendue on research a ectremely low compared to other Asian countries). 27 Annex C they reduce insecticide use by more. Therefore, the jury is still out on efficacy and efficiency of the FFS approach in Vietnam. There is what appears to be genuine enthusiasm for FFS at the village level. This may be evidence of perceived significant incremental benefits by farmers which have simply not been picked up by the surveys due to methodological problems related to the controls. It may reflect the fact that any costless (to the farmer) technology-based extension assistance is more than welcome and this is the only option being offered. It may also reflect simply the satisfaction with the social status conferred on those who attend first. There is also enthusiasm for FFS at the Provincial and District level. This may reflect a perception that FFS is an effective and efficient extension system. However, given the level of donor funding earnarked for this particular approach it is also difficult to conceive of Provinces or Districts doing anything other than receiving it warmly. Sustainability of the overall program, and thus also the portion covered by the project, is questionable so far with only about 4% of farmers attending FFS over 10 years and with little possibility of raising the level of expenditure with governments own fumds. Indeed, once donors pull out the rate of FFSs per year is likely to fall unless they are paid for by farmers. There is also uncertainty about the extent to which "booster shots" for previous FFS attendees will be needed to maintain skills in either the process of learning or the technologies themselves. While sustainability and efficiency concerns are not unrecognized by some involved in the program, a frequent question is: "But what else is there?" As noted above, the experiments evaluated by IRRI in the Mekong Delta using mass media have achieved over 90% coverage over three years at lower cost with significant impact on insecticide use7. In these experiments significant impact on insecticide use was achieved and between 1994 and 1997 15 provincial govenmments initiated their own programns along similar lines. The total costs for all the 15 provinces are reported as being US$15 1,000 to distribute 340,000 leaflets and 35,000 posters, and to organize 1,390 demonstrations and broadcast a radio drama 1,550 times. All this, it is claimed, reached to some degree about 2 million farm households. About 356,000 farmers were invited to participate in the experiment. While difficult to compare in a C:B ratio context with the FFS technique, on the face of it this appears to be a much lower cost technique with significant impact since the mean number of sprays applied declined from 3.35 per farmer in the first 40 days after spraying to 1.56 at 18 months after the campaign implementation. Moreover, a year later applications were still at the relatively low level of 1.76. While FFS protagonists would argue that such an approach does not deliver the benefits of an FFS, so far that is not strongly supported by the data. Such lower cost approaches may be an important avenue for further exploration. Resource constraints may dictate either modest technology gains and skills for 90% of farmers or substantial gains for 5% - or maybe something in between. At present, sustainability only appears to be realistically within reach if much lower costs can be achieved - which probably calls for a much greater share of mass-media- or much greater cost recovery from farmers. Unfortunately, neither the project being evaluated nor the broader donor-supported FFS/IPM program that is still on-going appears to have experimented enough to understand the costs and benefits of options at different farmer cost levels. Experiments might include several alternatives of depth versus coverage such as more or fewer visits, longer or shorter sessions, broader or narrower technology elements, larger or smaller groups, and, in particular. larger or smaller share of expenditure on mass media. Greater program experimentation would have enabled better "ranging" to approach closer to some level of optimal efficiency and sustainability. 7. see Communications And Behaviour Change In Rice Fanners Pest Management: The Case Of Usng Mass Media In Vietnam by Ecalads, Heong, Huan, and Mai. Journal of Applied Connmicadions, Vol 83, No. 1, 1999 IMAGING Report No.: 24291 Type: PPAR