Madagascar - Irrigation Rehabilitation Project

Document of The World Bank FOR OFFICLAL USE ONLY Report No. 5379-MAG MADAGASCAR IRRIGATION REHABILITATION PROJECT STAFF APPRAISAL REPORT April 3, 1985 Eastern Africa Projects Central Agriculture Division This document has a restricted distribution and may be used by recipients only - t! perfornance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS/EQUIVALENCES MONETAIRES* Currency Unit M Halagasy Francs (FMG) US$1.00 FMG 650 (at first quarter, 1985) FMG 1,000 = US$1.54 (at first quarter, 1985) WEIGHTS AND MEASURES/POIDS ET MESURES 1 are (a) = 0.0247 acres 1 hectare (ha) = 2.47 acres 1 kapoaka (285 grams)* = roughly 10 oz. 1 kilometer (km) = 0.62 miles 1 square kilometer (km2) = 0.39 square mile 1 kilogram (kg) = 2.20 pounds 1 liter (1) = 0.26 US gallon = 0.22 Imperial gallon 1 tonne (t) = 2,20b p.ounds ABBREVIATIONS AAA Parastatal Construction Company for Agriculture AHVR Aire de Mise en Valeur Rurale (Area for Rural Development) ASO Agricultural Sector Operation BTM Bankin' Ny Tantsaha Mpamokatra (Rural Development Bank of Madagascar) CCCE Caisse Centrale de Cooperation Economique (French Development Fund) CFDT Compagnie Francaise de D6vRloppement de Fibres de Textiles CIRVA Circonscription de Vulgarisation Agricole (Group of Fivaisan) COROI Comptoir de Commerce et de Representation de l'0cean Indien (State Marketing Company) DP DirectiGn de Programmation DFP Direction des Finances et du Personnel DIR Direction de l'Infrastructure Rurale (Irrigation Services) DVA Direction de la Vulgarisation Agricole (Directorate of Agricultural Extension) EDF European Development Fund, or FED Fonds Europeen de Developpement ERR Economic Rate of Return FAC Fonds d'Aide et de Cooperation (French Government Agency for-Grant Aid) FAO Food and Agriculture Organization FIDA Fonds International pour le Developpement Agricole FIFAMANOR Dairy and Agricultural Extension Projects (Projet d'Assistance Norvegienne au D&veloppement de l'Agriculture sur les Hauts-Plateaux) FNDE Fonds National de D6veloppement Economique (Investment Budget) GOPR Groupement Operational pour la Productivite Rizicole (Operational Group for Rice Productivity) * The kapoaka is a Nestle's milk tin which is commonly used to measure rice in rural areas; there are about 3.5 kapoakas of rice to the kilogram.  FOR OMCLAL USE ONLY ABBREVIAlIONS (ccat 'd) ICB InternationaL Competitive Bidding IDA ITternational Development Association IEAD International Fnd for Agricultural Development FOFIFA National Center for Applied Research oD Rural Developmnt (also CEBIRADER) KOBA State Wheat Mill LCB Local Competitive Bidding NPAEF M1instare de la Production Animale et des Enux et Fordts (Ministry of Livestock, Fisheries, and Forests) !PARA NinistAre de La Production Agricole et de la RUforue Agraire (Minlstry of Agricultural Production and Agrarian Reform) KwP Ministtre des Travaux Publics (Ministry of PubLic Works) ODR Opfwation, de Development Rizicole (Project to Develop Rice Production 0 and N Operatiot and Maintenance SAMAGOKY SocihtS pour l'Ak &ageuent et la Hise en Valeur de la Vallfie du Bas-Nangoky (Society for the Development of the Lover Mangoky River Valley) T and V Training and Visit Method of Extension NEC Water manaemet Comm4ttee OVERNMONT ADKMIISTRATION Fokontany . village Firaisau-pokontanzy group of Fokontany (or Firaisana) - (former canton) Fivondronaw-pokontany - group of Firaisana (or Fivondronans) (former sub-prefecture) Faritany group of Fivondronana (former province) FISCAL YE Government of Madagascar January. 1 - December 31 SAMANGORY January 1 - December 31 PADDY: RICE CONVERSION One ton of paddy gives about 670 kg rice. One ton of rice is derived from about 1.5 tons of paddy. I This document has a restriced distribution and may be used by recipients only in the perfonnhnce of | Lheir of iia duties Its contents may not otherwise be disclosed without World Bank authorization.  - i - MADAGASCAR IRRIGATION REHABILITATION PROJECT STAFF APPRAISAL REPORT TABLE OF CONTENTS Page No. I. *BACKGROUND A. Project Background ............................ 1 B. The Agricultural Sector ...................... I C. Bank Group Operations in Agriculture .......... 5 II. THE IRRIGATION SUBSECTOR A. Irrigation Practices *......................... 7 B. Physical Deterioration of Schemes ............. 8 C. Institutional Deterioration of Schemes ........ 9 D. Subsector Institutions .**.....**........*. 10 E. Medium Schemes: Maintenance, Operation, Finances, Land Tenure .......... *. ....... 12 F. The Mangoky Scheme .... o....00 ..* .............. 13 G. Subsector Issues ..... *..... ...*. ........... 14 III. THE PROJECT A. Rational for Bank Group Involvement ........... 15 B. General Description .*..... ... .... ............ 16 C. Detailed Description .* ............. ...... 17 Management Suppor: for MPARA Civil Works: Medium Schemes Rehabilitation of the Mangoky Scheme Participation of Co-financiers D. Project Costs a ....................... 23 E. Proposed Financing ....................... . 25 F. Procurement and Disbursement Profile .......... 27 G. Budgetting, Accounting and Disbursement Procedures 30 H. Environmental and Health Impact .................. 32 IV PROJECT IMPLEMENTATION A. The Sequence of Scheme Rehabilitation ............ 33 B. Ad hoc Coordinating Committee ............ r.. *.... 36 C. Scheme Finances, Operations and Maintenance ..*... 37 D. Organization of Work Planning ................... 38 E. Training and Manpower Development 39 F. Monitoring, Evaluation and Reports ............... 41  - li - Page No. V. ECONOMIC ANALYSIS A. Benefits and Rates of Return .... ............ 42 B. Risks and Sensivity Analysis ................ 44 C. Proposed User Charge for 0 and N and Efficiency 45 VI. AGREEMENTS REACHED AND RECOMMENDATION Tables in the Text Table I Summary Project Costs 24 Table 2 Proposed Project Financing 26 Table 3 Procurement Project Financing 28 Table 4 Economic Analysis: Economic Rates of 46 Returns, Basic Data, and Assumptions by Irrigation Scheme, First Tranche  ~IP~ TABLES. AN, aiAEaS AND MAPS Detailed Project Coscs Page No. Table I Sc Rebabilitation - SovLina 50 Tble 2 Schme Rehabilitation - BelDty 52 Tab]l 3 Schefm Rehebiticm- Behara 54 Ta 4 Sc Rhbelitation - HagkY 56 T3ble 5 AU Sdhm RehabiLitation 58 Table 6 ILnstitutim alildfrg. MPARA. 59 Table 7 Projert Coaenms by Year. Incuding 62 Table 8 Suz5y =oim=s by Project Corent 63 F=cmxc Aktalysis Table 9 Clculation - the Emnac Price of Pady at the FamrGate 64 Table 10 Faxm Budgets Per Ha of Irrigaced Paddy Productio 65 Table 1 1 Aloptlcin-es. Yields, Benefits and Costs, per Fkanomic 66 Hectire, *r E3dsrIrg aid Future Tecdial Padckges Table 1.2 E Rate of Petumn Cacul.alons 67 Table 13 Saaaoky PRehabitacica, Ecom2ic Cost akd Benefits 68 Tahle 14 Samngk:y Projected Average Ha and Yields by Crop ard 69 lype of zltivatio in 1986 arid Succeeding Years Table 15 Sanang*ky- Projection of FinawJAl Rehabilitation (1984-88) 15A: Projected I;ncome Lacement. Lgricultran 70 Production Deepart 15B: Consolidated Projected Ircwe Statezmet 71 X: Projected Profit ad Loss Statement 72 15D: Projected Balance Sheet 73 Table 16 Samazpky: Ix and ExperditLre Aunc 1982-1983 74 Table 17 Samagoky: Balaoce Sheet, at Deceder 31, 1983 75 Other Tables Table 18 Basic Data mn 27 TrrIgation Sches 77 Table 19 Projected Cash Fno of Goverrxent 78 Table 20 Estimces Schedule of Dishrwcs 79 Table 21 Beterioration in Paddy Oltivaticn, Soaina, Belamoty, Behara 81 ke I Materials Aailable in the Project File 82 kArm 2 Irrigation Legslation 84 Oarts ard Maps Chart 1 Project Organization 86 Chart 2 Minsrisy of ArIcultural Productica aid Agrarian Reform 87 (Orgawization Chart) Chart 3 Directorate of Rural Infrastructure (Orgamization Chart) 88 Chart 4 Project Iqplentation Schadule 89 Map I Ma r,Lton of rium Schemes (IBBD 18285) Map 2 Map of Behara, BeIlty, ard Soarina Schemes (ID 18286) Map 3 Map of Saragoky (MR 18387)  - iv - There is a separate Implementation Volume (EAPCA) with the following content: 1. Detailed Economic Analysis 2. Co-financiers Aide-Memoires 3. Euroconsult, Rapport de Syntheses, Projet de rehabilitation des petits perimAtres  v MADAGASCAR IRRIGATION REHABILITATION PROJECT Credit and Project Summary Borrower: Democratic Republic of Madagascar Amount: SDR 12.6 milliou (US$12 milliom equivalent) Terms: Staudard IDA terms Project Descriptiou: (i) Objectives: The Project would help develop a systematic approach to rehabilitaciou of 116 existing medium scale irrigation schemes (142,000 ha under canals) and is the first phase of a long-term program of their rehabilitation. A key project objective is to develop new institutions for self-admiaistration and to assise the farmers to provide self-financing to ensure adequate future maintenance of the schemes rehabilitated. tu addicion one Large scale irrigatiou scheme i.-uld be rehabilitated and management of the scheme improved. Components: a) strengthening of existing institutions to select the priority irrigation schemes to be rehabilitated; b) rehabilitation in the first phase of 16 medium irrigation schemes covering about 17,100 ha.; c) rehabilitation of the large scale Samangoky cotton and rice scheme. - (ii) Benefits: By increasing rice production 32,000 tons per year the project will reduce Madagascar's dependence on imported rice and enhance income for a group of over 2X of the farm population whose income per capita is less than half the country average. (iii) Risks: Project sustainability depeads oa water users effectively operating and maintaining the irrigation schemes. This risk is being reduced by requiring subscantial wacer user pareicipation at several stages: user review of engineering design; user commitment to self-finance; user prepayment of part of maintenance costs before rehabilitation is completed; TA to support these activities as well as formation of water user associations; and developme.t of individual scheme manuals for operation a-d maintenance. Part of :rne benefics are dependent on the development and diffusion of higher yielding new variecies and of increased double cropping. Nevertheless, selec:ion of schemes for rehabilitation would be such that even if the benefits from better varieties and double cropping are not realized, the economic rate of return would exceed 14%.  - V1 E.stimacd Proj ecr Costs percen. of - USS millin aforeip % of cotal local fore toral eharge bselie csrts gD Supprt to MtIz.rzy of igriculruwe 0.9 3.0 3.9 78 18 Sch1. Rehabita 4.3 8.9 13.2 68 60 SM .R R.eabilia 3.0 1.8 4.8 36 Z2 Total Basel'io Coss: 8.2 13.7 21.9 63 100 Physical cirsgrries 0.9 1.8 2.7 63 13 Price ropn^i 2.3 2.7 5.0 53 22 Tbtal Projecc Goscs 11.4 18.2 29.6 61 135 FIotmcfrmg Plan. sch total mA, Cr EDF FAC M faaHrs US$ (million) 29.6 12.0 6.0 4.0 0.1 7.1* 0.4 Perceczage 100 41 20 14 - 24 1 Iwesed tW!9, resultig frem project eqm&iures are estiuted at 54.7 mfillfIo. Estim Dirrseis us$ mi 11 nr .Pl o eulative 1986 1.4 1.4 1987 2.5 3.9 1988 3.2 7.1 1989 2.8 9.9 1990 1.4 11.3 1991 0.7 12.0 F .cc Rate of Return: 24% Staff Appraisal Reprt: ps: MM 18285 18286 18378  MADAGASCAR IRRIGATION REHABILITATION PROJECT STAFF APPRAISAL REPORT I. BACKGROUND A. Project Background 1.01 For more than a decade the Government of Madagascar and the Bank have discussed a project to rehabilitate the medium size irrigation schemes, which are an important part of Madagascar's Irrigation system. Because of the tumultuous political and ecoromic circumstances in Madagascar during this period, preparation bas been slow and sporadic. Agreement on basic project objectives was reached in 1979. By 1982 a preparation report was completed, which concentrated on rehabilitating five medium irrigation schemes as a pilot endeavor. A pre-appraisal mission, including representatives of the French 'Caisse Centrale de Cooperation Economique", visited Madagascar in November 1982 and reco_mended: (i) cost reductions of about 25% for the five by reducing low priority irrigation investments (canal linings, water inlets, weir designs); (ii) that the project's complex organization required further study; and (iii) that an inventory of the medium size schemes should be carried out to determine their priority for future rehabilitation. The Government agreed to an additional study to cover the above items. The study was carried out and the resulting report was submitted in January 1984. It became the basic document supporting the appraisal mission of April 1984, which was led by A. de Largentaye, and included J. Meerman, J. Weijenberg, J. Porter and H. Boumendil (consultant). Representatives of the Caisse Centrale de Cooperation Economique (CCCE) and of the European Development Fund (EDF) accompanied the mission and participated in discussions. This was followed by two post appraisal missions led by J. Meerman in July and October, 1984, again with participation of the CCCE and the EDF. B. The Agricultural Sector General 1.02 Madagascar, an island more than 1,500 km long, is slightly larger than France, although its climate, topography and soils are all far more diverse than those of France. Rainfall, for example, varies from that of tropical wet-forest to near desert conditions. With a population of 10 million, the country is sparsely populated but there is great pressure on productive land in some regions. About half the population is below the  -2- poverty line. Rice, most of which is produted under irrigation, is the most important crop. (In 1983 total production was 1.4 million tons, about 140 kg per capita.) It is followed by cassava, most of which is consumed oa the farm. For 1980-82, coffee exports contributed almost half of Madagascar's total export earnings. Cloves and vanilla accounted for most of the remainder. Other important cash crops, not currently exported, include sugar, cotton and tobacco. The cattle herd, of about 10 million head, is almost entirely owned by herders - accounting for a large part of the popuLation - who employ traditional, extensive husbandry systems. As a result, productivity and off-take are low. Beef exports are a fraction of former levels, notwithstanding increased foreign demand. Noonetheless there is excellent potential for augmenting production through improved herd management, and for iacreasing output of other forms of livestock, notably pigs and poultry. 1.03 Most of the estimated 1.5 million farm families--about 80% of the population-are small holders engaged in subsistence farming. A traditional land tenure system prevails with individual, heritable rights to cultivated land and collective rights to pasture. Smallholder farming techniques range from irrigated rice cultivation involving elaborate terracing and partial water control to slz3h and burn cultivation and semi-nomadic cattle raising. There is some large-scale mechanized farming, usually oa Government-owned farms. During the 1970s, the Government's role expanded substantially; most noteworthy were the nationalization of larkecing of foodcrops (especially rice), of the sugar industry, of input supply, of meat processing, and of export marketing. Irrigation Subsector 1.04 Past agricuLtural development has concentrated on irrigation with heavy investment by both Government and external agencies. Out of a total cultivated area of roughly three million ha, official irrigation systems serve about 250,000 ha, nearly all of it to produce paddy. In addition, about 700,000 ha of traditional paddy fields have some form of water coatrol. At present, the key issue in this subsector is deterioration: Madagascar's decaying irrigation infrastructure is yielding far less than it should, in part as a result of failure to maintain the schemes which in turn results from the deterioracion in management, and the lack of budget funds and foreign exchange for inputs and equipmeat. Although the traditional paddy fields have fewer problems, their yields have dropped because of reduced supply of inputs, and negative price and market incentives. Looking to the far future, it is probable that further development of irrigation, drainage, and flood control, will be concentrated on relatively large schemes siace most areas suitable for small-scale farming (terraced rice fields in small valley boctoms) have already been developed. The scope for further devel_pment, however, will remaia uncertain untit basic cost and benefit data are generated on potential areas for development.  -3 - 1.05 With assistance from cofinanciers, the Ministry of Agriculture (MPARA) has begun the rehabilitation of the larger-scale networks, which once produced most of the marketed rice and which suffer the most acute management problems.. The proposed Project would begin the rehabilitation of the medium-sized schemes by improving management arrangements, improving operation and maintenance, and by carrying out rehabilitation. These rehabilitation efforts are likely to occupy the Government for many years. It would also provide assistance to one large scale network, SAMANGOKY which has been supported by IDA but stands in urgent need of rehabilitation investments and a new approach to management. Recent Sector Performance 1.06 In production of the staple, rice - as with other crops - the Government's marketing monopoly and price controls from the early 1970s to 1983 led to wholesale disruption of marketing. By the early 1980s, official prices were so low and official marketing so disorganized that production stagnated and much rice entering into trade was sold on the parallel market. Simultaneously, lack of modern inputs and credit, deteriorating extension service, and poor irrigation scheme management and maintenance, left farmers few means to increase production, even if price and market incentives had been adequate. Hence, notwithstanding Government allocation of some 40% of the public investment budget for rice production in recent years, production stagnated. Paddy production in Madagascar, which grew 5% per year in the 1960s and increased from 1.2 million tons in 1960 to almost 2.0 million tons in 1970, did not grow thereafter, in spite of the growth of population and the labor force. Madagascar, a net exporter of rice until the early 1970s, thereafter imported increasing quantities, reaching a peak of 350,000 toos in 1982, over 20% of total consumption. Since then, reduced rice imports reflect changed policy but also an overwhelming foreign exchange shortage: 183,000 tons were imported in 1983, and about 110,000 tons were imported in 1984. New Agricultural Strategy 1.07 By 1982, the Government accepted as a fact the failure of agricultural development policies and the need for basic policy change. For 1983-85, it has set the restoration of production of key crops to levels reached in past years as an overall goal. For rice, the goal is a return to self-sufficiency. For the past two years the Government has been discussing with the Bank, and other aid agencies, how to achieve these. The main elements of Government's new strategy, as accepted by the donor community, were announced at the April, 1983 meeting of the Consultative Group for Madagascar: Short Term - concentration on rehabilitation of agro-industrial enterprises and large-scale production units, including the very large parastacaL irrigation schemes;  I - - emphasis squarely on rehabilitation through existing agents, including private enterprises, with priority to rice, cotton, coffee, and livestock; - regular price reviews aad increases to ensure adequate producer incentives, particularly of rice, cotton, coffee and sugar; - rehabilitation of rural infrastructure, particularly rural access and feeder roads. Longer Term - concentration oG smallholders and other privateLy owxed production units; corollary, sharp curtailment of investmnet in large scale, and particularly Governmeat-owned and operated, farming ventures; - a development strategy promotirg many crops and taking into accouat regional variations; encouragement of diversification of production systems; - conversion -of unsuccessful and unprofitable parastatal ventures through sales to private operators and use of management contracts; - reduction in state marketing; increased reliance on commercial companies; and - improvemenLt of agricultural extension in key production areas and reestablishment of agricultural research. 1.08 By 1984 there was some progress in implementating the overall strategy. The prices of rice and beef were alLowed co rise to market clearing levels; and free trade in these foods was re-introduced. Sigmificant increases in producer prices and improvements ia price setting mechanisms were made for cotton. Studies were completed on rice and export crops to provide the basis for :heir pricing and marketing reform. Input supplies have been imported with external aid. Input subsidies are being phased out, and fertilizer distrLbution by commercial firms has re-started. Veterinary product supply has improved, again with external aid. A review of ways to revive extension services is underway and under Bank supervision IFAD is financing a Training and Visit (T and V) extension program. A detailed review of the agriculture investment program has been carried out. New legislation for irrigation water charges has been enacted. In 1983 the Ministry of Agriculture reduced staff by about 3,000.  Sector Lending Strategies 1.09 The Bank Group lending strategy for agriculture and specifically for irrigation is consistent with the Government strategy discussed in para 1.07. The Bank Group accepts that Madagascar has a strong comparative advantage - at realistic exchange rates - in the production of many agricultural commodities including irrigated rice. Consequently the Bank Group supports the Government goal of self-sufficiency in rice production to be achieved within several years. To exploit this comparative advantage requires policy evolution, again along the lines set out above. The emphasis on rehabilitation of both capital and institutions, including the rehabilitation of the Ministry of Agriculture (MPARA) is highly desirable. In the context of such rehabilitation, there is the need for a new institutional division of labor in which the burden on Government is much reduced through increased reliance on suallholders as the group which will realize production increases; increased reliance on private entities to provide inputs and to market output; and increased reliance on market clearing decentralized prices to guide the movement of resources between sectors and within agriculture itself. The proposed project would help implement this strategy. It is based on increasing smallholder production at the medium sized schemes through their permanent rehabilitation. C. Bank Group Operations in Agriculture 1.10 The Bank Group has supporced fourteen agricultural projects in X2dagascar. Five projects have been completed: The Lac Alaotra Irrigation Project (US$5.0 million), the Morondava Irrigation Project (US$15.3 million), the Beef Cattle Development Project (US$2.8 million), the First Forestry Project (US$13.0 million) and the First Village Livestock and Rural revelopment Project (US$9.6 million). Eight are under implementation: the Mangoky Agricultural Development Project (USS12.0 million); the Agricultural Credit Project (USSI0.0 million); a Technical Assistance Project of studies to improve irrigation on the Plain of Antananarivo (US$2.3 million); a Second Forestry Project (US$20.0 milLion); an Agricultural Institutions Technical Assistance Project (US$5.7 million); a Second Village Livestock Project (US$15.0 million); the Lac Alaotra Rice Intensification Project (USS18.0 million) and a Cotton Development Project ($17.8 million). In addition, the Bank has appraised and is supervising the IFAD-financed Highland Rice Project (US$30.0 million) to increase rice production by providing agricultural inputs and by reorganizing and strengthening agricultural extension services using the Training and Visit system of extension. 1.11 Four agriculture projects financed by the Bank Group have been completed and audited by the Operations Evaluation Department: the Lac Alaotra Irrigation Project, the Beef Cattle Development Project, the Morondava Irrigation and Rural Development Project, and the Village Livestock and Rural Development Project. The Audit Report and a recent Impact Evaluation Report on the Lac Alaotra Irrigation Project criticized both poor project design and supervision, and insufficient attention to socio-economic and institut!tnrta aspects of the project. These lessons have been taken into account in the project design of the Lac Alaotra Rice  -6- Intensification Project, in which great emphasis is given to institution-building, to farmer involvement, to producer incentives and to self-management to reduce the burden of recurrent costs on the Government. The lesson of the Beef Cattle Project where large-scale cattle ranching failed dismally, have long since been reflected in our approach to agriculture operations in Madagascar. The PPARs for the Morondava Irrigation Project and the Village Livestock Project emphasize the enormous project implementation problems during the tumultuous 1970s. The Morondava project was a profoundly disappointing experience with cost overruns and poor management the most serious problem. The Village Livestock Project fared much better and established a sound basis for veterinary and livestock services in a key cattle producing region. 1.12 A_sociated with the recent policy and institutional changes has been a resurgence of the lending program of which a key element is the Agricultural Sector Operation (ASO) which is now nearing appraisal. It is designed to support continued policy changes which would help lead to an agricultural sector characterized by market-clearing pricing and decentralized, competitive marketing of inputs and production. These changes would contribute to the s' tess of the proposed Project, since project success requires substantial increase in produCtion by smallholders in response to increased market opportunities on the one hand; and on the other hand to increased availability of inputs. To support the many policy changes, the ASO would finance import of the agricultural inputs and equipment most needed for restoration of agricultural production, of both rain-fed and irrigation systems. 1.13 Agriculture's renaissance and its future development requires initiative, support and management from effective agricultural ministries which in turn require stability, good internal organization, good leadership, and. many skilled and dedicated people. The World Bank Group has been supporting such institutional development thorough the First Agricultural Institution Project. Much of the analytical underpinning for the recent Ministerial reorganization in 1982, for example, was promoted and financed through that project. Analysis carried out under the project has also helped the Ministry in its decisions on certain policy reforms, in the rice sector for example (para 1.08). The project has also financed a study of parastaEal performance in agriculture leading to a number of recommendations which are beginning to be implemented: some management audits and reorganizations are underway. It has financed a study of issues and constraints in agricultural research, and recommended several changes in management anc organization. This process of institutional development will be reinforced through the Second Institutional Development Technical Assistance Project now in preparation. In addition to continuing the ongoing work, through the project, Government will seek to (a) come to grips with the many problems of reforming the parastatal enterprises; (b) to make agricultural research address the country's agricultural production problems; (c) improve procurement and disbursement procedures; (d) develop a capacity for carrying out the agro-economic analysis MPARA needs for operations and for investment planning to promote agricultural development.  -7- It. THE IRRIGATION SUBSECTOR A. Irrigation Practices 2.01 Irrigation is practiced throughout Madagascar on about a third of the island's cultivated land. Most irrigation is carried out by small-holders on schemes which vary in size from a few ha to several thousand, but there are also several large parastatal schemes using modern techniques of water control and maintenance. The kinds of irrigation which are feasible vary according to the three main ecological regions of the country: the Highlands, the west and the narrow east coast. Because of the high altitude, in the Highlands the dry season, or austral winter, is cooL which greatly reduces the yields of rice and many other crops grown at this time. Hence, in the Highlands most agricultural production occurs during the rainy season. The agricultural population of the Highlands are highly skilled traditional cultivators. Population growth on the limited cultivable land available has led to substantial emigration to the western part of the Highlands where land of somewhat lower quality is available. Soavina-a scheme completed in 1962 but now to be rehabilitated under the first tranche of the project-is one such westward settlement, composed principally of immigrant rice growers from the densely populated faritany of Ambositra. The second region, the west, is hot; rainfall is usually under 700 mm annually; and the dry season is very long. There, farming conditions are excellent in the river valleys, the -baiboho, where much Elood recession cultivation is practiced. There are also several large irrigation schemes in the west. Livestock rearing in this area is widely practiced by populations which are somewhat nomadic. The Androy of Behara, another scheme to be rehabilitated under the first tranche, are an example. Many such cattle owners, however, are slowly becoming sedentary farmers. The settlers in the Onilahy valley, which includes Belamoty, a third rehabilitation scheme also in the west, are an example of livestock owners who have become cultivators within the last century. In the west, the main irrigated crops are rice and cotton, while pois .du cap (butter beans) are produced for export. The third region, the warm and wet humid east coast, has a highly productive agriculture. Westbound winds drop much of their moisture in this area, before they reach the Highlands. Annual rainfall in excess of two meters is the rule. Certain areas, such as the ports of Toamasina and Moroantsetra, receive three meters or more. Agriculture is relatively intensive, and most export crops, cloves, vanilla, coffee, are produced in this area. Rice irrigation is practiced in river bottoms; rainfed rice is also common. 2.02 Throughout the island, family farms are small, nearly always less than ten hectares and frequently less than a hectare, including both rice paddies and the -tanety- or rain-fed fields. Animal traction is common. - Rapid population growth has contributed to the small size of individual farm plots, especiallv in the Highlands where population has been sedentary for centuries-. In tile recently settled areas of the west, plots and farms tend to be larger.  -8- 2.03 Rice is by far the most important irrigated crop, although there are small areas of irrigated cotton and sugar cane. Abouc two-thirds of the population cultivates at least some paddy. Traditional yields of paddy on irrigated land are generally close to two tons per hectare in the warm (rainy) season, the so called second" seasoa because it begins late in the caleadar year, and about 1.5 tons in the dry season ("first" season). As noted, in the Highlands, low temperatures in the dry season make rice-cultivation a risky and low yielding crop. In. the lower parts of the Soavina scheme, at an altitude of 1100 m, some double cropping is practiced aLthough the length of crop cycles may require laad preparation of the second season before completion of the harvest of the first season. Irrigated dry season yields in Soavina are typically 1.3 t of paddy per hectare. L/ On average rainfed rice yields about 1 t paddy per hecrare, although yields may be subscantially higher if such rice is grown in the semi-irrigated conditions of river bottoms or in the water-retainirg paddies of the deteriorated schemes. Paddy yields on the large parastatal schemes (para 2.06) probably exceed 2.5 c on average, and may exceed 3 t. This is however much below their potectial because of substantial deterioration of these schemes. B.. Physical Decerioration of Schemes 2.04 The deterioration of the schemes, both large and small, may take the form of a catastrophic event. The pressures--literally--on che various irrigation works are heaviest during the Periods of maximum river flow during the high rainy season. Followiag storms, the dikes, canals, intake works, any of these, may suddenly collapse because of the water pressure, with consequences such as iaundation, croP land covered with sand, and greatly reduced flow-through into the irrigatioa canals. The non-parastacal schemes have largely relied on their own devices for repairs after such events. The repairs have oftea been inadequate so that the ability to withstand future shocks is dirairjshed. Hence with the pasSage of time the probability of disaster and che economic loss associated With any given period of high water-flow has increased. There is also 510w deterioration which gradually reduces the hectares which can be irrigated in any scheme: canals gradually silt up with sand carried into the System; they also lose water at leaky seams and breaks along their courses. As a consequence the capacity for carrying wacer by any given canal system has diminished. Finally, the drainage systems - drains and outlets of various kinds - have deteriorated so that land has become marshy and unsuitable for cultivation or may be inundated at flood-crests, when the river backs into the drainage svstem or breaks through proteccive dikes. 2.05 In the three schemes of che Project's first tranche (para 2.01), the combined effect of these various kinds of deterioration has reduced paddy irrigation to nearly half of design capacity. In Soavina, of the 1,200 hectares served bv canals, about 200 hectares cannot now be used in the wet season because of regular back-up inundationr by the river. Another 100 ha are lost because they have been covered with river sand. Ia 1/ One ton of paddy equals about .65 cons of rice.  -9- Belamoty and in Behara, sand in the canal-system has reduced their capacity for carrying water so that much of the down-stream land ia the schemes does not receive enough water to permit paddy irrigation. For the three only about 60Z of the areas served by canals can now be irrigaced (Annex Table 21). C. Institutional Deterioration of Schemes 2.06 The Government of Madagascar has classified irrigation schemes into four types: (a) Family rice fields developed and maintained by farmers themselves, may cover as much as 500,000 ha and range in size from one are to perhaps 5 ha. They involve terracing of valley bottoms, diverting water from streams and simple storage dams. (b) Small schemes, developed through local initiative with and without Government support. Their size varies from 5 ha to 1,000 ha; the total is about 300,000 ha. In many cases these traditional systems are operated by local commuaities, and require few nom-labor inputs for their maintenance. (C) Medium schemes were largely built with financing from FNDE (Fonds National de Developpement Economique, that is the investment budget), although frequently FNDE resources had their origia in foreign aid. The surplus rice produced in such schemes is usually sold at local markets. There are about 116 such schemes with a total area under canals of about 142,000 ha (an average of 1,224 ha); in part because of deterioratioa of the irrigation works, actual amount irrigated is about 79,000 ha. In theory at least, the Ministry of Agriculture has primary responsibility for running these schemes with some support from the local authorities. The proposed Project would coacentrate on rehabilitating about 16 of these schemes. (d) Large schemes were designed and constructed with external financing, and are operated and maintained by parastatal authorities reporting to the Ministry of Agriculture (MPARA). Land irrigated is about 80,000 ha and most of che surplus rice production is consumed in the large cities. The most importaut schemes are SOMALAC (Lac Alaotra area), FIFABE (Marovoay area), SAMANGOKY (Maagoky scheme), SODEMO (Morondava scheme) and the Andapa Basin Authority. 2.07 The large and medium schemes, (c) and (d) above, form the official system. Their management, maintenance, operation, repair, and finances are responsibilities of the Ministry of Agriculture (MPARA). In the case of the large schemes, SAMANGOKY, SOMALAC and so forth, these responsibilities rest with the associated parastacals. Low output prices, poor Labor discipline, inflated cost structure, including substantial  - 10 - overmanning, inadequate maintenance, weak cost recovery and shortage of experienced middle management have brought extremely high financial losses, sharp decreases in output and grave deterioration in the physical works. The main large irrigation schemes are receiving external assistance for physical rehabilitation and management support. IDA is currently active in two of these: the Lac Alaotra Rice Intensification Project which includes substantial civil works to reconstruct irrigation and drainage structures. The rehabilitation work at SAMANGOKY is described in para 2.16. 2.08 The problems in the medium schemes are-similar, if less spectacular, to those of the large schemes. TIeir operation, maintenance and repairs have been the responsibility of the XPARA acting through its Directorate of Rural Infrastructure (DIR). Maintenance and repairs of these schemes have been defective in part because of inadequate financial allocations. But deterioration has increased further because of the failure to encourage cultivators to carry out that maintenance of which they are capable. The traditional irrigation schemes usually are small and are managed by the farmers themselves. They do not have the management problems of the larger schemes. Since the traditional schemes have substantial potential for increasing production, they have also attracted Government and external support. The resultiag projects have been limited to the improvement of water inlet structures and, in certain cases, drainage outlets, whose construction is beyond the technical ability of the farmers. At least 2,000 traditional schemes have been improved in this manner, affecting more than 80,000 ha of irrigated land. The FED has been the largest foreign donor in chis area. D. Subsector Institutions 2.09 GovernmeLt reorganized the Ministry of Agriculture (MPARA) in 1982. Four of the five resulting Directorates would be heavily involved in the project and are therefore described here. The Directorate of Rural Infrastructure (DIR), the successor to Genie Rural, has three central services (Chart 2). Its Service de l'Irrigation is in charge of planning and implementing engineering works in all Government-managed schemes. The Service includes a "division des &tudes(design division), a rehabilitation division and an irrigation infrastructure division. The rehabilitacion division is in charge of supervising rehabilitation work, whereas the irrigation infrastructure division, with the assistance of three expatriates, is in charge of supervising studies and works on new schemes. Oa large schemes, managed by parastatals, the DIR is responsible for supervising feasibility and other studies, inviting and evaluating tenders and preparing contracts, and supervising irrigation work. The Directorate icself has usually prepared the detailed designs and supervised the contractors for medium schemes. It is clear that the DIR has much experience and competence in managing firms to carry out the engineering aspects of scheme rehabilitation (design work, execution, supervision). Notwithstanding the existence of the Rehabilitation Division, following the transfer of most construction and land preparation equipmeat of the mobile brigades of Genie Rural to a parastatal, the Ministry virtually has ceased  - 11 - to carry out scheme repair and maintenance. The Service de Gestion des Reseaux Hydrauliques (Irrigation Scheme Management Service) is responsible for operation, maintenance and development of the medium irrigation schemes. Its functions include managing the DIR's irrigation personnel at the medium schemes, coordination with other agricultural services, and assistance in the yet to be established scheme-level water management committees (para. 2.14). In 1983 the budget for operation and management of the service was FMG 600 million ($US1.1 million), almost all of which went to pay staff. The Service de l'Ame'nagement Rural operates the Toamasina grain silo, and several edible oil mills. Each of the above services also has corresponding provincial level services. In turn the six regional services, one in each faritany, are responsible for the 17 administrative circonscriptions'. The Directorate's staff of 1,400 includes 33 engineers, many trained in Strasbourg, France, as well as three French expatriate engineers. Many senior level staff are recent graduates who lack experience and need professional guidance. The DIR's experience in setting up and supervising civil works, and its grass roots organisation would be important positive elements in implementing the proposed Project. 2.10 The Directorate for Agricultural Extension (Direction de la Vulgarisation Agricole, DVA) has a total staff of about 4000 (ldd 1984) and a budget (in 1983) of FMG 927 million. At the national level, the Directorate has three services: Plant Protection, Support to Extension, and Coordination with Research. As with the DIR, these services are also found at the faritany levels. At the next lower level there are the "circonscriptions de vulgarisation agricole" (CIRVA). Below the CIRVAs are extension zones (corresponding approximately to fivondronana) and finally the extension cells, with each cell having - ideally - five extension agents. Usually one or two cells work on a medium irrigation scheme. There are about twenty agronomists based in Antananarivo and about fifty in the provincial services and CIRVA. Zone chiefs often have six years of secondary school and extension agents usually have three. With over 80% of the DVA budget used to pay salaries, the extension service lacks funds for transport and field activities. Hence the field services are out of touch with the central services and their activity is neither monitored nor evaluated. Applied research is very limited, and in the last decade, DVA has developed few, if any, new extension themes. Lack of agricultural inputs, of improved seeds and of agricultural credit have compounded the problems of the extension services. The IFAD Highland Rice Project, active in two CIRVA in the Highlands (Antsirabe and Ambositra), is providing operating and transport support and a new approach (training and visit) for extension work. Under the proposed project, the DVA would work intensely to increase yields on schemes to be rehabilitated. Hence the proposed project would provide considerable technical assistance to the DVA. 2.11 The Programming Directorate (DP) prepares the investment programs in agriculture. It has received IDA support under the First Agricultural Institutiops Technical Assistance Project which provides technical a3sistance for training in this task. Technical assistance is also being  - 12 - provided to improve the Directorate's monitoring and evaluation of projects. The statiscical service within the Directorate of Programming is charged with processing agricultural data generated by the agricultural extension service. The DP Is to be the coordinator for the proposed project (paras 3.05-3.08). 2.12 MPARA's recencly created Directorate for Finance and Personnel 's responsible for establishing the Ministry's budget, for administering and coatrolling its execution, and for managing all external financial resources supporting agriculture. (This would include the proposed project.) Under the First Agricultural Institutions Technical Assistance Project, the Directorate also is receiving technical assistance to carry out these tasks. E. Medium Schemes: Maintenance, Operations, Finances, Land Tenure 2.13 Most medium schemes were built between 1945 and 1970, typically during the colonial period (Annex Table 18.) The approach to cheir operation has been a ihighly centralized one in which the cultivators on the schemes were initially considered as tenants or wards while the mobile construction brigades of the Ministry of Agriculture's Genie Rural were to perform required repairs and maintenance. The effectiveness of these brigades dropped dramatically during the 1970's, and they have been phased out. In addition a "chef de reseau", was posted to each scheme --and still is-- to provide scheme management. Scheme maintenance has been very defective and largely limited to the cleaning of the primary and secondary canals by the water-users. The irrigation works (water intakes, desilting basins, weirs, dikes, drains, overpasses, access tracks) have coasequently deteriorated. When damaged, they have frequently not been repaired, although in most cases had funds been available, purchase of equipment and supplies would have been possible. Cost recovery has not been practiced on the schemes due in part to past Government policy and the power of local interest groups, but perhaps more recently also because -of Government inability to collect the money. The problems of handling receipts through the public fiaance and accounting systems are grave and the tax collection mechanisms ia the rural areas are very weak: the national property taxes once collected at the local level have largely disappeared, as well as the tax rolls according to which they were levied. 2.14 By 1981, the widespread deterioration of the medium schemes had touched off a discussion within Government towards developing a viable alternative approach to scheme operation and maintenance (O and M). The result was the body of legislation of 1981-1982 (Annex 2). This legislation foresees establishment of scheme level Water Management Committees (WMC) consisting of Government representatives, including the DIR's "chef de reseau' and representacives elected by the users. Ideally, the WMC takes decisions based on discussions in which user-representatives would voice their demands; 0 and M proposals are presented by the -chef de reseau'; users receive guidance for management of the tertiary and  S - 13- quarternary networks. A dina - a traditional solemn contract - would spell out the mutual obligations of the Government and the WMC. In brief, WMC are to manage water-distribution, and provide for adequate scheme operation, maintenance and repairs. To make this possible, the legislation stipulates that user charges for water are to be collected under Government authority and would be administered by the WMC under the direction of the on-site -chef de r&seau" but Government would also substantially subsidize O and M. Paradoxically, the laws also include provision for collection of charges to cover some part of the costs of scheme rehabilitation. IDA's position is that both the proposed subsidy and the proposed development tax are undesirable (paras. 4.07-4.08). The laws do not give the WMC or the users, say as an association, the right to make contracts, collect funds, or disburse them. Nor do they delegate this to local government. To the present the WMC have not been organized except in perhaps four schemes; and these four have not been effective. The need to develop institutions for scheme 0 and M remains, since failure to develop such institutions is the basic reason why there has been such widespread deterioration at the medium schemes. This is addressed under the proposed project. 2.15 Most land is cultivated by the de facto owner on the majority of the schemes to be rehabilitated under the Project, and ownership rights have evolved according to Madagascar's traditionil land tenure rights. In some cases, such as Soavina (para 2.01), which was developed under the Government's program of "Aire de Mise en Valeur Rurale (zone for rural development), land tenure contracts between the Government and the users were established at the time the scheme was constructed but are now out of date. In selecting schemes for rehabilitation under the proposed Project, land tenure arrangements and any issues specific to the scheme would be reviewed. The impact of the proposed rehabilitation on land values-and the likely community production response to these opportunities will be assessed initially during scheme-selection and evaluated in the associated base-line survey (para 3.07). F. The Mngoky Scheme 2.16 As mentioned, the proposed project would assist the rehabilitation of the Lower Mangoky River Valley Development Agency (SAMANGOKY), the large parastatal irrigation scheme in the fertile delta of the Mangoky, on the western coast. Of SAMANGOKY's 7,770 ha under canals, 7,000 can be irrigated productively about equally divided between paddy and cotton cultivation. Deterioration at SAMANGOKY, as a direct result of poor maintenance, is illustrative Of the kinds of physical deterioration found on the medium schemes. The Mangoky scheme is a gravity irrigation network supplied by a river intake. A dike for protection from river flood waters suffered a break of several hundred m in 1983 and required emergency repairs to avoid a second break in the coming rainy season. The intake on the Mangoky river feeds water into a main canal through a desilting basin. As the dredger for the desilting basin has broken down, the basin and much of the primary canal have silted up and are no longer effective. Irrigation is through concrete-lined and prefabricated concrete canals,  - 14 - most of which are above ground. Many of these leak and require resealing. The drainage system is excensive, but has become in part ineffective. The resulting high water table has caused soil salinisation and reduced yields of cotton. 2.17 SAMANGOKY has suffered large annual losses which continue (Annex Table 15). To restore financial balance, the scope of SAMANGOKY's tasks and responsibilities have been reduced: Civil works for extension of the irrigation scheme, which were to be executed by SAMANGOKY, have been contracted out. SAHANGOKY's cropping activity is being reduced and acreage is leased to large enterprises, such as SOTEMA, a cotton textile company. These will be responsible for all aspects of land preparation. In rice production, farmers are substituting draught oxen and human labor for mechanical land preparation by SAMANGOKY. In irrigation, the responsibilicy nf SAMANGOKY will be limited to operation and maintenance of the main it:rigation and drainage canals. Farmers will be required to maintain I..ertiary and quaternary canals and to pay higher water-fees than in the pAst. SAMANGOKY's staff is gradually being brought down to a reasonatle level. In late 1984 total employees were 460, Less than half of the tot-lI of two years earlier. 2.18 SAMANGOKY is also introducing individual standard contracts which should encourage production, by giving cultivators incentives to improve their plots. It gives usufruct rights to the surviving spouse or one of the adult children of farmers who paid regularly their water fees provided the heirs continue to meet such payments. The demand for these new contracts far exceeds the number available and only farmers who have paid their debts to SAMANGOKY are being considered for the new contracts. In the case of cotton, farmers with a history of low yields (less than 1.8 t per ha) are being discouraged from signing contracts, since at lower yields the high water charges would mean a relatively low return to their labor. G. Subsector Issues 2.19 The main current issues peculiar to the sub-sector can be summarized as follows: (i) Planning. The Government needs to establish a general strategy for dealing with the many issues involved in irrigation, flood control, and erosion conctrol. This strategy should guide the Government in carrying out daily activities, in dealing with the recurring hurricanes and in development programs. A plan consistent with the strategy is also ceeded for carrying out effectively the many rehabilitation investments which donors are likely to support ia the near future. This is particularly urgent as an aid to improved coordination of external financing. The proposed Project would contribute substantially toward this goal (paras 3.01 and 3.02).  - 15 - (ii) Operation. Operation of the schemes can be improved at all levels. A strategy is needed to clarify institutional arrangements, help strengthen and coordinate Government services to the schemes, and develop appropriate management systems. A basic consideration here is the optimal distribution of tasks between Government and others. Hitherto, Government and parastatals have clearly been overburdened. But alternative arrangements for irrigation maintenance and operation (0 and M) remain to be developed. The proposed Project would move from administrative procurement to greater reliance on decentralized market activities for provision of agricultural inputs and irrigation services. It would also institute scheme level management of 0 and M. (iii) Finance. Insufficient cash flow and inability to acquire inputs are a serious and pervasive problem, most clearly seen in the larger schemes. The legislation of 1981-82 to introduce irrigation cost recovery on the official schemes needs to be completed. In part because of the strong political resistence which its engenders, adequate cost recovery merely to fund 0 and K remains one of the risky elements in the proposed Project. To reduce this risk, several institutional innovations are included to support'scheme level finance of 0 and M at the schemes to be rehabilitated (paras. 4.06-4.10). (iv) Economic considerations. Investment cost per hectare depends on the degree of rehabilitation and improvement, including the degree of water control desired. There has been considerable discussion in Madagascar on the trade-off between more capital intensive irrigation works, with greater demands for sophisticated management compared to simpler improvements which are less demanding of 0 and M but less promising for boosting production. This issue would be addressed in the proposed Project's studies of the individual schemes. It is believed that priority should usually be given to rehabilitation rather than new developments, because of presumably lower costs per benefit gained. This is probably true, but the possibility of developing new schemes with superior benefit-cost ratios should not be dismissed as a long run consideration. III. THE PROJECT A. Rationale for Bank Group Involvement 3.01 Production of irrigated rice on the medium-sized schemes makes an important contribution to one of Madagascar's immediate agricultural goals: food self-sufficiency. In part the means to this goal are to consist of restoring infrastructure and returning to agriculture's production frontier", that is attain the level of production implied by efficient use of all existing resources. In this perspective the proposed  - 16 - Project has potential which goes well beyond the 17,100 ha to be rehabilitated through it in that it is designed to develop a systematic approach to scheme rehabilitation which could be progressively applied to the estimated 140,000 ha in medium schemes and part of the 300,000 ha in the small schemes. Hence it also could directly benefit a substantial part of the farming population. In order to realize this approach the Project includes a heavy technical assistance component at MPARA to develop ministerial machinery to select, prepare and implement scheme rehabilitations, both during and after Project implementation. This machinery is to be designed so as to encourage other external financing agencies to participate in the overall rehabilitation program, assuring a good economic rate of return to any funds so invested. In its emphasis on self-finance and self-management at the scheme level, the Project is to develop a general alternative to the reliance on government management and funding which has led to the maintenance failure and widespread deterioration in the infrastructure of the schemes. This approach, if successful, would assure adequate operation and maintenance for the medium schemes. The Project is, therefore, the pilot phase in a long term program for rehabilitation of the medium schemes, and perhaps some of the small schemes as well. B. General Description 3.02 The project would support the rehabilitation of and about 16 medium sized irrigation networks, irrigating a total area of about 17,100 ha, and of the large-scale SAMANGOKY scheme. It would include an important institution-building component to reinforce planning and management capacity, for the Ministry of Agriculture and for individual schemes, and specific irrigation rehabilitation works. For each network, the rehabilitation cycle would start with a planning exercise, involving both MPARA technicians and on-scheme representatives. It would include a socio-economic review leading to a feasibility study and conclusions on scheme-management arrangements, and then execution of civil works per se. The result of this exercise would be, for each scheme, a rehabilitated irrigation network and agreed mechanisms for financing and maintaining it on a self-sustaining basis in the future. Project investment costs also include certain expenditures in support of agricultural development (preparation studies, travel support to extension staff), although most such expenditures are to be financed through parallel programs like the on-going IFAD-supported Highland Rice Project, or the FED technical assistance in the Toliary region, and through input-purchases by farmers themselves. Output increases including that of SAMANGOKY would be achieved by expanding the area under irrigation in the schemes, by increasing yields per ha and by additional double cropping. 3.03 Reconstruction would normally be designed to restore existing structures and to avoid further deterioration of the schemes, but in some instances would also include correction of design flaws (for example improvement of desilting structures). Works to improve water control above the original standa-ds would be undertaken if they prove economically justified and if it is probable that users will have the higher degree of discipline required for successful management and increased production. Detailed engineering design is being completed on the three first schemes  - 17 - selected under this process which would be rehabilitated during 1985 and 1986. These will be a model for the additional rehabilitation works, which would nonetheless, be the subject of individual planning exercises. The expectation is that rehabilitation programs and management arrangements will vary substantially, according to local circumstances, from relatively sophisticated schemes involving substantial capital investments, to relatively inexpensive repair works to support a simple gravity scheme. Criteria governing choice of technology appropriate to each scheme have been discussed and would be reviewed at negotiations; the central issues would be ability to maintain schemes and the economic return to proposed investments. 3.04 The Project includes, therefore, the following components: (a) Management support for the MPARA, primarily through a broad-based technical assistance program to support the several dimensions of the rehabilitations (civil works, scheme finances and 0 and M, agricultural development); (b) Civil works for reconstruction of medium irrigation schemes including pre-investment studies and planning for scheme operations and maintenance to be carried out by consultants; and Cc) Rehabilitation of the Mangoky scheme, including technical assistance, rehabilitation works, maintenance equipment and studies. C. Detailed Description 3.05 Management Support for MPARA. MPARA, the executing agency for the project, would manage implementation through four of its five Directorates and a Coordinating Committee, most of whose members are MPARA Directors (para. 4.05). (Project organization is shown in Chart 1.) This Committee has overall steering responsibility to assist a Project Coordinator, in the MPARA Programmes Directorate (DP), who as the general manager of the Project, would be responsible for overall project implementation, planning, management and coordination. He also serves as Secretariat to the Coordinating Committee. Line responsibility for the functions set out below will be handled by the MPARA Directorates indicated: (a) Directorate of Rural Infrastructure (DIR) --engineering design, civil works, and on-scheme arrangements for self-finance and management;  - 1.8 - (b) Extension Service (DYA) -al' aspects of agricultural development related to the project; and (c) Finance Directorate (DFP) -financial management and controls. 3.06 Effective management by MPARA's directorates and the Coordination Committee is of critical importance for successful impleamentation of the project. But MPARA's shortage of experienced personnel, office-space, office equipment, vehicles, and support for staff travel (official travel indemnities for civil servants can cover only a small part of per-diem expenses) would greatly reduce Project effectiveness. The Project therefore includes measures to alleviate each of these shortcomings where it bears most directly ou implementation. The proposed technical assistance to MPARA of 29 man/years would be on board for four years and includes: (i) a senior advisor to the Project coordinator to assist him in financial control, preparation of the annual Project budgets and work programs, review of feasibility studies, and contract preparation and management; (ii) an agricultural economist to assist the DVA in organizing and supervising the baseline agro-economic studies (para. 3.07' on schemes to be rebabilitated, including training of interviewers; formulating and reviewing the agricultural development plans for each scheme; introducing effective monitoring and evaluation mechanisms; preparing annual work plans and reports for the directore's activities under the project, including working out support activities with cooperating services; (iii) a training/extension specialist in the DVA to re-train the local extension services for the various planning tasks (ii above) and their implementation; (iv) a senior advisor to the DIR to assist in supervising the engineering consultants firms, in reviewing the list of priority schemes for future rehabilitation, in reviewing preliminary and detailed engineering designs and in preparing the directorate's annual work program and reports under the project; (v) a rural institutions advisor to the DIR to assist water management committees in organization and operations (para. 4.09); (vi) two senior irrigation engineers to assist provincial irrigation services in appraising, selecting and supervising the consulting engineers for irrigation rehabilitation works; and  - 19 - (vii) an advisor, in the DP, in project evaluation and monitoring (para. 4.14). 3.07 Under the Project those members of the extension services working on schemes t;o be rehabilitated will receive additional training, guidance and equipment from MPARA to carry out or to help carry out the baseline surveys and work out the agricultural development plans [para 3.06 (ii)]. The baseline surveys will include a brief history of the scheme; data on production, population, production techniques and inputs used; land tenure situation; marketing channels, input availabilities, prices, storage capacities, family incomes, principal cash crops. These surveys will be used in the economic feasibility study, and as basic inputs in developing a Five Year Plan of agricultural development of the scheme. In addition there will be an annual implementation plan based on the Five-Year Plan (para. 4.11). Implementation of the agrecultural development plan will require the cooperation of existing organizations, including market-oriented operators, for its execution and will use resources independently available. Government has agreed that the Five Year Agricultural Plan for each scheme to be rehabilitated would be submitted to the relevant cofinancier for review. Governmet also agreed that IDA would review and approve the Five Year Agricultural Plans for the schemes to be financed under the IDA credit. 3.08 The agricultural development plan of each scheme would define the technical packages and their delivery system. In the Highlands, agricultural technical packages are well known as a result of intense research carried out until ten years ago under the GOPR program. 2/ There are also technical packages which remain cost effective for areas outside the highlands area. The agricultural research institute, FOFIFA, and the FAO-financed Madagascar Fertilizer Program are doing research in the area to confirm the validity of the existing packages and to develop new ones. The Norwegian-financed FIFAMANOR dairy and wheat extension Project and the French-financed KOBAMA flour mill project are conducting research on cultivation of wheat as an off-season irrigated crop in the Highlands. 3.09 Capital investment under this component includes pick-up vehicles, four-wheel drive sedans, motorbicycles, bicycles, construction of offices and office equipment (Annex, Table 6). 3.10 Civil Works: Medium Schemes. Preliminary engineering designs have been completed for three schemes, Soavina, Belamoty and Sehara, with 3,100 ha under canals. Their rehabilitation will constitute the first civil works tranche. Rehabilitation works will normally be limited to essential i.rigation structures including primary and secondary canals, but not tertiary and quarternary canals which are entirely maintained by the users. For the first phase program, the civil works would include repair of: 2| 'Groupement Operationel pour la Productivite Rizicole, a FED supported program during 1965-73 to provide increased rice production through extension work and dissemination of technical packages.  - 20 - (i) 1.5 km of flood protection dikes (1,500,000 m3 of earth-fills); (ii) river intakes and drain outlets; (iii) desilting structures (cleaning and enlargement); Civ) 40 km of canal embankments (100,000 m3 of earth-fills); (v) concrete lining of 20 km of canals (100,000 mi2); (vi) water-level control structures; and (vii) scheme service roads. The content of the civil works for the remaining 13 schemes would vary by scheme and would be defined during the planning for scheme rehabilitation (paras. 4.01 to 4.03). 3.11 Rehabilitation of access roads would not be financed under the Project. If access is a serious problem, however, the scheme in question would not be rehabilitated until repair of the needed access road had begun. Agreement on this point was reached at negotiations. The Ministry of Public Works, which is a member of the ad hoc Coordinating Committee, will be consulted in situations where access is a problem. 3.i2 Reihabilitation of the Kangoky Irrigation Scheme: The most urgent works in the Mangoky scheme are being carried out in 1984-85 with funds from the Mangoky Agricultural Development Project. These include re-outfitting the dredge; closing the breaks in the protective dike; and reconstruction of a secondary canal. The proposed Project would finance: ij the cleaning of the desilting basin, of the headworks, and of the access canal; Cii) resealing the concrete section joints, primarily for the 100 km of raised canals which require special polyethylene and polyurethane seals; Ciii) rehabilitation of 45 km of service roads; (iv) repair of the river-intake gate and of irrigation gates; (v) improving the main drain outlets; (vi) equipment to make possible improved maintenance of irrigation structures (spare parts for the floating dredger in the de-silting basin; power shovels to clean drains and canals; specially equipped boats to mow weeds growing in the larger drains; graders, trucks, motorcycles and other vehicles; small compressors and other small civil works equipment, and painting equipment);  - 21 - (vi) technical assistance in cost accounting. 3.13 The three senior technical assistants-hitherto financed by the CCCE--will be financed in late 1985 through 1987 under the IDA supported Cotton Development Project. They include an agronomist who is responsible for agricultural operations; an engineer responsible for planning and carrying out the maintenance of the irrigation system; an engineer responsible for restoring the repairs garage and spare parts depot to effective operations. Under the proposed Project the fourth long-term technical assistant is being financed to develop a system of analytical cost accounting for SAM&NGOKY and to oversee its implementation- involving training on the job - and integration into SAMANGOMY's operations. (To the present it is difficult for SAMANGOKY to distribute its costs by categories such as maintenance, extension, agricultural production, marketing and so forth. Determination of costs of services and inputs received by different producers - farmers by size of plot, large farms and so forth - is also a time consuming and ad hoc cxercise). 3.14 In complement to the physical restoration and basic reorganization, SAMANGOKY's management has redefined and reduced its objectives (para 2.17). SAMANGOKY now provides water for irrigation (which requires managing and maintaining the irrigation network); provides extension services; and sells agricultural inputs and equipment. Managemnat has also worked out a Five Year Work Program with a view to obtaining full recovery of operating costs (para 3.15). The updated version of this plan was reviewed at negotiations. It requires a further reduction in SAMANGOKY employees from 408 at the end of 1984 to 313 by 1988. It includes payment by Government of the local costs of project items to be imported in 1985 (FMG 183 million); a capital contribution by the Government (FMG 597 million); conversion of part of Samangoky's short term debt into a 10 year loan to the BTM (FMG 260 million) for FMG 1,040 million in total. It also includes annual payments from Government for maintenance of the headworks structures (FUG 245 million in 1986, and FMG 67 million in succeeding years). SAMANGOKY would be profitable by 1988 (Annex, Table 15 ) given valuation of capital at historical prices. 3.15 Several elements in the Five-Year Rehabilitation Plan are crucial. In 1984, 46% of SAMANGOKY's useful land was cultivated. In the Plan 97% is to be cultivated by 1986, while because of double cropping, ha cultivated would be 112% of ha available. By 1988, ha cultivated should be 124% of ha available (7,700 ha), at which time most rice land would be double cropped. In addition, substantial increases in cotton yields and, to lesser degree, rice yields are foreseen. To achieve these objectives SAMANGOKY is radically reorganizing its relations with its clients. In addition to individuals, private and public farms are entering into long-term contracts with SAMANGOKY in which SAMANGOKY supplies land and  - 22 - water against a fixed rent. These would account for about 60Z of total cultivation in 1988. (In late 1984, such contracts covered five firms for 2,150 ha). Individual cultivators would still till 40Z of SAMANGOKY's land in 1988. They are being selected on the basis of past performance (para 2.18) and will pay higher water charges than in the past. !i the case of paddy these will be 500 kg per ha per crop cycle. Average yields at SAMANGOKY are 3.6 t per ha. This makes returns per ha, after payment of the user charge, far higher than returns on Highland irrigation where yields are 2 t per ha. Notwithstanding the continuing fiscal burden of SAMANGOKY, Government believes that further increases in user-charges are not currently feasible, perhaps because of the radical changes in cultivator relations at SAMANGOKY which are still underway (paras 2.17-2.18). At negotiations it was agreed that execution of this plan including the cash contribution of FKG 597 million to Samangoky would be a condition of disbursement for the Samangoky component of the project. 3.16 Participation of Cofinanciers. The principal cofinanciers (CCCE, FED and IDA) have worked together intensively through Project preparation and appraisal and have expressed agreement in detailed joint aide-memoires on Project objectives; on division among cofinanciers of financing; on the mode of operations from selection of scheme through completion of the civil works. 3/ 3.17 IDA would be the principal supporter of institution-building. It would finance the technical assistance to MPARA and some of the preparation studies for future rehabilitation. IDA, along with the other cofinanciers, would also finance reconstruction of medium schemes. Initially the FED would concentrate on rehabilitation in the faritany of Toliary, and the CCCE in the Highlands. The rehabilitation program, updated every year, would specify the distribution of studies and civil works among cotinanciers. The Project has been designed with sufficient flexibility so that if an additional cofinancier wishes to participate, it will be feasible to engage an additional consulting firm to carry out the studies needed to permit expanded operation. It would also be possible to use the same sequence (selection, studies, construction, execution of works) for some of the traditional schemes should this prove desirable at a later time. 3/ The basic aide-memoire of the co-financiers is contained in the Project Implementation Volume.  -23- D. Project Costs 3.18 As indicated in Table 1, total Project costs are estimated at FKG 22.1 billion (US$29.6 million), of which 61X are in foreign exchange. Physical contingencies equal to 13% of base-line costs are provided for. Local inflation is assumed at 14Z per year. International inflation is taken at SZ and 7.5% for 1985 and 1986, and 8% annually for 1987-1989. The exchange rate was adjusted for these different rates to preserve its purchasing power parity of mid-1985. It moved, therefore, from FM(' 650 to the US dollar in 1985 to FMG 846 in 1989. Consequently, combined five year baseline costs in US dollars were inflated only 22Z for price contingencies. No provisiou for physical contingencies has been included for staff and equipment for which numbers are clearly established. Because value added tax and duties on locally procured imported equipment and material are substantial, taxes are estimated at 16% of total costs, and burden principally the civil works category. Taxes and duties on technical assistance, vehicles and equipment directly imported by MPARA or SAMANGOKY are excluded. Costs by project component are detailed in Annex Tables 1-8. 3.19 The unit or mean investment costs per ha of medium scbeme rehabilitation for a total of 17,100 ha are assumed equal to those estimated for the first tranche rehabilitation at the schemes of Behara, Belamoty and Soavina. Total costs for the medium scheme rehabilitation are estimated at about $800 per ha. in 1985 prices. The unit costs by category are detailed in Annex Table 5.'  lbble 1. Sumry Project Costs Share of Foreign ProJect Conut loa Fore Total I Foreg Tbtal Ba sote Etxce FM Mill1, US mi llionX MPARA 1M iement Support Co2olnt: Ctvil Works 113 210 323 0.2 0.3 0.5 2 65 Epqtriate Staff Studies, Consultant Fees, hAdits 114 1,024 1,138 0.2 1.6 1.8 8 90 Vehicles, Equlpment and Furniture 43 385 428 0.1 0.5 0.6 3 90 Operating Costs 285 357 692 0.4 0.6 1.0 5 56 Sub-total 555 1,976 2,531 0.9 3,0 3.9 18 78 Scheus Rehabilitation Ctonunt Civil Works 2,414 4,483 6,897 3.7 6.9 10.6 48 65 Ccuultant Fees, Pe-inyesta nt Studies, Supervision-of Wblsk 292 1,165 1,457 0.4 1.8 2.2 10 80 Scheim Maintenaze and Operatirg Costs 62 155 217 0,2 0.1 0.3 2 29 Sub-total 2,768 5,803 8,571 4.3 8,9 13,2 60 68 SMANOY Qonnt Civil Works 464 862 *1,326 0.7 1.3 2.0 9 65 Expatriate Staff, Studies, Consultant Fees, Atlxts, 3B 10 48 0.1 0.0 0.1 1 20 Vehicles, Equipuent O 276 296 0.0 0.5 0.5 2 93 Worklng Capital 1,486 OOD 1.486 2.2 0.0 2.2 10 00 Sub-total 2,008 1,148 3,156 3.0 1.8 4.8 22 36 la Mmej aut 5,324 8,927 14,251 8,2 13.7 21.9 100 63 Physical rxtirancis 619 1,IU8 1,807 0.9 I., 2,7 13 bb Price Qontigwncles 2,516 3,570 6,086 2,3 2,7 5.0 22 53 Total Project Gosts 8,466 13,685 22,152 11.4 18.2 29.6 135 61 (taxes) (3,536) (4.7) 21  - 25 - E. Proposed Financing 3.20 The IDA Credit of US $12.0 million would finance 41% of Project costs (Table 2). For the management support component (MPARA), this would cover 65Z of civil works (para 3.09); 95Z of expatriate staff and studies (para 3.06); 75Z of vehicles, equipment and furniture (para 3.09); and 36% of operating costs (travel allowances, para 3.06). Of the scheme rehabilitation component the IDA credit would cover 30X of rehabilitation works (para 3.10) and 40% of preinvestment studies and associated consultant fees (para 4.04). Finally, the Credit would cover 65% of the rehabilitation works on the Mangoky Scheme (para 3.12) and 100% of consultant fees, vehicles and equipment (paras 3.12, 3.13). with the exception of funds for assistance to the Mangoky Scheme, which would be passed onto SAMANGOKY as a grant, the Government of Madagascar would pass on all funds to the Ministry of Agriculture (MPARA). A PPA request is being processed, primarily for funding for the T.A. Team to MPARA and for engineering studies for scheme rehabilitation to enable timely Project start-up and preparatory studies needed for the second tranche of rehabilitation. 3.21 Co-Financing. Under parallel arrangements CCCE would finance 25% of vehicles and equipment for the institution building component (para 3.09), and FAC would finance the monitoring and evaluation advisor (para 3.06, vii). The EDF allocation for 1985 would be ECU 5.3 million (US$4.0 million) to cover the cos0s of rehabilitating the Behara and Belamoty Schemes in 1985-86; and three additional schemes, also in Toleary Province, in 1987 and 1988. In 1984 CCCE has committed FF 24 million (US$2.4 million) to the scheme rehabilitation component to cover the-costs of the Soavina Scheme in 1985-86 and the pre-investment studies of four addition.al schemes for rehabilitation in 1986 and 1987. CCCE has also scheduled in 1986 about FF 36 million (US$3.6 million) for rehabilitation of four additional schemes in 1986-88. 3.22 Government's Share. Over the five year implementation period, Government's participation is estimated at 10% of total project costs, net of taxes, as indicated below (US dollar millions): financial contribution for civil works 3.9 working capital, SAMANGOKY (para 3.14) 2.2 incremental operating costs of MPARA 1.0 7.1 taxes 4.7 net contribution 2.4 This amount is reasonable, given the extreme scarcity of Government budget resources which is expected to continue throughout the project period. IDA approval of the annual budgets and financing plans of the Project components would be a condition of disbursement for uncommitted expenditures financed under the IDA credit. Assurances were obtained at negotiations that the Government would make available funds required for project implementation. The level of budget support required and provisions for it would be specifically reviewed each year as part of th.e annual work plan exercise.  Table 2. Proposed Project F'ianeing Project C9!!!ent: Totcal IDA - cS FAC C4 Beneficiaries iDA Tui milLion 2 iWARA t!Mgment &~icQoe CiviU Wbrks 0.6 0.4 - _- _j.2 - 65 Expatriate Staff, Studies,65 Obnsultant Fees, Audits 1.5 1.4 - - 0.1 - - 93 Vehicles, Equlpsnt and Fuwniture 0.5 0.3 - 0.2 - - - 70 OperatLng Costs 1.1 0.3 - - 0.8 30 Scke Relmbilitation Cboponent Civil Works 12.7 3.8 2.7 4.1 - 2.1 - 30 CDnAultant Fees, Pre-invesutnt Studies, &zpervision of Works 2.0 0.8 0.5 0.7 - - - 40 Schi. HaLnteranoe and Operatfrg Cbsts 0.4 - - - - - 0.4 00 Civil Works 2.3 1.5 - - - 0.8 65 BEpatriate Staff, Studles, Consultant Iees, h dits 0.1 0.1 - - - - - 100 Vehicles, Hquipwant and Iurnictre 0.5 0.5 - - - - - 100 WorkiLg Capital 2.2 - - - - 2.2 - 00 Refurdving PPA 0.8 0.8 - - - - - 100 Oxbtiroencies (price) 4.9 2.1 0.8 1.0 0.0 1.0 - 43 Ibtal 29.6 12,0 4.0 6.0 0.1 7.1 0.4 41 Percentage 100 41 14 20 (1 24 1  - 27 - F. Procurement and Disbursement Profile 3.23 Procurement procedures are summarized in Table 3. All civil works to be financed by IDA (US$10.9 million including GOM contribution) would be under bidding documents, divided into lots to be proposed by the consulting engineers and allowing bidding for one, several or all of the lots. Procurement would be through international competitive bidding and domestic civil works contractors would be allowed a preferential margin 7*Z orer the lowest evaluated bid of competing foreign contractors. Sut-contracting to small regional contractors would be encouraged on the basis of rules clearly stated in the tender documents. IDA would review: (i) bidding documents and bid award procedures; and (ii) the bid evaluation and proposed contracts prior to contract award, for all civil works estimated to cost over US$100,000. International consultants to be financed by IDA (US$1.8 million) for the institution building component are being recruited following IDA guidelines. Consulting services for pre-investment studies and supervision of works to be financed by IDA (US$1.1 million) would be procured from local engineering firms, possibly in joint venture with international firms, following procedures acceptable to IDA. Four annual lots of vehicles would be procured through international competitive bidding or shopping among all locally represented manufacturers with adequate repair facilities. Office equipment and furniture would be procured based on price quotations from at least three local or foreign suppliers. Grouped vehicles, equipment and furniture purchases exceeding US$100,000 would require international competitive bidding following IDA guidelines and prior IDA approval of bidding documents, bid evaluation and draft contracts. Purchase of specialized irrigation and drainage maintenance equipment (US$0.3 million) for the Mangoky scheme from a single manufacturer would be allowed. Assurances on these procedures were obtained at negotiations.  - 28 - F Taohl 3. ProazeEEit Procedues ERXd&tur -CategorDes ICB (7R 1 D, [ NA. U1L --US$ Id IWIcx- )mAM mmwer-