Sri Lanka - Village Irrigation Rehabilitation Project

Document of The World Bank FiUE COPY FOR OFFICIAL USE ONLY Report No. 3363-CE STAFF APPRAISAL REPORT SRI LANKA VILLAGE IRRIGATION REHABILITATION PROJECT April 30, 1981 This document has a restricted distribution and may be used by recipients only in the performence of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS US$1 = Rs 18.0 Rs 1 = US$ 0.0556 UNITS AND EQUIVALENTS I acre = 0.405 hectare (ha) 1 long ton = 1.016 metric tons 1 meter - 3.28 feet (ft) 1 bushel of paddy = 20.87 kg 1 cube - 100 cubic feet 1 cubic foot = 0.0283 cubic meters FISCAL YEAR January 1 - December 31  FOR OFFICIAL USE ONLY ABBREVIATIONS AND ACRONYMS ASC - Agrarian Service Center CO - Cultivation Officer DA - Department of Agriculture DAS - Department of Agrarian Services DCB - Decentralized Budget DCS - Department of Census and Statistics DI - Department of Irrigation ERR - Economic Rate of Return FFHC - Freedom From Hunger Campaign GPS - Guaranteed Price Scheme "for paddy) ICB - International Competitive Bidding LDO - Land Development Ordinance LHG - Low Humic Gleys MADR - Ministry of Agricultural Development and Research MLLD - Ministry of Lands and Land Development PMB - Paddy Marketing Board RBE - Red Brown Earths RD - Rural Development RWS - Rotational Water Supply TA - Technical Assistant TS - Tank Supervisor VIR - Village Irrigation Rehabilitation UP - University of Peradeniya WMS - Water Management Specialist GLOSSARY Akarawela - recently bunded irrigable paddyland Anicut - diversion scheme Chena - shifting (slash and burn) cultivation Purana - traditional permanently bunded irrigable paddyland Maha - northeast monsoon, October-February Oligama - rainfed paddyland Vel vidane - irrigation headman Yala - southwest monsoon, April-September This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   SRI LANKA VILLAGE IRRIGATION REHABILITATION PROJECT Table of Contents Page No. I. SECTORAL BACKGROUND ....... .............1................. The Economic Context ................................... I Agriculture in Sri Lanka ................................ 1 The Irrigation Subsector ................................ 3 Village Irrigation and its Context ..... ............ 4 Village Irrigation Development Programs ................. 5 Bank Group Involvement in Irrigation .. .................. 6 II. THE PROJECT AREA ........................................ 7 Introduction .................................... 7 Physical Resources ....................... 7 Agricultural Development ................. 10 Production Constraints and Risks .......... .. ............ 12 Status of Water Management in Sri Lanka .. ............... 13 Agricultural Institutions .............. .. ............... 14 Inputs and Services .................. .. 16 III. THE PROJECT .......................................... 18 Project Objectives ...................................... 18 Summary of Project Features ............................. 18 Civil Works .......... ................................... 19 Water Management and Maintenance ........................ 23 IV. PROJECT COSTS, FINANCING AND PROCUREMENT ........ ........ 25 Project Costs ........ .............. ..................... 25 Financing ........................ ....................... 25 Procurement ............................................. 26 Disbursements ....... * .............. ..................... 28 Accounts and Audits ................... .................. 28 V. ORGANIZATION AND MANAGEMENT ............................. 29 Overall Project Management ............. .. ............... 29 Civil Works .... ......................................... 30 Operation, Maintenance and Water M4anagement ............. 32 Training .......................... 35 Technical Assistance .................... 36 Monitoring and Evaluation ............................ 36 This report is based on the findings of an IDA mission which visited Sri Lanka in August/September 1980 comprising Messrs D.J.W. Berkoff and D. Leeuwrik (IDA) and V. Branscheid and Y.K. Murthy (Consultants) and of a follow-up mission in January/February 1981 comprising Messrs. D.J.W. Berkoff and J. Kampen (IDA).  Table of Contents (cont'd) Page No. VI. PRODUCTION, PRICES AND FA1M INCOMES .. ................... 37 Agricultural Development with the Project ............... 37 Marketing and Prices ....... .... ......................... 39 Farm Incomes .......................................... . . 40 Cost Recovery ............................................... 0 .......... 43 VII. BENEFITS AND JUSTIFICATION ..... ......................... 44 General ................. ........ .............................. 44 Employment Effects . ..................... ......... ....... 44 Foreign Exchange Effects .... ............................ 45 Environmental Impact .. .................................. 45 Benefits to the Economy and Economic Analysis ........... 45 Project Risks and Sensitivity Tests ..................... 46 VIII. RECOMMENDATIONS ......................................... 48 ANNEXES ANNEX 1 - Agricultural Supporting Tables and Charts Table 1 - Supplementary Irrigation Requirements of Crops by Soil Type Table 2 - Water Balance of Walagambahuwa Tank Table 3 - Area Sown Under Mlinor Irrigation: 1975-1980 Table 4 - Area Harvested as Percent Area Sown Under Minor Irrigation: 1975-1980 Table 5 - Average Yields Under Minor Irrigation: 1975-1980 Chart 1 - Probability of Occurrence of Periods of Seven Days of Consecutive Rainless Days and M4oisture Stress Chart 2 - The Dry Zone Catena ANNEX 2 - Project Schedules, Sample Cost Estimates and Organization Table 1 - Approximate Number of Schemes Proposed for Rehabilitation: by District: 1981-85 Table 2 - Analysis of Civil Works Costs: Based on Preliminary Estimates for Schemes to be Rehabilitated in 1981 Table 3 - Sample Cost Estimates for Typical Tanks: by Method of Construction Table 4 - Sample Cost Estimate for Typical Anicut Chart 1 - Project Management and Implementation Chart 2 - Department of Agrarian Services: Organization Chart (Existing) Chart 3 - Department of Agrarian Services: Organization Chart (Proposed) ANNEX 3 - Project Cost Tables Table 1 - Project Cost Estimates and Schedule of Expenditures Table 2 - Estimates of Average Cost of Civil Works at Current Prices and Approximate Number of Project Schemes Table 3 - Construction Equipment List and Procurement Schedule Table 4 - Transport Vehicles List and Procurement Schedule Table 5 - Miscellaneous Equipment List and Procurement Schedule Table 6 - Estimated Schedule of IDA Disbursements  - iii - Table of Contents (cont'd) ANNEXES (cont'd) ANNEX 4 - Criteria for Selection of Schemes for Rehabilitation ANNEX 5 - Supporting Tables and Charts: Economic and Financial Analysis Table 1 - Cropping Intensities and Cropped Areas Assumed for Economic and Financial Analysis Table 2 - Prices Used in Financial and Economic Analysis Table 3 - Calculation of Construction Conversion Factor Table 4 - Crop Input Requirements and Yields Table 5 - Typical Budgets for Representative Farms in the Dry Zone Table 6 - Typical Budgets for Representative Farms in the Intermediate Zone Table 7 - Economic Cash Flow Table 8 - Relative Importance of Project Cost and Benefit Streams with Switching Values Table 9 - Sensitivity Analysis: Difference Between With and Without Cases: Yields and Cropping Intensities Chart l(a) - Locus of Present Values of Total Project Costs and Benefits Chart l(b) - Locus of Total Project Cost and Benefit Switching Values ANNEX 6 - Supporting Documents in Project File   SRI LANKA VILLAGE IRRIGATION REHABILITATION PROJECT STAFF APPRAISAL REPORT I. SECTORAL BACKGROUND The Economic Context 1.01 Per capita income in Sri Lanka remains low (US$220 approx.), and until recently, the economy was characterized by low rates of growth in output and employment. In contrast, social progress has been considerable: literacy is high, basic health facilities are widely available, population growth rates have declined to about 2% p.a. and average life expectancy approaches that of much more prosperous countries. Ambitious social programs have, however, increasingly been called into question by economic stagnation, and the Govern- ment that came to power in 1977 has, therefore, implemented a series of major reforms designed to revive the economy through dismantling controls over resource allocation; establishing realistic relative prices and incentives; and launching an ambitious public sector investment program. The economy's response has been encouraging. 1/ GDP grew by 8% in 1978, 6% in 1979 and a further 6% in 1980. Nevertheless, by 1980 it had become clear that the Government's strategy, in particular its emphasis on major long-gestation, capital-intensive investments, carried with it certain risks. Overheating of the economy was reflected in a sharp increase in the rate of inflation (which probably exceeded 30% in 1980) and the emergence of a severe balance of payments problem. Domestic inflation has been aggravated by adjustments to administered prices and adverse external developments, but recourse to infla- tionary finance in an attempt to maintain public investment has also been a significant factor. The Government has begun to address these problems, notably through adjustments to the investment program (including a 25% across- the-board reduction in investment allocations in the 1981 budget). Neverthe- less, further measures will almost certainly be required if the successes of the Government's policy reforms are to be maintained. Among these measures will be the need to emphasize quick-yielding investments (such as the proposed Village Irrigation Rehabilitation (VIR) project) to sustain medium term growth, while adjusting longer-term large-scale investments to the absorptive capacity of the economy. Agriculture in Sri Lanka 1.02 Sri Lanka is endowed with resources and a diversity of agro-climatic conditions which permit a wide range of agricultural pursuits. Agriculture remains of central importance to the economy and accounts directly for more 1/ For a full review of the economic situation see the latest economic report: "Sri Lanka: Key Development Issues in the 1980s" Report No. 2955-CE, HIay 20, 1980.  - 2 - than 25% of GDP, 50% of employment, 70% of export earnings and about 40% of government revenue. In addition much of the activity in the manufacturing, transport, and service sectors is directly or indirectly related either to the supply of agricultural inputs or to the processing and marketing of agricultural outputs. 1.03 The island can be divided into two climatic zones which determine the primary characteristics of agricultural production. The wet zone (average annual rainfall in excess of 1,900 mm) is situated in the southwest quadrant of the island and accounts for about 30% of the land area, with population densities as high as 700 per sq km. Rainfall comes with both the northeast and the southwest monsoons, permitting the double cropping of rice and other field crops as well as the cultivation of perennials, notably tea, rubber and coconuts. The dry zone (rainfall 900 to 1,900 mm) covers most of the remain- ing land area to the north, east and southeast. In contrast to the wet zone, rainfall is concentrated during the northeast monsoon or maha season (October through March) and cropping during the southwest monsoon (April through September) is only possible where irrigation is provided. Population densi- ties are much lower than in the wet zone, averaging only 75 per sq km. A transitional band between the wet and the dry zones is known as the inter- mediate zone. 1.04 A total of about 2.25 M ha is under permanent cultivation (34% of the land area), of which about 65% lies in the wet zone. Paddy accounts for about 0.65 M ha (divided fairly equally between the two zones), tree crops for about 0.93 M ha (concentrated in the wet zone) and other perennial crops for about 0.04 M ha. The remaining 0.62 M ha comprises mixed, rainfed farming (mainly in the wet zone) mostly in the form of small gardens around the homestead. In addition, up to 1.0 M ha in the dry zone is, in any one year, under shifting cultivation (locally known as chena). 1.05 In line with general economic developments, the rate of growth in agricultural value-added declined from 3.4% p.a. during the 1960s to only 1.9% p.a. between 1970-76. Coconuts, tea, and rubber fared particularly poorly, but there was also a striking decline in the rate of growth of paddy output-- from 6% p.a. (following the introduction of the new high yielding varieties) to only 0.9%. Erratic weather conditions, declining commodity prices, a rising effective tax burden, and a dual exchange rate system which discrimi- nated against traditional exports, all combined to depress producer incentives. This was compounded by the uncertainties caused by a long drawn-out land reform (1972-75) and a deterioration in the institutional support for produc- tion programs. Although the minor field crop sector (pulses, oilseeds, chillies, coarse grains) was reasonably successful, even this was primarily based on artificially high prices created by the effective ban on imports. 1.06 As with other sectors of the economy, the initial response to the Government's policy reforms was encouraging. Agricultural value-added rose by 10% in 1977 and a further 5% in 1978--the first sustained increase since the late 1960s. In 1979, the rate declined to about 2% due to drought, and a further drought limited growth during the current year. However, the overall increase during the period has been substantial, largely due to increases in paddy production which rose from an average 1.4 M tons between 1970-76 to  - 3 - successive records, reaching 2.1 M tons in 1980 despite the drought. Gener- ally favorable weather, the initial impact of the Mahaweli program, a 21% increase in the government procurement (GPS) price in November 1977, and the general effects of the policy reforms all played a part. In contrast to the paddy sector, production increases in the tree crop sector have been limited, due primarily to past neglect. The performance of minor field crops has also been poor, being adversely affected by import liberalization. The recent removal of subsidies on imported wheat and the introduction of a support price scheme for some minor crops should, however, bolster future production. The Irrigation Subsector 1.07 Rainfall in the wet zone is fairly well distributed throughout the year. Paddy can therefore be grown under rainfed conditions, and irrigation is normally limited to run-of-the-river diversions. Double cropping is widespread, with overall intensities averaging about 170%. In contrast, in the dry zone irrigation is normally essential for paddy cultivation. Most existing schemes are based on storage tanks designed to supplement rainfall in the maha season, with residual supplies being used for limited yala cropping. There are a total of about 180 major schemes which have a storage capacity of 2,000 acre ft or more (irrigating nearly 400,000 acres) together with a large number of smaller village schemes. Although often based on ancient systems, many of the larger schemes have been developed through colonization programs designed to relieve population pressure in the wet zone. In contrast, most of the smaller tanks are associated with traditional (purana) villages which often have a long history of continuous settlement. 1.08 In addition to the development of tank irrigation, past colonization programs have included a number of more complex river basin developments of which the most important were the Gal Oya and Uda Walawe projects. These have now culminated in the ambitious Mahaweli Ganga Development Program which aims to harness the perennial supplies of the largest Sri Lanka river system, not only for development within the Mahaweli basin itself, but also for transfer to neighboring catchments in the dry zone. It envisages the ultimate develop- ment of 900,000 acres of new land (effectively doubling the total irrigated area) and 970 MW of hydroelectric generating capacity. In addition to greatly expanding the irrigated area, it will link up a number of existing schemes, supplementing their limited catchment yields with assured perennial supplies. The present government has greatly accelerated the Mahaweli program which is now expected to account for more than 30% of 1981-85 public investment. 1.09 While the major emphasis in the irrigation sector is on the expan- sion of the irrigated area, there is a growing realization that the utilization of existing systems, and even of newly developed systems, remains well below their potential. This reflects on the one hand inadequate maintenance of irrigation facilities and their consequent deterioration, and on the other the wasteful use of water due to inefficient water control and management. Despite its preoccupation with the Mahaweli program, the government has begun to give greater attention to smaller scale projects designed to make more efficient use of existing systems. The resources available for this purpose are limited but a number of projects have been developed (some with Bank assistance), and this objective also lies at the heart of the proposed Village Irrigation Rehabilitation (VIR) project.  - 4- Village Irrigation and its Context 1.10 Minor irrigation schemes are defined as having a command area of 200 acres (80 ha) or less. Most were constructed by the proprietors with government aid confined at most to the provision of headworks and sluices. There may be as many as 25,000 such schemes throughout the country, with an irrigable area in excess of 400,000 acres. Of these about 15,000 are village tank schemes located in the dry and, to a lesser extent, the intermediate zones. In contrast, the 10,000 or so anicut (diversion) schemes lie largely in the wet zone. Many schemes have deteriorated or been abandoned, and the Ministry of Lands and Land Development (MLLD) estimates that only about 50% are in working condition at varying degrees of efficiency. 1.11 Under current water management practices, water is usually allowed to flow from field-to-field with few control structures, limited field channel development and head sluices that are seldom operated systematically in relation to rainfall or irrigation needs. Continuous flow leads to water losses and underutilization of the available land and water. Since it is difficult for individual farmers to change this system, increased guidance by government will probably be necessary to make the best use of the available resources. 1.12 Operation and maintenance have traditionally been the responsibility of the farmers themselves, with government support confined at most to arrang- ing pre-cultivation meetings, resolving water disputes, or undertaking major repair works. Management under the traditional system was controlled by the vel vidane, an irrigation headman who usually owed his position to inheritance or leadership qualities. Under his guidance, the farmers (at least in theory and often in practice) implemented an agreed and coordinated cultivation program below the tank. In return he received certain payments in kind (traditionally half a bushel per cropped acre). With the passage of the Paddy Lands Act in 1958, its subsequent amendment, and its replacement by the Agricultural Lands Act of 1973, the government attempted in various ways to replace the traditional system with elected committees under official sponsorship. During the 1960s, this system appears to have functioned fairly well, but further administrative changes weakened the effectiveness of the elected committees, with the result that the traditional system of control was undermined without being replaced by an effective alternative. Recognizing this, the present government has abolished the system of elected committees, replacing it with one which in some ways returns to traditional practice. Responsibility for day-to-day operation has (at least in theory) been returned to an irrigation headman (now elected) assisted by an appointed Agrarian Services Committee and officials of the Department of Agrarian Services (DAS). More than 4,000 cultivation officers (COs) have been appointed, and public funds have been provided through DAS to restore physical facilities. The COs are, however, generalists with wide responsibilities for the local provision and coordination of agricultural inputs and services, and their role in minor irrigation has been limited. Under the VIR project, technical support for minor irrigation provided by DAS will be strengthened with the aim not only of restoring physical facilities but also of introducing improved agricultural, water management and irrigation maintenance programs.  - 5 - 1.13 While the decline of traditional management partly explains the deterioration and underutilization of village irrigation systems, the fact that such systems form only one element in a complex agriculture, increasingly orientated to rainfed farming, also has effect. With time and population growth, paddy holding sizes have declined, while the developed area below the tank has tended to increase, aggravating the risks of cultivation. To offset these trends, and acutely aware of his cash, labor, and farm power constraints, the farmer has increasingly allocated his effort and resources to upland and chena cultivation, a process that was given added impetus during the 1970s by the high prices for upland crops (para 1.05). Rice cultivation continues to play an important role in village life, providing subsistence and stability, and conferring a certain social status on those who own paddyland. Neverthe- less, evidence is accumulating that, even in large rice-based colonization schemes, unauthorized chena cultivation can be the main source of income, incurring the least cash risks and providing the most satisfactory distribu- tion of monthly returns. 1.14 Chena is destructive of the land resource, and declining fertility, weed competition and soil erosion lead to low yields per unit area. The Gov- ernment is actively seeking to discourage chena, and this is an important rationale for re-establishing tank-based irrigated agriculture as a focus for promoting stabilized upland cropping. Rehabilitation of existing schemes involves fairly limited risks. In the case of new village irrigation develop- ments, however, given the substantially higher costs involved, more comprehen- sive area development planning is required to ensure that balanced agricultural systems are developed comprising both dryland and irrigated cropping. Even in established systems, there are undoubtedly complementary measures that could be taken, separately or in association with irrigation rehabilitation programs, to stabilize upland agriculture in lands bordering purana villages and to implement GOSL's objectives of discouraging--and if possible eliminating-- chena cultivation. Village Irrigation Development Programs 1.15 Over the years, GOSL has implemented a number of village irrigation rehabilitation programs under various budgetary heads. These have included a food-for-work program (supported by the World Food Program and implemented by the Ministry of Rural Development) which undertook repairs to some 6,000 village tanks between 1974-79, and projects financed under the Decentralized Budget. It was not, however, until the present government came to power that a systematic approach to village irrigation was adopted. Three types of program have been, or are to be, implemented: (a) Rural Development Projects, implemented through the Ministry of Plan Implementation, but including as a major component the rehabilitation of both minor and major irrigation schemes. The first (for Kurunegala District) was supported with an IDA credit (Credit No. 891-CE). Others have been implemented for Hambantota (Norwegian assistance), Matara (Sweden), and Nuwera Eliya (the Netherlands). A second IDA project has recently been approved for Puttalam and Matale (Credit No. 1079-CE). Further projects are envisaged for Badulla (IFAD) and Vavuniya and Mannar (possibly IDA).  -6- (b) Irrigation Department Programs, implemented under the Depart- ment's own budget. The first major program was implemented in 1980 and included the rehabilitation of about 150 schemes in six districts. This was essentially the first phase of the proposed VIR project which originated in proposals submitted by GOSL to the 1979 meeting of the Sri Lanka Aid Group. It was prepared between October 1979 and July 1980 by the FAO/World Bank Cooperative Program. (c) Dry Zone Farming Projects to be implemented under the budget of the Ministry of Agriculture. The first is to be for Anuradhapura and will be supported by the ADB. In contrast to most other programs this project includes not only tank rehabilitation, but also settlement of new areas by tank- based communities within the context of overall land use planning. 1.16 Although these various programs are financed under different budget heads, the civil works component in each is the responsibility of the Irrigation Department. Measures have been taken, therefore, to ensure that implementation of the proposed VIR project complements other programs. The dry zone farming project areas (notably Anuradhapura) have been excluded from the VIR project, while the districts covered by rural development projects will only be included once the minor irrigation components of such projects have been completed. Bank Group Involvement in Irrigation 1.17 The Bank's lending strategy has increasingly focussed on the agricul- tural sector which is expected to account for about 50% of the total FY80-85 program. The irrigation sector has been mainly supported in the context of the Mahaweli program with one project completed (Ln 653-CE/Cr 168-CE), one under implementation (Cr 701-CE, US$19.0 M) and a third recently appraised (for a credit of US$90.0 M). A technical assistance credit is supporting the long term planning of Mahaweli development (Cr 979-CE, US$3.0 M). In addi- tion to Mahaweli, IDA is supporting a Tank Irrigation Modernization project covering five major tanks in the dry zone (Cr 666-CE, US$5.0 1M). Finally, the two rural development projects discussed above (Cr 891-CE, US$20 M and Cr 1079-CE, US$33.5 M) include major components covering the rehabilitation of both minor and medium sized irrigation schemes. Two small irrigation projects have been completed--Lift Irrigation (Cr. 121-CE, US$2.0 M) and Drainage and Reclamation (Cr. 168-CE, US$2.5 M)--and performance audit reports prepared. The audit reports identified design deficiencies, in particular as a result of over-optimism of the response of beneficiaries but also due to inadequate attention to research and extension services and, in the case of the Drainage and Reclamation Project, to unusual soil conditions. As a consequence, net benefits have been lower than envisaged at appraisal, although the revised rates of return remain satisfactory. Over the last several years, project preparation activities have been directed towards obtaining a fuller understanding of the systems and institutions expected to have significant impact on project benefits. In particular, an Agricultural Extension and Research Project (Cr. 931-CE) has been initiated which would complement the activities proposed under this and other projects. In general, ongoing projects are progressing satisfactorily.  - 7 - II. THE PROJECT AREA 1/ Introduction 2.01 The project is concerned with the scattered, irrigable areas com- manded by small village tanks and anicuts. Two major dry zone districts (Anuradhapura and Polonnaruwa) are excluded but, apart from these, the project area comprises essentially the whole of the dry and intermediate zones, together with minor parts of the neighboring wet zone (see Hlap 15450). In addition to the main zonal divisions, Sri Lanka has been subdivided into 22 agro-ecological regions according to elevation, topography, climate, and major soils associations (Map 14076R). Most of the project area is covered by the four dry zone regions, which are characterized by flat-to-undulating terrain, and the three "low country" intermediate zone regions. Limited parts of the hilly, more steeply dissected regions of the "mid-country" also fall within the project area, while eastern Ratnapura extends into the true wet zone. Physical Resources 2.02 Climate. The main climatic features of importance to the project are those of the dry and intermediate zones. Rainfall characteristically shows a bimodal pattern with two well defined seasons. The greater part of the rain falls during October through January (maha) as a result of the fIorth East Monsoon. Lesser rains occur during MIarch through May (yala), the period of the South West Monsoon. A short dry season occurs in February and a longer and more pronounced dry period stretches from June through September. Map 15453 illustrates how the duration of wet and dry periods vary throughout the project area. Annual rainfall at the level of 75% probability of occurence varies from as low as 650 mm in the northwestern and southeastern parts of the dry zone to about 1,200 mm in the mid-country areas of the intermediate zone. On average, the most drought-prone districts experience serious droughts once in every three or four years during yala and once in every five to seven years during maha. The great variability of rainfall from year-to-year and the frequency of dry spells are dominant features influencing agricultural per- formance, both under rainfed and tank-irrigated conditions, and result in widely fluctuating crop areas and yields (Annex 1, Tables 3-5). 2.03 Because of strong, dry winds and high temperatures, rains during yala can be largely ineffective in the lower rainfall areas. They play a more important, if somewhat unreliable role in the intermediate zone and account for a major part of total rainfall in the wet zone. Temperature imposes no constraint on agriculture. Maximum and minimum temperatures average about 0 0 89 F and 77 F, respectively, with little variation about the mean. Provided adequate water is available, crops can be grown the year round. 1/ This discussion takes into account data derived from the 1973 Agricul- tural Census and other statistical sources, a socio-economic survey undertaken by the University of Peradeniya (UP) as part of project preparation and studies undertaken at various times by the Agricultural Research and Training Institute (ARTI).  - 8 - 2.04 Soils and Land Capability. The main soil types and associations are shown on Map 14077R1. Reddish Brown Earths (RBE) and their hydromorphic variant Low Humic Gleys (LHG) occur most widely in the dry zone, with an estimated area of more than 2 M ha or about 30% of Sri Lanka's total land area. HIost of the tanks are situated on these soils. Tanks are found to a lesser extent on the 0.4 M ha of alluvial soils scattered throughout the project area. Red and yellow latosols cover about 0.3 M1 ha in the northern Coastal region. They are as a rule highly permeable and unsuitable for tank construction. Other coastal soils have a combined area of about 0.6 M ha, and comprise non-calcic brown soils, sandy regosols and some black clay, saline and alkaline soils. 2.05 The most common soils in the project area (the RBE and LHG), occur in a catenary sequence in the gently undulating-to-rolling topography of the dry zone, which is a composite of minor watersheds and drainage basins (Annex 1, Chart 2). Drainage variations differentiate the soils, with the RBE being generally well drained at the top of the slope, moderately-to-imperfectly drained on the mid-slopes, and replaced by the poorly drained LHG on the lower slopes and in the valley bottoms. RBE are residual soils, while the LHG contain elements from the upper slopes transported through the erosive action of water. They are therefore heavier than the RBE and have a layer of calcium carbonate at about 3 ft depth. Infiltration rates for the LHG are much lower and moisture holding capacity higher. A high water table during most of the crop season and impeded drainage restrict land use possibilities. The RBE on the other hand are more amenable to diversified cropping with upland crops. Both soils are easily saturated, are sticky and plastic when wet, and hard when dry. The period that they can be worked easily with bullock-power after rain or irrigation is very short. 2.06 The depth of surface soils to impermeable bedrock is variable but in most instances not great. Water tables in the upper slopes are well down, rising to within 3 ft of the surface only in years of exceptional rainfall. In these areas, internal drainage during the wet season is good. In the middle and lower slopes, water tables may rise to the surface for short periods during the wet season, but are well down during the dry season. The LHG soils in the valley bottoms remain poorly drained for a large part of the year with water tables rising to the surface. Even during the dry season water tables may stay within a few feet of the surface (3-8 ft at Maha Illuppallama). Lateral seepage from the well drained areas of a catchment to lower portions can be considerable, thereby greatly affecting crop performance and apparent water use in the lower areas. 2.07 Water Resources and Hydrology. Outside the areas influenced by major schemes, the seasonal pattern and unreliability of rainfall make water the most limiting resource in the dry zone. Rainfall is occasionally aug- mented by water from anicuts, shallow dugwells and tubewells, but storage tanks of varying size represent easily the most significant supplementary source of water in all areas outside the major schemes, the Jaffna peninsular and the northern latosol region. These tanks allow a more efficient utiliza- tion of available rainfall and greater stability in annual crop production, as well as supplying water for domestic purposes and stock.  2.08 The distribution of tanks is governed by landscape, soil suitability, and social factors. Most village tanks are constructed across valleys on the less permeable (LIHG) soils. Ihigh tank densities are found in Anuradhapura, northern Matale, and Vavuniya districts, and in transition areas between the dry and intermediate zones, notably in Kurunegala whiclh has easily the highest number of tanks. Individual tank sites are influenced as much by social as by technical factors and are not always in the most favorable position in relation to the hydrology of their associated catchments. They are generally shallow, with an average depth above sluice level of about 10 ft when filled. Catch- ment areas are small, although tanks are sometimes built in cascade along the more substantial valleys. Streams are non-perennial and flow for relatively short periods following rain. Commiand areas are highly variable in relation to catchment area, storage capacity, and extent of reservoir. Irrigation potential and efficiency of water use can also vary tremendously, from the larger tanks that provide water in all seasons to small tanks which do not fully irrigate their command areas even during the maha. 2.09 Unfortunately, little is known about the hydrology of small tanks and reasonably complete hydrological information is available for the last four years for only one tank, at Walagambahuwa. This settlement itself is thought to be fairly typical of some 3,000 similar villages occupying a total rice growing area of about 100,000 ha. Rainfall averages about 1,200-1,500 mm, with 70% received during the maha season. The 75% probability of rainfall expectancy based on monthly totals is, however, only 750-850 mm. Annex 1, Table 2 summarizes relevant data on the characteristics of the tank and its catchment area. Although general conclusions can hardly be drawn from one tank, and Walagambahuwa is unusually favorably served both in terms of tank area to command and capacity, the following points emerge: (a) During maha the contribution of runoff to tank inflow is roughly twice that of direct rainfall. In yala, the contribution of runoff is negligible. (b) Catchment water yield varies widely from year to year in maha (5-25%), while it is more even, but negligible for irrigation purposes in yala (3%). The wide variations in maha are-due to differences in rainfall duration and intensity, changing vegetative cover, soil conditions and moisture status of the catchment over time. (c) Tank water losses are from evaporation, deep percolation, bund leakage and overflow. For the Walagambahuwa tank, total water losses range from 5-10 mam per day. Such high losses severely restrict the irrigation potential of small, shallow tanks in yala. 2.10 In addition to its hydrological investigations, the Department of Agriculture (DA) has tested improved agronomic programs over four years at Walagambahuwa (see para 2.23), which suggest that rehabilitation and improve- ments to existing tanks will have an appreciable impact on agricultural production in the dry zone, in particular through assuring supplies during the  - 10 - main maha season. If water management programs can be implemented to make best use of the maha rains, then significant supplies in some years can be transferred for irrigation in yala, especially in the higher rainfall areas of the intermediate zone. While this provides a reasonable basis for justify- ing the fairly modest rehabilitation works proposed, a more accurate assess- ment of irrigation potential requires further investigation to establish the implications of the wide range of conditions under which the tanks are oper- ated. DA has already initiated hydrologic studies in anticipation of the VIR project and these will be supported during the course of the project's implementation. Agricultural Development 2.11 Settlement Pattern. In the typical village, the traditional purana (or permanently bunded paddy land) is bordered by akarawela--irrigable land that has been more recently cleared (sometimes under official village expan- sion programs) and to which irrigation is extended in a wet year. Beyond the akarawela, further paddy areas may exist. These are primarily rainfed (olagama) but may receive drainage water and be partially irrigated when water is available. Purana lands have prior right to the tank's supplies with the akarawela next in line. Given this tendency to develop land beyond the tank's original command, it is unusual in the dry zone for even purana land to receive significant yala supplies, while maha supplies themselves may have become less reliable as demands on the available storage increase. In the interme- diate zone, yala cultivation is more important mainly because of the more favorable rainfall pattern but also because higher settlement densities have limited the extent to which additional land can be developed. Beyond the command of the tank or anicut lie the house plots and gardens, upland cropping areas and (in the intermediate zone) tree crop plantations (especially coco- nuts). Further afield and often a long way from the village lies the chena land. 2.12 Land Tenure. Land tenure tends to mirror this settlement pattern. Owner occupiers predominate on the purana land, especially in the long-settled, more densely-populated areas, while akarawela is more often allotted by Gov- ernment under the Land Development Ordinance (LDO), especially in the less populated districts of the dry zone. Encroachments are also significant, particularly in chena areas. GOSL's policy is to regularize encroachments (the unauthorized cultivation of state land), subject to certain criteria and limits, and this will tend to stabilize overall agricultural activity. Sharecropping is relatively infrequent on minor tank schemes, although there is some evidence that cash constraints and other factors are leading to the spread of leasing arrangements. 2.13 Size of Holding and Fragmentation. Average paddy holdings range from as low as 0.4-0.6 ha (1.0-1.5 acres) in the wet zone to perhaps 2 ha (5 acres) or more in the less populated areas of the northern dry zone. Of the nine districts covered by the UP survey, it was only in the intermediate zone districts of Kurunegala and Monaragala that less than 25% of the area was occupied by farms above 2 ha in extent, while average operational holdings of 3.6 ha and 4.8 ha were registered in Vavuniya and Mannar respectively. These latter estimates are, however, probably greatly overstated. Fragmentation  - 11 - of paddy lands is another common feature. This is often associated with the distinction between purana, akarawela and oligama lands, and reflects the history and social cohesion of the traditional village (with each family participating in successive new developments). This pattern distributes risks and allows relatively equitable access to the communal resources of the tank. The UP survey (excluding Anuradhapura) suggests an average of about 2.7 plots per owner. 2.14 Cropping systems and intensities. Cultivation of short duration upland crops (4 months or less) is possible under rainfed conditions in maha. In the intermediate zone, rainfed yala cultivation of short-term, drought- resistant crops such as sesame may also occasionally be possible. Cultivation of wetland rice is, even in maha, more problematic. Ten inches of rain are generally required for wetland preparation and weed control, so that sowing or transplanting is not possible before late-October/November, being frequently delayed until December or even January. Since the rains cease in January, late-sown crops and late-maturing varieties have to be completed with supple- mentary irrigation. During yala, rice cannot be grown without irrigation except in the wet zone. 2.15 Village irrigation schemes have therefore been developed to provide security to the subsistence rice crop. Paddy is the universal crop during maha and is greatly preferred during yala. Where water is scarce, and rice cultivation difficult, small areas of irrigated subsidiary crops are sometimes grown during yala, while vegetable cultivation based on dugwells and/or seepage from the tank may also occur on a small scale. Under the betma system, rights to cultivation of purana land during yala are temporarily distributed in relation to the stored water available at the end of maha. Such a system is an impressive way of ensuring equity, but favors yala paddy cultivation over the more extensive cultivation of subsidiary crops. In the intermediate and wet zones, yala cultivation is much more prevalent, and in some mid-county and up-county areas, vegetable cultivation is increasing in significance at the expense of paddy. Overall intensities under minor irrigation schemes vary considerably from scheme-to-scheme and year-to-year (Annex 1, Table 3), but given the evidence available, they seem likely to have the following broad orders of magnitude: Maha Yala Total Wet zone 90-95% 80-90% 170-185% Intermediate Zone 80-90% 20-50% 100-150% Dry Zone 65-90% 0-15% 65-105% 2.16 An important feature of rainfall uncertainties is the prevalence of crop failures. These can be either full or partial and are reflected in both harvested areas and yields. Annex 1, Table 4 summarizes published data on crop failure for the last five years. For Sri Lanka as a whole, they averaged 6% for minor irrigation in maha and 13% in yala. The rates of failure can, however, be significantly greater in the dry zone.  - 12 - 2.17 Yields. Crop yields are generally low for all crops including paddy. All-island paddy yields under minor irrigation average about 45-55 bushels/acre during maha (2,400-2,800 kg/ha) and perhaps 40-50 bushels/acre during yala (2,000-2,600 kg/ha). The difference between the two seasons reflects adequacy of water supplies since, other things being equal, yields in the yala season would be higher due to more favourable growing conditions. Yields in the project area are similar to these national average. They may be higher in the dry and intermediate zones than in the wet zone, but considerable variations occur from district to district. Schemes which require rehabilitation will-- other things being equal--tend to have lower-than-average yields. Production Constraints and Risks 2.18 The allocation of effort between irrigated and upland crops, the timing of cultivation, and the technologies that farmers adopt, are related to a complex of inter-acting factors that include: (i) the variability of, and risks associated with, rainfall; (ii) the availability of cash and labor; (iii) the farm power situation; and (iv) the farmer's ability to change his practices given his level of knowledge and his inter-dependence with others. These combine in a way which leads the farmer to give attention first to upland and chena cultivation, and only subsequently to paddy cultivation which is usually postponed until well into maha. As a result, the early maha rains are not fully utilized for paddy and the crop matures late in March/April, long after the maha rains have ended. From January onwards, the paddy crop requires a full supply of irrigation water and, as a result, most tanks deplete either before the irrigation season is over, thus endangering the crop, or at best are fully depleted by the end of the maha crop. This situa- tion has been further aggravated by the continuous expansion of the developed area below the tank and by the deterioration in the irrigation facilities. Since rainfall alone is insufficient to support a yala crop, and runoff is generally insignificant (see para 2.09), rainfall in yala (200-250 mm) is not utilized at all. The overall effect is to reduce intensities below the potential represented by the available water resources, while the risks associated with uncertain water supplies discourage the adoption of high yielding technologies. 2.19 The Risks Associated with Rainfall. Rainfed crops must be sown early since the rains cease in about mid-January and the chances of crop failure increase as sowing is delayed. In contrast, the risks associated with irrigated paddy are reduced by a delay in cultivation. If rice is sown at the beginning of the season before the tank has filled, it is grown essentially under rainfed conditions, subject to the same periods of moisture stress as upland crops but even more liable to damage due to its greater sensitivity. Since great variations in stored water occur as a result of rainfall vari- ability, a farmer who plants early cannot be certain that there will be adequate irrigation water available from the tank to mature his rice crop. If, however, he waits with his sowing operations until the tank has filled (normally December), he can be certain that there is a sufficient supply to ensure survival of his crop even without further rainfall. In the event that the tank does not fill, the planted area can be adjusted accordingly and in particular akarawela and olagama lands can be left fallow. By reducing the risk of crop failure, the farmer is more willing to finance the substantial  - 13 - outlays required for paddy cultivation. Furthermore, such outlays tend to be lower if planting is delayed since farm power requirements and labor for weeding are both reduced if sufficient water is available (from prior rain- fall or the tank) to flood the land. 2.20 Farm Power. Early cultivation under dry or semi-wet conditions requires substantially more power than mudland preparation. Although farm power availability has much improved over the last few years due to the rapid introduction of farm tractors, running costs and hire charges for tractors have risen rapidly. Inevitably, farmers adjust by reducing power require- ments, chiefly by "trading" water supplies for power. This trade-off is an important one. While cash outlays per acre of crop may be reduced by such practices, they also contribute substantially to the failure to exploit the full potential of the overall water resources available. 2.21 Weed Control. Weeds are a major problem under both irrigated and rainfed conditions. Cultivation practices have, therefore, evolved to minimize weed competition at least cost to the farmer's available resources of labor and power. Chena land is abandoned after at most two seasons, with weed growth as much as declining fertility responsible for this rapid rotation. In paddy areas, farmers use standing water to control weeds. Not only does this encourage delays in planting but it is also an important contributing factor to the practice of continuous irrigation based on field-to-field flows which leads to significant waste and the early depletion of the tank. Status of Water Mlanagement in Sri Lanka 2.22 Irrigation practices adopted on minor schemes in Sri Lanka appear in many respects to be wasteful. 1/ It must be recognized, however, that they represent a largely rational response by the individual farmer to the risks he faces and to the trade-offs possible between water on the one hand and the costs of farm power, labor, and agricultural inputs on the other. Given the interdependence of farmers within the command area, individuals cannot affect significant improvements on their own and changes must be introduced on the basis of complete schemes. Such changes are unlikely to emerge without stimulus from outside, and are unlikely to be successful if they fail to respond to the constraints on farmer behavior outlined above. 1/ Although not as wasteful as in some major schemes for which water diver- sions as high as 20 ac ft/acre have been registered and where they are commonly as high as 6-8 ac ft/acre. The problem of excessive water use in major schemes is being tackled under a major water management project implemented by the Department of Irrigation with USAID assistance in the Gal Oya scheme, and under the IDA-supported Tank Irrigation Modernization Project for five major tanks in the northern dry zone. Water management programs also feature in the Mahaweli areas. W4hile minor schemes face different problems, these ongoing programs may nevertheless have signi- ficant implications for the minor irrigation sector.  - 14 - 2.23 Numerous studies on water management are available in Sri Lanka, mainly focussed on major schemes. Most recommend the introduction of short- season varieties and early cultivation during maha to make better use of the available rainfall. Experience at Walagambahuwa and elsewhere has shown that a paddy crop sown with the early rains in October can in most years survive on rainfall until the end of December. Since a four month variety sown in October matures in February, only 1-1/2 to 2 months of irrigation from the tank is required compared to 3 to 4 months for a late sown crop. Storage provides therefore more certain protection for the completion of the maha crop, and water is more likely to be available for yala. The introduction of irrigated upland crops in yala would then make the best use of the limited water supplies since inevitably these will be insufficient in most minor schemes for a full yala rice crop. 2.24 The general objectives of water management programs are therefore well established. To encourage the wider adoption of such practices, guide- lines have been issued by the GOSL authorities which emphasize the efficient use of rainfall through early planting and the adoption of short season varieties (para 2.27). 1/ 2/ A more systematic approach to water management forms an important element of the VIR project, which it is hoped will strengthen DAS support for minor irrigation (para 3.14). Agricultural Institutions 2.25 Sri Lanka's agriculture is served by a profusion of government insti- tutions and organizations at the national, district, and local levels. About ten ministries are directly or indirectly involved with agricultural adminis- tration. The most important of these for the project are: (a) The Ministry of Lands and Land Development (MLDD) which is responsible for irrigation development outside special areas; the operation and maintenance of major irrigation schemes; land settlement and administration; and forestry. (b) The Ministry of Agricultural Development and Research (MADR) which is responsible for promoting field crop production (including rice and subsidiary crops) and minor export crops. 2.26 The Department of Irrigation (DI) in the MLDD is responsible for all irrigation and drainage development outside the areas under the authority of the Mahaweli and River Valley Development Boards and for the operation and management of all schemes over 200 acres in extent. In addition to irrigation works financed under its own budget, it is also responsible for implementing irrigation under the decentralized budget and the district rural development 1/ Ministry of Agricultural Development and Research, and Ilinistry of Lands and Land Development: "Cultivation Timetable", June 1980. 2/ Ministry of Agricultural Development and Research: Cultivation Rules Formed Under Section 42(i) of the Agrarian Services Act, No. 58 of 1979.  - 15 - programs. It will be responsible for the main civil works component of the VIR project and for overall project coordination under the Project Steering Committee. 2.27 The Department of Agrarian Services (DAS) in the MADR is respon- sible, under the Agrarian Services Act of 1979, for a wide range of activities concerned with agrarian management, servicing of farmers, coordination of Government agencies and increasing agricultural productivity. It is charged with the maintenance of the Agricultural Lands Register, the administration of tenancy regulations, the organization of cultivation meetings, the supply and distribution of agricultural inputs, and the operation and maintenance of minor irrigation works. It maintains a network of 536 Agrarian Service (AS) centers throughout the country, organized on a district basis with district offices under the direction of an Assistant Commissioner (Annex 2, Chart 2). An AS Committee, appointed by the Commissioner or his deputies, is associated with each AS center and includes both Government officials from relevant departments and farmer representatives. A DAS official serves as secretary to the committee and is supported by approximately ten COs who live in the villages (being locally recruited) and who are responsible for direct contact with the farmers. The Centers themselves represent a focus for the distribution of agricultural inputs, and provide office space for other Government agencies. 2.28 The DAS's role in the operation of minor irrigation is at present primarily fulfilled by the COs and other DAS local staff. They arrange for the election of village irrigation headmen, organize cultivation meetings in association with the farmers, and promote the implementation of agreed culti- vation timetables together with the extension service. In addition, however, DAS employs a staff of technical officers whose responsibilities have been largely confined to minor irrigation construction and repairs. A budget is provided annually to DAS for this purpose (Rs 10 million for 1981). Priorities are established in association with the COs, AS committees and the District Administration, and the actual work is carried out almost entirely by small scale labor contractors. Responsibility for minor irrigation was returned to DAS fairly recently and the department has yet to build up its full complement of experienced staff. 2.29 The Department of Agriculture (DA) in the MADR is responsible for research and promotional activities concerned with field crops, as well as for agricultural training. The Department is receiving IDA support in strengthen- ing its activities through the Agricultural Extension and Adaptive Research (AEAR) Project (Credit No. 931-CE). 2.30 The DA conducts research on paddy, other annual field crops, and minor export (tree) crops. It maintains a Central Rice Experiment Station, a Minor Export Crops Research Institute, and a network of five Regional Research Centers (RRCs) and twelve branch stations. Research at the Walagambahuwa tank settlement is undertaken by DA staff based at the Maha Illuppallama Regional Research Center. Sri Lanka has a strong rice breeding program, and local rice HYVs were widely adopted by farmers well before IRRI varieties were developed. However, only in recent years has DA stepped up research on other field crops suitable for irrigated agriculture. National Coordinated Research Programs have been launched for a number of crops, in association with international and bilateral research organizations. New  - 16 - varieties of such crops as chillies, green gram, black gram, pigeon pea (dhal), and cowpea should be forthcoming over the next few years. Under the AEAR project, an adaptive research program incorporating trials on farmers' fields is being developed, based on the Regional Research Centers and a network of 24 Adaptive Research Centers. These will be staffed to cover all Sri Lanka, and will be further supplemented by the use of mobile research teams. 2.31 The Extension Division of DA has about 2,300 field officers at District, Division and local levels, and in addition provides direct support to a number of other agencies including the Mahaweli Development Authority. The AEAR project envisages the rationalization of the current extension service through the establishment of a unified "Training and Visit" (T&V) extension system for all crops (excluding rubber and tea) and animal husbandry (excluding veterinary and breed improvement services). The existing number of extension staff is being increased by a total of 1,260 (960 field level workers, 103 agricultural instructors, 139 subject-matter specialists and 56 agricultural officers). This would bring down the number of farm families per field extension worker from 1,200 to about 600. The organization and staffing being established under the AEAR project would complement the rehabilitation works provided under the VIR project and no additional support is proposed. Inputs and Services 2.32 The policy changes introduced by the present government have had as one of their major objectives the liberalization of domestic trade, and the promotion of the private sector. This has affected both inputs and the marketing of outputs, and has greatly expanded the alternatives open to the farmer. Nevertheless, Government subsidies on inputs and procurement policies for outputs continue to affect the farmer's position. 2.33 Agricultural Input Supplies. Fertilizer is imported by the Ceylon Fertilizer Corporation, and in very small quantities by the private sector, and is distributed through multipurpose cooperatives, AS centers and (increas- ingly) private traders. Fertilizer consumption for paddy has recently grown significantly stimulated by the liberalization of trade and the introduction of a subsidy in September 1979 (which on urea was as high as 85% of the cif price). This subsidy has recently been sharply reduced, following the decision to limit the overall financial burden in 1981 to Rs 1,000 M and price increases of up to 100% have occurred. Paddy prices have, however, also risen and remain favorable for fertilizer use. Fertilizer availability will be strengthened with the completion of the urea plant now under construction, and with the development of CFC's network of regional stores. Provided overall price incentives are maintained, and provided trade liberalization continues, it seems unlikely that fertilizer availability will represent a significant production constraint. Agro-chemicals are imported by the private sector as well as by the Ceylon Petroleum Corporation. Pesticides and herbicides are usually formulated locally and distributed through a well developed network of private dealers. Although the Government controls the range of formulations, there is a need for more effective communication with the trade to ensure timely availability of adequate quantities of the required chemicals. About 10-15% of the total annual requirement of paddy seed is produced by DA on its 42 seed farms and in conjunction with private certified seed growers.  - 17 - Improved seed production has received support from IDA (Cr. 595-CE) as well as bilateral sources. It is unlikely, however, that substantial improvements in the supply of seed for field crops other than rice will occur in the near future. 2.34 Agricultural Credit. Institutional credit is handled primarily by the two state-owned commercial banks, the Peoples Bank (PB) and the Bank of Ceylon (BC). PB operates through Multi-purpose Cooperative Societies and their banking divisions called Rural Banks. BC operates directly, in partic- ular through its AS center branches. National agricultural credit policy is established by a Rural Credit Advisory Board, chaired by the Governor of the Central Bank, with the Central Bank's Department of Rural Credit functioning as the secretariat. 2.35 The bulk of lending is short-term (6 months) production credit at a 9% interest. Such loans are available for paddy and most subsidiary crops, (for a maximum of 4 ha of paddy and 2 ha of subsidiary crops), and are dis- bursed partly in kind and partly in cash. Medium-term loans are also avail- able from both banks for purchase of machinery and equipment at an annual interest rate of 14%, repayable over periods from 3 to 5 years. Restrietions on obtaining these loans vary with downpayment, but generally require that the borrower owns at least 1.2 ha of cultivable land and has the joint guarantee of the dealer and a close relative. High default rates on agricultural loans are a serious problem. The situation was aggravated during maha 1977/78 when GOSL instructed the banks to relax lending criteria. To improve loan recovery, lending criteria were tightened in 1979, leading to a large fall in the volume of lending. The banks are trying to regularize the position, but the overall situation remains somewhat unsatisfactory, although it is being tackled with support from an ADB credit. 2.36 Crop insurance is available to farmers under the 1973 Agricultural Insurance Law and is designed to indemnify farmers against loss, stabilize farm income, and promote agricultural production. It is in theory compulsory for those receiving official credit for paddy but can be extended to other crops and livestock. The scheme has been effective only to a limited extent although in principle it could play a significant role in minor irrigation schemes for which the uncertainties and risks of paddy cultivation are rela- tively great. 2.37 Marketing and Processing. GOSL operates a guaranteed price scheme (GPS) for paddy, as well as for a number of other crops such as pulses and coarse grains. 1/ The operation of the GPS system is discussed in Chapter VI. Most paddy is milled in the private sector, either directly or on behalf of the PMB. Private investment in storage facilities appears to have lagged behind its expanding role in trading, and the PMB will continue to play an important role in stabilizing seasonal price fluctuations. Considerable expansion in PMB storage facilities is currently under way, assisted by USAID. 1/ GOSL pricing policy is discussed in Volume II of the 1980 Economic Report: "Sri Lanka: Key Development Issues in the 1980s - Agricultural Producer Incentives."  - 18 - Fluctuations in the price of field crops have been even more pronounced than those for paddy, reflecting the limited role that the PMB has so far played in the marketing of these crops. Nevertheless, major upward adjustments to the price of flour and other factors appear to be contributing to a signifi- cant recovery in production, and further increases can be expected if programs to promote irrigated yala cropping of these crops prove successful. III. THE PROJECT Project Objectives 3.01 The project aims to increase agricultural production and farmer incomes in existing village irrigation schemes by (i) financing the rehabil- itation of village tanks and anicuts over a five year period (1981-85); (ii) strengthening the major government institutions involved in village irrigation; (iii) initiating a systematic water management program; and (iv) establishing evaluation programs designed to assist project implementation and the prepara- tion of further programs in the village irrigation sector. Summary of Project Features 3.02 (a) Civil works component (i) Rehabilitation of about 1,200 village irrigation schemes by the Department of Irrigation, including repairs to and remodelling of tank bunds, sluices, spillways, anicuts and irrigation distribution systems, including the pro- vision of appropriate field structures; (ii) Modernization of working schemes by the Department of Agrarian Services, with a view to facilitating the introduction of water management programs; and (iii) Provision of survey, drawing, construction, maintenance and quality control equipment in support of the civil works program, together with the necessary transport vehicles. (b) Operation and Maintenance Components (i) Strengthening of DAS's capacity to service the opera- tion and maintenance of minor irrigation through incremental staff and the provision of office, survey and construction equipment, and transport vehicles; (ii) Introduction of systematic water managemaent programs on project schemes; and  - 19 - (iii) Support for ongoing training programs in basic tech- nical skills as well as for new water management training programs to be provided at the I4aha Illuppallama Research Institute. (c) Project Evaluation and Investigations (i) Monitoring of project impact on agricultural production by DAS field staff, and the Sri Lankan Freedom from Hfunger Campaign (FFHC), with support from the Department of Census and Statistics (DCS) to check the yield data collected by DAS staff; (ii) Systematic data collection and evaluation programs by the Land and Water Use Division of the Department of Agriculture, covering the hydrology of village tanks, soils, land use and related information; and (iii) Regular socio-economic evaluation studies of project impact by the University of Peradeniya as a basis for introducing modifications to project design during proj- ect implementation and to assist in further planning. Civil Works 3.03 Conditions vary widely, from schemes so damaged that they are aban- doned to those that are operating but for which rehabilitation is desirable, either to maintain or to increase production. Despite this variation, certain features of rehabilitation would be common to most schemes: (a) Investigations and designs, including hydrological studies to establish the catchment area, runoff and water yields, surveys of the tank bed to determine present and potential reservoir capacity; surveys of the command area under the tank for agricultural and water management planning; designs of the tank bund (including remodelling and/or raising), spillway, sluice and canal structures; and cost estimates of the com- plete rehabilitation works; (b) Clearing the tank bund and borrow areas; scarifying, stripping and benching of the bund; earthwork to build up the bund to the designed section; protective turfing for the rebuilt reaches of the tank bund; and the gravelling of embankments used as farm or access roads; (c) Remodelling of the spillway and the provision of a crest wall; (d) Replacement of junction block-type sluices with tower-type sluices using Hume pipes and screw hoist cast iron gates;  - 20 - (e) Alignment of the main canals and field channels; desilting, re-sectioning and improving the main supply channels by the DI and the field channel system by the farmers with the advice of DAS; and the rehabilitation or provision, as appropriate, of control structures, drainage works, turnout structures and measuring devices. An assurance has been obtained that all plans and designs for facilities provided under the project, including those for water management, would be in accordance with specifications and guidelines acceptable to IDA. Initial sets of guidelines for (i) planning and designing, (ii) maintenance, and (iii) implementation of water management programs have been prepared and discussed with the GOSL authorities. 3.04 Construction Program. The civil works component to be implemented by the DI represents essentially a five year time slice of the ongoing program for rehabilitation of minor irrigation works which was initiated in 1980. An analysis of 375 preliminary estimates prepared by the DI for this program is given in Annex 2, Table 2. After excluding those schemes which failed to meet the agreed cost criterion (Annex 4), the following characteristics emerge (at 1980 prices): Average command area: 26.5 ha approx. (65 acres) Average cost per acre: Rs 4,150 Average cost per scheme: Rs 270,000 3.05 These averages have been used to estimate the approximate number of schemes to be implemented under the DI component of the project, taking into account the allocations for the project under the revised five year investment program 1/ and expected price increases. The resulting program is presented in Table 3.1, with further amplification given in Annex 3, Table 2. Table 3.1: APPROXIMATE PROGRAM OF REHABILITATION WORKS BY THE DEPARTMENT OF IRRIGATION 1980 1981 1982 1983 1984 1985 Total Surveys & Investigations (no.) 150 160 340 420 400 - 1,420 Schemes Rehabilitated (no.) - 65 160 250 360 365 1,200 Command Area Rehabilitated (ha) - 1,700 4,200 6,550 9,450 9,600 31,500 3.06 The tentative program proposed would rehabilitate about 1,200 schemes with a total command area of about 31,500 ha (78,000 acres) over the period 1981-85. Annex 2, Table 1 provides a provisional distribution of the civil works program by district. This distribution is based on: (i) the total number of schemes in each district with a command area between 20-200 acres; (ii) their present condition; (iii) other village irrigation rehabili- tation programs being implemented or planned; and (iv) the capacity of each 1/ The five year program was revised in early 1981 to take into account the budget cuts proposed for 1981 and the resource outlook for the period 1981-85.  - 21 - DI Range office to undertake additional work. The IDA-supported Kurunegala Rural Development project is scheduled for completion in 1983. A substantial program under the VIR project for this important district is proposed for the period 1983-85, and this accounts for a major part of the increase in the number of schemes in the latter years. Expenditures on minor irrigation in Puttalam and Matale under the RD-II project will peak in 1983, so that their inclusion in the VIR project would be delayed until 1984. The RD project in Hambantota is being supported by Norad and commitments have been made for the period up to 1982. A modest program under the VIR project for Hambantota is proposed for 1983-85. 3.07 Actual costs of individual schemes can vary greatly. The number of schemes that can be financed within the funds allocated may therefore differ from the tentative estimate of 1,200 presented in Table 3.1. Similarly, the allocation by district may vary depending on the performance of each DI Range office. The proposed monitoring and evaluation studies should permit the adoption of a flexible approach and no difficulty is envisaged in adjusting the annual construction program. Assurances have been obtained from GOSL that IDA will be consulted annually to consider any revisions that may be necessary. 3.08 Highest priority ought to be given to schemes providing the greatest economic returns to improvements and repairs, full account being also taken of any interactions between neighboring schemes in the same catchment. For instance, first priority should be given to those schemes out of use due to recent damage but which could be fully restored with relatively little investment or those for which major damage is likely within a few years. In contrast, long abandoned schemes that would require almost complete reconstruc- tion or those which are comparatively "safe" should be given lower priority. Given their large number, however, it would be impracticable to conduct a full economic evaluation of each and every scheme. Following the precedent estab- lished by previous rural development projects, 1/ criteria have therefore been agreed in an attempt to ensure the selection of reasonably dependable schemes. These criteria are based on current cost data and the most recent design procedures of the DI for village tanks, taking into account all avail- able hydrologic information. 2/ The criteria, given in Annex 4, will be kept under review during project implementation, particularly in view of the results of the Physical Resource Evaluation Study (para 5.23). 3.09 Status of Design. A tentative list of 375 possible schemes was presented to the appraisal mission in September 1980. Of these, a few had been taken to the stage of final designs but for the majority only preliminary estimates were available (Annex 2, Table 2). The DI had scheduled a program 1/ An initial set of criteria was established under the Kurunegala Rural Development project (see the SAR, Report No. 2292a-CE). Revised criteria have recently been agreed during negotiations for the Rural Development II project (Matale and Puttalam districts). 2/ Irrigation Department Technical Note No. 2 "Design of Embankments for Village Tanks and Medium Reservoirs," 1980.  - 22 - of detailed designs for about 150 schemes in preparation for the 1981 program. Although some delays have occurred, it is understood that more than enough progress has been made to ensure the satisfactory implementation of the proposed 1981 program (about 65 schemes - Table 3.1). To ensure that this momentum is maintained the DI has relaxed overtime, travel and related restrictions specifically for the investigation and design of minor schemes. Subject to the success of this measure during the current year, it is proposed that it should be extended in future years. 3.10 Construction Methods. For most tank projects, earthwork on the bund represents the majority of project costs. In carrying out the civil works program, major emphasis would therefore be placed on minimizing the earthwork required. Guidelines for carrying out the civil works program, and for evaluating design alternatives, have been issued to DI field staff. Every effort would be made to complete the rehabilitation cycle over a period of two years, with surveys and investigations completed in the first year and the rehabilitation works themselves within the following construction season. The emphasis in carrying out the civil works program would be on labor intensive methods. Where possible, labor contractors would be employed, and this would be the predominant method of construction in districts where adequate labor is available. Excavation and earthwork involved in rehabilitation of the field channel distribution system would be undertaken using farmer labor on a voluntary basis. 3.11 In some districts, labor is relatively scarce and more expensive, and the labor contracting industry less developed. Contractors with machinery are available, but some additional machinery would be required to ensure expeditious completion of the work in the northern dry zone (Mlannar, Mullaitivu, Trincomalee and Vavuniya) and the south east (Badulla, Monaragala and Hambantota). For those schemes to be rehabilitated with machinery provided under the project, a balanced unit of construction equipment will be supplied, giving maneuverability in narrow locations, and adapted to the short leads and lifts normally involved (Annex 3, Table 3). The units will be concentrated on those schemes involving relatively large amounts of earthwork, particularly in the northern dry zone where the terrain is gently undulating and bunds tend to be long. To the extent possible, tanks would be selected in clusters to facilitate the maximum utilization of construction machinery and to minimize transportation time. Although the units would be allocated on a district basis, they would move between districts in response to the workload and to ensure their full utilization. It has been assumed that these construction units would be sufficient for approximately 250 schemes over the period 1981-85, equivalent to about 20% of the total. Taking into account the employment of contractor machinery in a number of districts, it has been assumed that 40% of all the work undertaken will be by machinery, with the balance undertaken by labor-intensive operations. 3.12 Supporting equipment and services. In addition to construction equipment, the project would provide DI with transport vehicles for supervi- sion and investigations; surveying and drawing equipment for carrying out the expanded program of surveys and designs; supplementary workshop equipment to strengthen the DI's equipment repair and maintenance capacity; and seven  - 23 - quality control field laboratories. The construction equipment and transport vehicles will be maintained at two departmental workshops, at Rambawa in Anuradhapura (being established under the IDA-supported Tank Irrigation Modernization Project Credit No. 666-CE) and at Amparai, respectively. Some additional workshop equipment to ensure mobility and to account for the special requirements of the project will be provided (Annex 3, Table 5). To ensure proper quality control during construction, it has been agreed that seven field laboratories should be established to cover all the districts involved in the project. Most districts are at present dependent on staff and testing equipment from the DI's head office in Colombo, and only minimal quality control is maintained for village tank work. A small field laboratory has, however, been set up in Kurunegala and a similar approach is to be followed elsewhere in the project area. In addition to providing physical facilities and equipment, the project provides for incremental staff and related costs incurred by the DI and the Ministry of Lands and Land Development. 3.13 Normal maintenance of the facilities rehabilitated under the project would be the responsibility of the farmers, supported by DAS. However, DI would ensure satisfactory performance of the headworks and structures rehabil- itated under the project for a period of two years. Water Management and Maintenance 3.14 The project would strengthen the minor irrigation activities of DAS to undertake three major tasks: (i) the planning and implementation of water management programs on selected schemes; (ii) the operation of a new water management service to liase with farmers and to maintain agreed water manage- ment programs; and (iii) the continuation of the present program of repair and maintenance. 3.15 Water Management. Water management programs would be introduced on tanks to be rehabilitated under the project and on a number of neighboring (operating) schemes. The aim would be to make optimum use of the available water resources comprising both rainfall and stored water. In the early years, the program would be applied in accordance with the evolving capacity of the new DAS organization. The extent to which specific features of the program are introduced would vary, depending, for instance, on the size of the scheme, the number of farmers involved and the potential for a yala crop. In general, the program would consist of the following: (a) Construction of field channels with control structures; (b) Introduction of a rotational water supply (RWS) system where appropriate; (c) Plowing and sowing operations as early as possible in maha; (d) Introduction of after-harvest plowing following either the maha or yala crop so as to keep the soil open for easier rainfall infiltration and plowing early in the following maha season;  - 24 - (e) Operation of the sluice to ensure stored water is used only to supplement rainfall during the maha and yala, with the sluice closed when irrigation requirements can be met by rain; (f) Closure of the sluice at night to prevent night irrigation; (g) Introduction of a standby rationing system when stored water supplies fall short of normal requirements; (h) Promotion of short-duration paddy varieties to reduce water requirements and ensure an early harvest; and (i) Promotion of irrigated upland crops during the yala wherever soils permit. 3.16 A preliminary water management plan would be prepared before the rehabilitation works are carried out so as to ensure that physical improve- ments are consistent with the operational procedures to be implemented. This program would be prepared by DAS Agricultural Planning Teams, comprising an agriculturist (seconded from the DA) and a technical assistant (TA). Assur- ances have been obtained from GOSL that twelve such planning teams wouild be appointed by December 31, 1981. They would work closely with the farmers, the DI and other Government officials. Implementation and further adaptation of the water management program would be the responsibility of the farmers under the elected village irrigation headman, supported by regular visits and technical guidance from DAS officials to be called the "tank supervisors." Given the variability from scheme-to-scheme, and in rainfall conditions from year-to-year, the irrigation headman and tank supervisor would be given guidance on procedures for reacting to such changing circumstances without undermining the basic logic and equity of the water management program (see Chapter 5). 3.17 Staff and Training. The number and experience of the technical assistants in the department's ongoing repair and maintenance program are inadequate, while water management will comprise a largely new activity. TAs receive basic technical training at Moratuwa University, south of Colombo, and at the Hardy Institute in Amparai, and these DAS training programs would be supported under the project. In addition, a new training program in water management for DAS staff charged with the implementation of the proposed program would be initiated by the DA at the Maha Illuppallama Research Insti- tute. Cooperation and consistency between DI and DAS training should be ensured. Emphasis in DAS training will be on practical issues and extensive field practice on neighboring tanks in Anuradhapura and Kurunegala will be provided. 3.18 Coverage of the Water Management Program. In the early years, schemes would be selected in convenient locations to facilitate supervision by the limited DAS staff, the numbers so selected being determined by the capacity of the DAS organization. For other rehabilitated schemes, the water management improvements proposed (e.g., early plowing, closure of the sluice when it rains, etc.) would be implemented by the farmers themselves with the assistance of the COs and extension personnel. Working tanks located close to those rehabilitated under the project could, however, be included in the water  - 25 - management program. Funds would be provided to enable DAS to undertake the limited civil works necessary to modernize such tanks so that the implementa- tion of the water management program can proceed in a flexible, appropriate and efficient manner. It is assumed that the first such modernization proj- ects would be implemented in 1983, allowing time for the DAS water management organization to be established and to gain experience on the projects reha- bilitated by the DI civil works program. For the purpose of estimating costs, Rs 50,000 has been allowed per scheme (at 1980 prices) to cover the reconstruction of the distribution system (consisting of supply and field channels) for the introduction of water management, as well as any minor repairs or improvements necessary to the headworks. It is assumed that a total of about 500 schemes will be modernized under the DAS program during the period 1983-85. 3.19 DAS would also continue to be responsible for undertaking maintenance and repair work where the work involved is beyond the capacity of the farmers, but insufficient to justify a full rehabilitation project by the DI. The normal limit for such minor works has recently been increased from Rs 50,000 to Rs 100,000 per scheme but, in most cases, the total cost is considerably less than this limit. The modernization of operating tanks prior to the introduction of water management would be undertaken by the staff involved in this ongoing program of repair and maintenance. 3.20 DAS capability to respond to emergencies (e.g., threat of a bund breaking) would also be strengthened under the project by the provision of limited construction equipment to be based at regional offices and under the control of the regional engineer (para 5.07) and maintained at the DI work- shops to be strengthened under the project. DI quality control field units would provide quality control support to DAS repair and maintenance work, in particular where significant earthwork on the bund is involved. IV. PROJECT COSTS, FINANCING AND PROCUREMENT Project Costs 4.01 Total project costs, net of taxes and duties, are estimated at Rs 784 million (US$43.6 million), with a foreign exchange component of about 22%. Taxes and duties are waived on machinery imported by the public sector. Base cost estimates are at mid-1980 prices, and exclude the contribution farmers are expected to make in construction of the field channel system (see para 3.03(h)). Price contingencies equivalent to 68% of base costs have been allowed at the annual rates of inflation given in the footnote to Table 4.1. The rapid rates of inflation envisaged in these estimates reflect current trends. The overall costs of the project have been adjusted to correspond to the allocations in GOSL's Medium-Term Public Investment Program established in January 1981 but the divisibility of the civil works program is such that there should be no difficulty in adjusting the annual program to take into account any unforeseen limitation of funds. Table 4.1 and Annex 3, Table 1 present summary cost estimates, and details are available in the project file.  - 26 - Financing 4.02 The proposed IDA credit of US$30 million would finance approximately 69% of project costs, equivalent to 100% of foreign costs and 60% of local costs. The remaining funds (Rs 244 million or US$13.6 million) would be provided by GOSL. The IDA share reflects the resource constraints facing Sri Lanka, the efforts the government is making to meet these constraints (e.g. by cut-backs in the overall investment program which also affected the VIR project) and the social priority attached to the project. The estimated schedule of expenditures is given in Annex 3, Table 1. The project would include up to US$1.0 M of retroactive financing for eligible project expendi- tures incurred from October 1980 onwards, to facilitate the smooth implementa- tion of the 1981 program, and to ensure the early launching of the monitoring and evaluation exercise. Procurement 4.03 Equipment and vehicles with an estimated cif cost of US$2.5 million will be procured through international competitive bidding (ICB) in accordance with IDA guidelines. A preference limit of 15% of the cif price of imported goods would be extended to local manufacturers in the evaluation of the bids. Due to strong individual preferences, motorcycles and bicycles supplied to DAS staff would be purchased by the individuals concerned. 1/ Small off-the-shelf items costing less than US$20,000 each, for which international tendering is impracticable, would be procured after comparing prices from not less than three independent suppliers in accordance with GOSL procurement procedures acceptable to IDA. Such purchases will not exceed the equivalent of US$.50 M. 1/ This procedure was followed in both the Kurunegala and RD-II (Matale and Puttalam) projects. GOSL makes a medium-term loan to officials who in turn purchase the relevant item from a local dealer. The loan is repaid out of official travel allowances.  - 27 - Table 4.1: SUMMARY COST ESTIMIATE % of Base Local Foreign Total Local Foreign Total Costs --------Rs M--------- ------US$M----------- Civil Works Surveys and Investigations 6.5 0.7 7.2 0.4 ... 0.4 1.5 Department of Irrigation /a 270.3 75.6 346.0 15.0 4.2 19.2 74.3 Department of Agrarian Services /a 26.3 3.9 30.3 1.5 0.2 1.7 6.5 Sub-total 303.3 80.3 383.6 16.9 4.5 21.3 82.3 Equipment and Vehicles Department of Irrigation 3.4 35.3 38.7 0.2 2.0 2.2 8.3 Department of Agrarian Services 0.8 6.7 7.5 ... 0.4 0.4 1.6 Sub-total 4.2 42.1 46.2 0.2 2.3 2.6 9.9 Incremental Recurrent Costs Department of Irrigation 9.0 1.5 10.4 0.5 0.1 0.6 2.2 Department of Agrarian Services 10.1 1.6 11.7 0.6 0.1 0.7 2.5 Sub-total 19.1 3.0 22.1 1.1 0.2 1.2 4.7 Training, Evaluation and Technical Assistance 10.9 3.4 14.3 0.6 0.2 0.8 3.1 Base Costs: Total 337.4 128.7 466.0 18.7 7.2 25.9 100.0 Price Contingencies /b 276.6 41.6 318.2 15.4 2.3 17.7 68.3 Total Project Costs 614.0 170.3 784.3 34.1 9.5 43.6 168.3 /a Including 10% supervision and administration. /b Price Contingencies as follows: 1980 1981 1982 1983 1984 1985 --------------% per annum------------- Local Costs 25.0 25.0 15.0 12.0 10.0 8.0 Foreign Costs 12.5 9.0 8.5 7.5 7.5 7.5 4.04 Civil works under the project will be individually small and scat- tered. Construction will be seasonal and it would be neither feasible nor economic to group schemes into large contracts suitable for ICB. Thle civil works under the project will therefore be undertaken mainly by local domestic contractors, with a minority of the work being undertaken by force account. The DI and the DAS will retain full responsibility for construction standards. For contracts, the normal practice is to award the works for minor irrigation schemes on the basis of small unit price contracts to registered and approved contractors, supervised by technical assistants and engineers and with a works supervisor permanently on site during construction. Given the low average  - 28 - cost, these procedures will be followed under the project although the poten- tial of local competitive bidding will be explored where this is warranted by the works to be undertaken, or by the possibility of grouping a number of schemes under one contract. Disbursements 4.05 Disbursements under the IDA credit would be made for: (a) 70% of the costs of civil works, excluding the costs of supervision and administration; (b) 100% of the cif costs of equipment and vehicles procured abroad; and 100% ex-factory or 80% at point of retail sale for equipment and vehicles procured locally; (c) 100% of specified DAS training costs; and (d) 100% of the costs of evaluation studies and technical assistance. 4.06 Disbursements for civil works and training would be made against statements of expenditures, submitted to IDA through the office of the project coordinator, and audited according to procedures acceptable to IDA. Support- ing documents will be retained by GOSL and made available to IDA supervision missions if required. Disbursements for equipment and vehicles, technical assistance and evaluation studies would be made against full documentation and subject to normal IDA procedures. It is anticipated that the final disbursements of the IDA credit would be made by December 31, 1986, about one year after the scheduled completion of project implementation. The estimated disbursement schedule is given in Annex 3, Table 6, and the credit would be allocated as follows: Amount of Credit Allocated in US$ Equivalent % Financed 1. Civil works and materials 23.0 70 2. Equipment: directly imported 2.5 100% cif costs locally procured 0.2 80 3. Training 0.4 100 4. Evaluation studies and technical assistance 0.8 100 5. Unallocated 3.1 n.a. TOTAL 30.0 69 Accounts and Audits 4.07 The DI and the DAS are subject to normal GOSL expenditure and audit- ing procedures, which are considered satisfactory. Project accounts would be maintained separately by these two departments in a readily identifiable form to enable them to be audited separately from overall departmental accounts by an independant auditor. Budgetary allocations would be included in the vote  - 29 - of the DI, the lead agency for the project. The Project Management Division would be responsible for preparing the annual budgets for civil works and for assembling the overall budget request in consultation with the Project Steer- ing Committee (see para 5.01); for releasing funds to the range offices of the DI, the DAS and other agencies involved in the project; for maintaining the consolidated district accounts for the civil works implemented by the DI; for assembling the audited accounts of DI and other departments and agencies to which funds have been released; for calling for tenders procuring and supply- ing machinery, vehicles, equipment etc. provided for in the project; and for submitting reimbursement claims to IDA. The corresponding unit in DAS would be responsible for preparing the annual budget request for the DAS components; maintaining consolidated district accounts for DAS expenditures; and for preparing and furnishing audited accounts with claims for reimbursement to the DI. Assurances have been obtained from GOSL that: (a) separate accounts would be established and maintained for the project by DI and DAS; (b) accounts and financial statements would be audited annually by an independent auditor acceptable to IDA; (c) financial statements would be submitted to IDA for review not later than six months after the end of each fiscal year, and certified copies of audited accounts and the auditor's report not later than nine months; and (d) the audit report would include, inter alia, a statement that funds disbursed against statements of expenditure had been used for the purposes for which they were provided. V. ORGANIZATION AND MANAGEMENT Overall Project Management 5.01 Committees at three levels will be responsible for the control and coordination of the project: (a) The National Committee for Village Tank Rehabilitation, established in 1980, is responsible for the coordination of all minor irrigation works, irrespective of the funding or implementing agencies. It will be responsible for over- all control of the VIR project, and for deciding the annual district programs in relation to other projects and programs. (b) The Project Steering Committee in the MLDD, chaired by the Secretary of Lands, and with representatives from the MADR, the Ministry of Finance and Planning, the Department of External Resources, the DI, the DAS, the Survey Depart- ment, the Water Resources Board, the Freedom from Hunger Campaign and the University of Peradeniya, will be respon- sible for (i) the approval of the annual district programs  - 30 - and budgets, (ii) allocation of funds to the Departments and Agencies involved, (iii) reviewing and monitoring progress, (iv) resolving implementation bottlenecks and (v) recommending modifications in project design. (c) The District Agricultural Committee in each district, chaired by the Government Agent and consisting of the Additional Government Agent (Lands), the Regional Deputy Director of Irrigation, the Assistant Commissioner of Agrarian Services/ Regional DAS Engineer, the Assistant Director of Agriculture with the Development Officer of the district office as secretary, will be responsible for (i) identifying items for inclusion in the annual district programs, (ii) sched- uling a program of investigations by DI and DAS officers, (iii) submitting an annual program of works for the approval of the Project Steering Committee and (iv) monitoring progress and resolving implementation problems. These com- mittees were established under the Irrigation Ordinance and have wide responsibilities for the coordination of agricul- tural activities in the districts. They will provide a link with the proposed District Development Councils and will ensure consistency with other programs e.g. those financed under the RD projects or the DCB. In some districts, a spe- cific sub-committee may be established for the VIR project but this is not thought essential for all districts. 5.02 An assurance has been obtained that the Project Steering Committee and District Agricultural Committees in all districts partici- pating in the project will be maintained through the life of the VIR project. Through the Project Steering Committee, the MLLD will coordinate the activities of the various departments and agencies involved, and will be responsible for overall policy review and management. A coordinator appointed in MLLD will have particular responsibility for the supervision of the agencies involved in project evaluation. Funds for the project will be included in the budget of the DI, which is the major implementing agency. A Project Management Division has been established in DI, and will be responsible for the implementation of the DI program, as well as functioning as the general secretariat for the project as a whole (para 4.07). A Project Management Division in the DAS, under the proposed Deputy Commissioner for Water Management, will supervise the water management and other programs to be undertaken by that department, and will liaise closely with the comparable division in the DI. Civil Works 5.03 Department of Irrigation. The DI will be responsible for the execu- tion of the main civil works through its present organization, supported by machinery, vehicles, equipment etc. provided under the project. The regional operations of the department are organized under 15 Range Officers, each under a Deputy Director (DD). For major construction works, the Range DD reports to the Senior DD (Major Construction) in Colombo. For all minor works, he reports to the Senior DD (Regional Works). Overall responsibility for all regional works is held by the Additional Director of Irrigation. In addition to the deputy director, each Range has a position for a chief irrigation engineer, and is sub-divided into a number of divisions, each under an irrigation  - 31 - engineer supported by technical assistants, draughtsmen etc. A division can either be confined to a major irrigation project, or have general responsibil- ity for irrigation in a defined geographical area, including that for minor irrigation construction. While the DI as a whole is suffering from a serious shortage of qualified engineers due to emigration and competition from the private and parastatal sectors, the shortages are less severe in the case of technical assistants. It will be these who will be primarily responsible for surveys and supervising construction under the project, while their produc- tivity should be increased by the relaxation of restrictions on overtime, etc. (para 3.09). Given also that financial constraints are likely to curtail the initiation of other major projects, it is judged that these offices will have the capacity to implement the proposed work program. Assurances have been obtained that the staff position will be reviewed annually in establishing the district programs, and that staff will be transferred between districts if and when this is required. Consideration at a future date may be given to the possibility of employing private consultants to design and supervise construc- tion if DI staffing constraints pose a serious problem. This approach is being tested under the IDA-supported RD II project (Cr. 1079-CE), and experience under this project will be closely monitored. 5.04 Implementation Program. In selecting, investigating and rehabili- tating schemes under the project, a regular annual cycle will be followed: (a) Initial identification of potential schemes would be under- taken by the District Coordinating Committee, taking into account the advice of relevant Government officers, and would be completed early in the year prior to that in which actual construction is to take place. The length of this list would be determined by the district allocations set by the National Committee. (b) Identified schemes would be inspected by the technical staff of the DI and a preliminary investigation report would be prepared, sufficient to establish whether or not the scheme is likely to conform to the agreed criteria. (c) Approval of the revised list would be obtained from the Project Steering Committee at the latest by April and the DI, in cooperation with the District Coordinating Committee, would prepare a schedule of full investigations for the period April through September. (d) At the commencement of the full survey, a meeting would be arranged at the site, involving the officers of the DI, DAS and DA as well as the farmers under the leadership of their elected irrigation headman (vel vidane). Proposals for rehabilitation would be discussed and agreement reached on the approach to be taken and the timing of investigations. (e) In tanks selected for the formal implementation of water management, the DAS Agricultural Planning Team would advise the DI survey team of the structural requirements for water management and on completion of the DI designs, the DAS  - 32 - officers would inspect the final designs to ensure com- patibility with the water management and operation program proposed. Technical approval will be the responsibility of the Deputy Director of Irrigation in the district, with overall technical supervision provided by the Project Management Office in Colombo. (f) A revised list of schemes to be constructed in the following year will be submitted by the District Coordinating Committee to the Project Steering Committee at the latest by end-November. Final approval of the schemes to be included will be given by the Steering Committee by end-December. (g) Construction will commence with the release of funds in the new year. Prior to construction, a meeting will be held at the site to explain the project proposals to the farmers and to arrange for their participation in the project. Construction would follow the completion of the main maha harvest, being phased over the period January through September to avoid the main rains and avoid interruption with agricultural production. Every effort will be made to complete work on any scheme within the course of one season. Particular attention would be given to maintain- ing the quality of construction through regular visits by the proposed quality control field staff. 5.05 The cycle would be repeated with the initial selection of schemes for the subsequent year. In preparing the tentative project schedules, it is assumed that each technical assistant involved in the project will be able to supervise construction of three schemes, and investigate a further four to five in any one year. Operation, Maintenance and Water Management 5.06 Department of Agrarian Services. Under the Agrarian Services Act of 1979, DAS has responsibility for the Government's role in all matters relating to the operation and maintenance of minor irrigation. To date, this role has been largely confined to (i) organizing elections of irrigation headmen, arranging pre-cultivation meetings, etc.; and (ii) carrying out maintenance and repairs. Under its present organization (Annex 2, Chart 2), arranging elections, etc. has been the responsibility of the divisional and cultivation officers based at Agrarian Services Centers. These are organized on a district base, with the district organization headed by an Assistant Commissioner reporting to the Commissioner of Agrarian Services in Colombo. The repair and maintenance activities are implemented by a staff of technical assistants, who report for administrative purposes to the Assistant Commissioner of the district but for technical direction to the Chief Engineer in Colombo. Overall direction and control of the repair and maintenance program is provided by a Deputy Commissioner located in Colombo. The number of technical assistants in place is well below the approved cadre. Of a nationally approved total of 444 technical assistants, only 290 are in place of which none are of Grade 1 status (70 approved) and only 116 are Grade II (214 approved). Of the latter, some are on secondment from the DI while they also include about 30 re-employed pensioners.  - 33 - 5.07 Under the project, DAS will be strengthened to facilitate the imple- mentation of a systematic water management program. Assurances have been obtained from GOSL that the necessary administrative changes will be under- taken in accordance with a program and timetable satisfactory to IDA, and that their effectiveness will be reviewed not later than end-1984 as a basis for assessing the long-term role of DAS and other agencies in support of minor irrigation. The revised organization proposed is shown in Annex 2, Chart 3. Seven regional offices will be established, covering the whole country, each under an engineer reporting to the Chief Engineer in Colombo. Three of these offices cover the majority of the initial VIR project area and will be sup- ported under the project. A fourth office is proposed for Kununegala and Puttalam which will participate in the VIR project as from 1983/84 (para 3.06). These offices will direct three main activities: (a) The planning and implementation of water management programs on identified schemes. Twelve agricultural planning teams will initially be established, each comprising an agriculturalist and a technical assistant recruited from DAS's existing staff. (b) The operation of a new water management service to liase with farmers and to maintain agreed water management programs. A new cadre of "tank supervisors" would be established, the numbers of whom would increase as additional schemes are brought into the water management program. (c) The ongoing repair and maintenance program. Existing staff would be strengthened through an expanded training program, and provision of additional vehicles and construction equipment. 5.08 Three divisions corresponding to these activities will report to the Regional Engineer. Although staff would be organized under separate divisions, there is a complementarity between water management (being concentrated during the main growing period) and repair and maintenance (being concentrated in the off-season). It will therefore be possible to provide some cross-support between the two programs. The district offices of DAS under the Assistant Commissioners will continue to be responsible for the administration of the Agrarian Services Law, including the management of the Agrarian Service Centers. 5.09 The Regional Engineers will be responsible for two or three dis- tricts. Their staff may be based at district offices to permit easy access to the field, but for all but administrative purposes will report to the Regional Engineer. A Deputy Commissioner (Water Management) in Colombo will direct the water management activities as a whole, and will function as the DAS coordinator for the project. A Water Management Engineer will be appointed, reporting to the Chief Engineer, who will be responsible for supervising the technical aspects of water management. The appointment of these two officers will be a condition of credit effectiveness. The officers will be assisted by an adviser on water management (para 5.19). The Deputy Commissioner (Minor Irrigation) and the Chief Engineer will continue to supervise the ongoing repair and maintenance program.  - 34 - 5.10 Selection of Schemes for Water Mlanagement. All tank schemes to be rehabilitated by the DI will in principle be considered for water management. In addition, DAS would identify appropriate working tanks in convenient locations to facilitate the overall efficient use of the water management staff (para 3.18). The cycle for approving the works involved in the latter schemes would be similar to that for the more substantial DI program. Final approval of the DAS program would be given by the Steering Committee before end-December, and funds would be released as appropriate in the new year. Construction would be supervised by the repair and maintenance wing and the DI quality control field staff would provide backup assistance. 5.11 Water Management Planning. Once a scheme has been identified for water management, the Agricultural Planning Team would prepare a plan following guidelines that have been prepared in draft and will be finalized by DAS and DA staff in cooperation with the water management adviser (para 5.19). Asso- ciated with the team in its work would be one DAS Tank Supervisor (TS), the local extension worker, the local cultivation officer and the irrigation headman (vel vidane). The findings and final designs of the teams would be transmitted through the Regional Engineer to the Deputy Commissioner (Water Management) for approval. The structural requirements and their location in the field would for rehabilitated tanks then be put forward to the DI through the office of the Regional Engineer for execution. Should other technical requirements for water management become apparent during the agricultural plan- ning cycle, they would also be submitted through the same office to the DI. 5.12 Implementation of Water Management. For implementation, the TS would be the focal point. Each TS would be allocated 10 to 15 tanks. After the investigation team has cleared and finalized a water management plan for a tank, and the Deputy Commissioner (Water Mlanagement) had given his approval to the proposed program, the TS would leave the team to implement the program. He would organize farmers groups and assist in the election of group leaders; lay out the field channel and drainage system; assist farmers in construction and setting prefabricated farm gates; place rain gauges and establish time schedule boards; and ensure that the water level gauge and measuring device are set appropriately. 5.13 Operation of Water Management. The operation of the system would be done by the same TS who had helped with implementation. His main task would now be to maintain the system of operation. To do this, he would follow a published schedule of visits to each tank. He would be responsible for advising farmers on the water supply system; operation of the sluices after rainfall; and utilization of saved water during yala. He would be available in case of any disputes which may arise during water shortages. During the early phase of the project, he would concentrate on implementing early cul- tivation in order to obtain an immediate impact. He would also assist farmers in maintenance. About 10 TSs would be supervised by one DAS TA Grade II officer, on a full time basis. The TA Grade II officer would be supervised by the Regional Engineer. In order to obtain easier supervision, tanks for any one TS mould be selected in clusters. Bicycles would be provided for each TS and motorcycles for TA Grade II officers. 5.14 The actual management and distribution of water would be decided by farmers. In the larger schemes, with rotational water supply (RWS), consider- able numbers of farmers would have to work together. This would be accomplished  - 35 - through the establishment of groups. Each group would consist of about 6 to 10 farmers. A basic criterion for forming a group will be the physical entity around one field channel. Each group would elect one farmer of the group as their group leader. All group leaders will be represented in a Tank Committee, chaired by the vel vidane. The group leader would be the focal point for individual farmers in case problems arise within the group's internal schedule. If the problem is due to a situation outside the group's influence, the group leader would raise his voice in the Tank Committee. Since water is provided on a rotational basis to each group, the group leader would help to ensure that each group member receives his share of water according to the pre-planned time schedule. 5.15 The operation of the system will be done by the vel vidane. He will operate the sluice and follow the rotational schedule as published on the display boards. Disputes in the allocation of water among farmers which cannot be solved by group leaders will be settled by him. The vel vidane will also be the vital link between farmers and government officers. In the case of smaller schemes, or where RWS is not introduced, farmer groups may not be needed and the vel vidane will be fully responsible for the operation of the scheme. 5.16 Repair and Maintenance. For the first two years after reconstruc- tion, the DI and DAS will be responsible for ensuring the satisfactory per- formance of civil works constructed by them. Normal maintenance will be the responsibility of the farmers, assisted by the TS. The latter would alert farmers to maintenance requirements, in particular to the need for preventive maintenance. The repair and maintenance wing of DAS, reporting to the Regional Engineer, will continue its regular program. For those schemes participating in the water management program, the visits of the TS will ensure feed back to the maintenance wing concerning emergencies or major repairs. To strengthen the capacity of the DAS to respond, the project would provide construction equipment which will be based at the new Regional Offices. It is hoped that over time, and once the current backlog in maintenance is overcome, the maintenance staff will increasingly turn to the implementation of the civil works required for the introduction of water management programs. Training 5.17 A crash program to provide training for 22 new TA recruits was implemented by DAS at Moratuwa University in 1980, and a further 100 will be trained during 1981. The training lasts only one year, compared to the two years previously provided for DI technical assistants. Once the current acute staff shortages are overcome, more substantial training programs leading to a full technical diploma will become the rule. For staff directly involved with the water management program, three levels of training would be provided at Maha Illuppallama: (a) Senior technical staff: two to four weeks with possibly a study tour abroad; (b) The Agricultural Planning Teams (agriculturalists and TAs) and TA Grade II (supervisors of the TS's): four weeks during the first year with regular follow-up refresher courses; and  - 36 - (c) Tank supervisors: four weeks during the first year, with refresher courses subsequently. 5.18 Arrangements would also be made to provide five day training courses to irrigation headmen (vel vidane) involved in water management programs. Short training courses of a few days would be given to DAS staff indirectly involved in water management: at Maha Illuppallama for Assistant Commissioners and Divisional Officers and in the districts for Cultivation Officers. Technical Assistance 5.19 A Water Management Adviser to be provided under the project, would coordinate the training program and help train the instructors from the Department of Agriculture in the techniques to be adopted under the program. He would assist in developing course material and in providing instruction at Maha Illuppallama. He would participate in the development of a few tanks near Maha Illuppallama as pilot projects to be used intensively for the training of tank supervisors. In addition, he would assist DAS in establish- ing the organizational framework for the water management program; refine the guidelines for implementing water management, prepare design manuals for the distribution system; and generally advise DAS staff. He would also provide assistance to the DA's physical resource evaluation (para 5.23). An assurance has been obtained from GOSL that the water management adviser would be employed by October 31, 1981 on terms and conditions acceptable to IDA. Terms of Reference for his assignment are available in the project file. For a period of two years, he would be full time on the project. At the end of this period, his assignment would be reviewed and a decision taken as to whether he should be extended full time for a third year or, alternatively, should return regularly during the remainder of the project period for a total of a further 12 months. Monitoring and Evaluation 5.20 A flexible approach to project implementation is recommended. Issues to be kept under review include: (i) the overall district balance; (ii) the methods of construction; (iii) the criteria for tank/anicut selec- tion; (iv) the implementation of the water management program; and (v) the impact of the project on agricultural production and its implications for further planning and implementation of village irrigation schemes. 5.21 Project Monitoring. The Project Management Division of the DI will be responsible for receiving and monitoring progress reports from the districts; summarizing and submitting monthly progress reports on the implementation of the DI civil works to the Project Steering Committee, and submitting quarterly reports to IDA. A project monitoring unit in the Water Management Division will have comparable responsibilities in DAS. 5.22 Agricultural Production Impact. The DAS in association with the Freedom From Hunger Campaign (FFHC) and the Department of Census and Statis- tics (DCS) will be responsible for monitoring the agricultural production benefits obtained. The FFHC has already developed and tested data collection procedures. Actual data collection will be undertaken by DAS COs, under the supervision of FFHC field staff. A base line survey of about 150 schemes  - 37 - scheduled for rehabilitation in 1981 is to be undertaken following the 1980/81 maha crop. These base line data will be followed up annually by a comparable survey of about 200-250 schemes. 5.23 Physical Resources and Hydrology. The Land and Water Use Division of the Department of Agriculture will undertake a physical resource evaluation study directed at improving the data base on the hydrology and irrigation potential of village tanks. For new village schemes the lack of data on small catchments could be a serious problem. Improvements to the data base may also suggest modifications to the storage/supply ratio included in the criteria for tank/anicut selection, and through adaptation of design criteria result in cost savings. The terms of reference for the proposed study are available in the project file. In addition to investigating the variability of rainfall runoff relationships, the study would: (i) conduct inflow, outflow and loss measurements on selected schemes in order to establish the magnitude of the tank and command area water balance components; (ii) classify the soils in the command areas and correlate these with the overall hydrology of the tank; (iii) evaluate crop and irrigation water requirements for paddy and upland crops; (iv) develop suitable water use models; (v) recommend cropping patterns appropriate for different conditions; (vi) review the accepted storage/supply criteria; and (vii) suggest minimum tank depth criteria. 5.24 Socio-economic Evaluation. The University of Peradeniya will under- take a regular annual socio-economic evaluation survey on a sample of about 20-25 tanks in order to evaluate the project's impact on agricultural produc- tion, farm incomes and employment. A base line socio-economic survey of about 1,400 farmers in 300 randomly selected tanks was completed in 1980. Further surveys would be coordinated with the physical resource evaluation. The study would also focus on the impact of rehabilitation on farm incomes as a basis for making recommendations for implementing cost recovery programs in the future. Terms of reference for the University's study are available in the project file. 5.25 An assurance has been obtained that MLLD will reach agreement, no later than December 31, 1981, with the agencies concerned (FFHC, DCS, UP and DA) to undertake the specified monitoring and evaluation studies under terms and conditions acceptable to IDA. VI. PRODUCTION, PRICES AND FARM INCOMES Agricultural Development with the Project 6.01 A combined irrigable area of about 31,500 ha (78,000 acres) under some 1,200 village irrigation schemes would benefit under the project from a better managed, more reliable water supply. In addition to higher yields and cropping intensities on existing irrigated land, it is expected that about 22.5% or 7,100 ha would represent land previously cultivated under rainfed conditions. This estimate assumes that most schemes will be small and working close to capacity, and that major design changes will be uncommon. Cropping intensities in the dry zone are low, being estimated to average only about 70% (irrigated and rainfed) in maha and 5% in yala. With the project,  - 38 - these are expected to rise to 85% (irrigated) and 20%, rates which correspond to those achieved on the more favorably placed working tanks. In addition, added security is expected to reduce the risk of crop failure from an average of 15% to 5% in maha and from 20% to 15% in yala. Cropping intensities in the intermediate zone tend to be significantly higher, being estimated at about 90% in maha (irrigated plus rainfed) and 25% in yala. These are expected to rise to 95% and 50%, respectively. In contrast to schemes in the dry zone, for which the main impact of rehabilitation will be felt in maha, the expec- tation is that benefits in the intermediate zone will primarily accrue from increased yala cropping. 6.02 Rice is virtually the only crop grown at present. The area under rice in maha is expected to increase from about 23,100 ha (of which about 5,000 ha is rainfed) to 27,200 ha with the project and in yala from 2,900 ha to about 7,500 ha. In total, the rice area receiving irrigation will there- fore rise by about 13,700 ha or 65%. On the lighter soils, some diversifi- cation is projected during yala, to the extent of about 2,000 ha of upland crops represented by chillies and cowpeas. The actual diversification achieved will reflect the success or otherwise of the water management program. More reliable water supplies, together with improved water management practices, the introduction of early maturing varieties, early sowing and other improved cultural practices are expected to result in an increase in average irrigated yields with the project from 2.25 tons/ha to 2.8 tons/ha in maha, and from 2.2 tons/ha also to 2.8 tons/ha in yala. In other words, yields are expected to rise from a relatively low level to those prevailing under the more suc- cessful currently operating tanks (para 2.17). Average yields of chillies and cowpeas which are representative of possible upland commercial crops are estimated at 1.2 tons/ha and 1.0 ton/ha, respectively. These are not expected to rise significantly during the project period due to the relatively small effort spent on research and seed production of these crops. At full develop- ment, crop production would amount to about 90,300 tons of paddy, about 1,000 tons of chillies and 850 tons of cowpea, compared to a current paddy produc- tion of about 46,000 tons. 6.03 Without the project, cropped areas and cropping patterns are expected to remain broadly constant. In practice, further deterioration of the physical facilities might be expected to reduce areas under irrigation, with a conse- quent increase in rainfed paddy or alternatively, the abandonment of command area. Some increase in average irrigated yields without the project is projected due to the progressive adoption of improved practices, with maha yields rising to an average of 2.6 tons/ha and yala yields to 2.5 tons/ha. Rainfed paddy yields are projected to remain constant. 6.04 Table 6.1 summarizes the overall agricultural impact of the project. The detailed assumptions are presented in Annex 5, Table 1. The estimates represent a reasonable response to the rehabilitation works themselves. Further improvements over and above those projected with the project are certainly possible. Much will depend on the selection of schemes for inclu- sion in the project and on the effectiveness of the water management programs implemented on selected schemes. The initial selection of schemes according to the agreed criteria will ensure that costs are contained in the early years. In the longer-term, the monitoring and evaluation studies proposed will increasingly provide a basis for modifying the selection process to concentrate on those schemes with the highest agricultural potential.  - 39 - Table 6.1: AGRICULTURAL IMPACT OF THE PROJECT X Increase Future Future With Project Present Without With Over Without Situation Project Project Project I. Command Area (ha) Rainfed 7,085 7,085 - - Irrigated 24,415 24,415 31,500 +29 II. Sown Area (ha) Maha: rainfed paddy 5,020 5,020 - Maha: irrigated paddy 18,095 18,095 27,170 +50 Yala: irrigated paddy 2,380 2,380 7,485 +162 Yala: irrigated upland - - 1,970 - III. Harvested Area (ha) Maha: rainfed paddy 3,855 3,855 - - Maha: irrigated paddy 15,950 15,950 25,805 +62 Yala: irrigated paddy 2,380 2,380 6,555 +175 Yala: irrigated upland - - 1,715 - IV. Production (M tons) Paddy 45,940 52,500 90,300 +72 Chillies - - 1,025 - Cowpeas - 850 - Marketing and Prices 6.05 Paddy. Recent developments have reflected GOSL's attempts to extricate itself from a costly rationing system designed to guarantee cheap food supplies to the population at large. Reforms introduced since 1978 have included the replacement of the ration system by food stamps. This has led to a major shift in rice marketing to the private sector, with the Government's role increasingly moving towards one of market regulation and away from distribution. Under this system, the Government stands ready to buy at the GPS price (effectively the floor) and--provided it has adequate stocks--sets the market ceiling at the price at which it sells imports. A recent decision to institute a public tender system should enable the Government to intervene at a price above the GPS price thus regulating the market without losing the efficiency gains from private marketing. 1/ 6.06 In November 1980, the GPS price was increased from Rs 40/bushel to Rs 50/bushel (Rs 2.4/kg) and a further increase to Rs 52.5/bushel occurred in early 1981. Since imports were being sold at a retail price of about Rs 6.0/kg in late 1980 (equivalent to about Rs 80-85/bushel of paddy) suffi- cient margin remains to allow active participation by the private sector. 1/ For a full discussion of the policy changes and their implications, see the 1980 CEM: "Sri Lanka: Key Development Issues in the 1980s," Volume II: Agricultural Producer Incentives" Report 2955-CE, May 29, 1980.  - 40 - Although imports are being sold on a cost plus basis, world market prices have risen significantly during the last year and are expected to continue to increase in real terms. The current economic price is therefore signifi- cantly above the GPS price even after the recent increase in the latter. For the purposes of the economic and financial analysis, it is assumed that this differential will narrow to reflect GOSL's policy of trade liberalization, but that farmers who sell primarily at the peak harvest will continue to obtain the GPS price (Annex 5, Table 2). To the extent that the GPS acts as a floor price, this is likely to result in a conservative assessment of farm incomes. 6.07 Chillies and Cowpeas. In October 1979, GOSL introduced a floor price scheme for eight minor field crops including chillies. The prices set were revised in April 1980 and the scheme was extended to include cowpea and green gram. The aim of the floor price scheme is similar to that for paddy--to stabilize the market and to provide a minimum guarantee to the farmer. The capacity of the Government to stabilize the market will depend essentially on the PMB's ability to purchase supplies during the harvest season, and on its flexibility in meeting changing relative supply positions among the different crops. A reduced PMB role in the rice market should release capacity for other crops, although a strengthening of the PMB's administrative organization and improved storage and marketing facilities will still be required. 6.08 Farm Inputs. To offset the impact of inflation, and to compensate for an unchanged GPS price at that time, GOSL increased the subsidy on fer- tilizer in September 1979 from 50% of the C&F price to 75%. With the recent increase in the GPS price, and in recognition of the heavy and growing burden on the budget, a policy decision has been taken to reduce the total level of fertilizer subsidies to Rs 1,000 M during 1981 and thereafter to consider phasing out the subsidy completely. For the financial analysis, initial price increases of 75-125% have been taken, and it is assumed the subsidy will be phased out by 1990 (Annex 5, Table 2). Price incentives for fertilizer use are currently--even after the recent price increases--very favorable but the effect on such incentives will clearly have to be kept under review in future price adjustments. Subsidies on other farm inputs are fairly modest. Cross-subsidization in-the petroleum market has kept diesel costs relatively low, but the diesel price was increased twice during 1980 and is now Rs 25/ gallon compared to Rs 13.5/gallon in early 1980. While an implicit subsidy may still exist, the effect on farm power costs is largely offset by import duties on import equipment. The liberalization of the trading regime has furthermore resulted in tractors and other privately marketed agricultural inputs being widely available, and--while their costs have increased relative to the GPS price for rice--physical availability appears no longer to be a serious problem. Annex 5, Table 2 summarizes the price assumptions used in the economic and financial analysis. Farm Incomes 6.09 The impact of the project on family incomes has been analyzed separately for representative farms in both the dry zone and the intermed- iate zones. Farm size in the latter tends to be significantly smaller than  - 41 - in the former, offset in part by the higher levels of cropping intensity. 1/ For the purposes of this analysis, an average paddy holding of 1.5 ha is assumed for the dry zone areas and 0.75 ha for the intermediate zone. Other representative fana sizes are illustrated in Annex 5, Tables 5 and 6. 6.10 Complicating any assessment of the project's impact on farm incomes are the wide variations in initial conditions from scheme to scheme, and the fact that paddy represents only one element in a complex agriculture. The with-project net value added per farm represents an upper limit in the impact on farm incomes for breached tanks or those which to all intents have been abandoned. The share of paddy in total farm incomes is equally variable, and the impact of rehabilitation on non-paddy activities is unclear. According to an analysis of more than 1,000 farm interviews by the University of Peradeniya, paddy accounted for about 60% of net farm incomes and 35% of total incomes in the crop year 1978/79. 2/ Despite substantial variations from district to district, no clear pattern emerged in relation to zonal characteristics. The relative share of paddy in total incomes is therefore based on these proportions in relation to the present situation, but income from non-paddy sources is assumed to remain constant in real terms following rehabilitation. 6.11 Table 6.2 summarizes the projected impact of the project on net farm incomes. While it is expected to have a significant effect on average incomes derived from the paddy tract (130% increase in the dry zone and 81% in the intermediate zone), its impact on total incomes is likely to be fairly modest given the importance of other sources of income. It is possible that, by competing for family labor and other resources, income from other sources will decline. However, this is likely to be offset by the greater stability created by a firmer paddy economy, and the associated promotion of stabilized upland and permanent tree crop cultivation. Of interest is the apparent 1/ According to the 1973 agricultural census, paddy holdings averaged 1.5 acres (0.6 ha) in the six districts of the project area characterized by a mixed zonal pattern (Annex 1, Tables 3-5) compared to 4.5 acres (1.8 ha) in the remaining districts lying fully in the dry zone. A comparable picture emerges also from the socio-economic base line study undertaken by the University of Peradeniya. 2/ The average structure of income per farm family in the University's survey was as follows (1978/79 prices): Rs/family % of Total Paddy 1,845 34 Other field crops 425 8 Permanent crops 590 11 Livestock 160 3 Total Farm Income 3,020 56 Non-farm Income 2,385 44 Total Family Income 5,405 100  - 42 - comparability of overall family incomes under the two representative farm budgets, since higher cropping intensities and a lower dependence on hired labor in the intermediate zone broadly offset the larger assumed paddy holding in the dry zone. This conclusion is broadly consistent with the results of the UP survey. 1/ 6.12 Even after the project, and allowing for a substantial proportion of income being derived from sources other than the paddy tract, average family incomes--even for larger paddy holdings (see Annex 5, Tables 5 and 6)--will remain low relative to those achieved for instance under the Mahaweli program. Assuming an average family size of 5.8 persons, per capita income for the ave- rage paddy holding will rise from about Rs 1,600 to Rs 2,200 (US$89 to US$122). These can be compared with an estimated average per capita income of about US$220 for the country as a whole. Table 6.2: PROJECTED NET FAMILY INCOMES /a % Increase With Future Future Project Over Without With Without ------- Rs -------- Dry Zone: 1.5 ha Income from Paddy Tract 3,140 7,215 +130 Other Farm Income 2,030 2,030 - Non-Farm Income 4,065 4,065 - Total Family Income 9,235 13,310 +44 Per Capita Income 1,600 2,295 Intermediate Zone: 0.75 ha Income from Paddy Tract 3,240 5,870 +81 Other Farm Income 2,095 2,095 - Non-Farm Income 4,195 4,195 - Total Family Income 9,530 12,160 +28 Per Capita Income 1,645 2,095 /a At an average of 1980 and 1990 prices expressed at 1980 currency values. See Annex 5, Tables 5 and 6. 1/ For instance, compare the characteristics of the two main districts surveyed Anuradhapura (dry zone) and Kurunegala (mixed zonal pattern) (1978/79 prices): Average Paddy Income from Holding Paddy Total Income Kurunegala 0.9 ha 1,950 5,730 Anuradhapura 2.04 ha 2,295 5,230  - 43 - CosL Recovery 6.13 It is estimated that total costs associated witht the provision of an adequate irrigation distribution system will amount to about Rs 2,225 per ha at 1980 prices (Rs 900 per acre), of which Rs 1,235 hiave been allowed for in the project cost table (shared equally between the provision of control structures and the earthwork associated with the main supply channel--Annex 2, Table 3). The balance represents farmer labor provided for the construc- tion of the field channel system. Farmers will also meet the majority of O&M costs, notably through the provision of labor for annual maintenance. It is estimated that on average, four cubes (400 cubic feet) of earthwork are involved per acre per year (the estimate provided by the DAS), of which one cube (Rs 125 per ha) represents the incremental cost that can be attri- buted to the project. In addition, farmers are expected to contribute an operational charge equivalent to half a bushel of rice per incremental cropped acre (the level set under the Agrarian Services Act to reward the vel vidane). On this basis, about 15% of the total present value of project costs will be directly contributed by the farmers (Table 6.3), equivalent on an annual- ized basis, and after allowing for returns to farmer labor and management, to between 9-11 of estimated project rent (Annex 5, Tables 5 and 6). Table 6.3: PRESENT VALUES OF PROJECT COSTS AND FAR1ER CONTRIBUTION Present Value at 12, Rs 11 investment Costs 265.4 Field Channel System 20.8 Incremental Maintenance 20.4 Incremental Operations 8.8 DAS Operations 20.4 Present Value: Total Costs 335.8 Farmer Contribution 14.9% 6.14 Given the relatively low incomes prevailing in the project area after rehabilitation, these rates of cost recovery are considered reasonable. No water charges are levied by GOSL on minor irrigation schemes, given their traditional form of ownership and management. 1/ It is not thought practica- ble to institute the collection of suchi charges, nor to introduce an alterna- tive cost recovery program, in the context of the VIR project in view of various factors, including: (i) their absence in the context of other village 1/ Under the Irrigation Ordinance of 1946, amended in 1968, GOSL is empowerec to levy charges only on major schemes (200 acres or more). Recent regula- tions have been issued designed to institute such collections, but per- formance has been mixed and the levels set are low (Rs 20-30 per acre depending on the intensity achieved). The Irrigation Ordinance is currently being revised and it is possible that the levels of water charges will be reviewed.  - 44 - irrigation programs (including those supported by IDA); (ii) the fairly limited impact the project is expected to have on average farm incomes; (iii) the variability of conditions and the inevitable uncertainties which will continue to be associated with minor irrigation; and (iv) the widely disbursed nature of the project area, the relatively large number of benefit- ing farmers and the substantial costs which would inevitably be incurred in administering any cost recovery program. The socio-economic evaluation study will, inter alia, review the actual impact of the project on farm incomes (para 5.24) and, subject to progress in collecting such charges on major schemes, progress in instituting other cost recovery programs in the context of major settlement projects (e.g. the Mahaweli program), and in the light of the revision of the Irrigation Ordinance, the possibility of instituting water charges in relation to benefitting minor irrigation schemes will be evaluated for the future. VII. BENEFITS AND JUSTIFICATION General 7.01 The project aims to increase agricultural production and farm incomes on existing--but underutilized--village irrigation schemes which have been allowed to deteriorate through inadequate maintenance and repairs over a number of years. The emphasis on rehabilitation and quick-yielding investments is appropriate in the light of Sri Lanka's evolving overall financial problems, and benefits from the project can be expected to flow as from the first year's investments. Some 20,000-25,000 farm families will have their incomes supplemented as a result of the project and will be provided with greater assurances against the risks of weather. -L. ect beneficiaries are widely dispersed throughout the country, in areas whizh frequently remain untouched by other government programs. The project should therefore help to reduce regional imbalances that may result from the emphasis on major new irrigation projects and reach some of the least advantaged communities in the country. The main benefits arising from the project are summarized below: Benefitting farm families 20-25,000 Benefitting command area 31,500 ha Increase in cropping intensity from 82.5% to 116.25% Increase in rice production 37,800 tons p.a. Increase in per capita income 43% Increase in farm employment 2.1 M mandays p.a. Increase in total value-added from Rs 140 M to Rs 250 M p.a. Employment Effects 7.02 Overall direct employment effects will largely arise from increased cropping intensities on land currently under irrigation and from the conver- sion of areas presently planted to rainfed paddy to irriated agriculture. At full development, employment would increase from about 2.7 M mandays to 4.8 M mandays per year, which is equivalent to approximately 8,000 additional full time jobs. Medium and larger farmers are expected to hire additional labor, mostly from rural workers currently underemployed, especially during the dry  - 45 - season. Construction of the civil works will also create a total of about 0.85 M mandays of employment during the course of project implementation, primarily for unskilled labor. Foreign Exchange Effects 7.03 Based on projected 1990 international prices ex-ressed at mid-1980 currency units, the annual gross foreign exchange value of incremental pro- duction of paddy generated through the project is estimated ac abouc Rs 183 M (US$10.2) at full development. Additional gross benefits equivalent -o Rs 22 1- p.a. (or 5% of the total) would be realized if the targets for diversifying into other yala crops achieved. Assuming that 50% of the latter would be in the form of tradeable commodities, and allowing for the foreign exchange costs of recurrent agricultural inputs (US$2.5 NI) and foreign exchange costs of the O&M costs of the irrigation svstem (US$0.1 M), the project's -denti- fiable direct foreign exchange impact would total about TJSS8.7 HI p.a. Environmental Impact 7.04 The project is expected to have a beneficial environmentai impact through its contribution to the stabilization of agriculture in the dry zone and through the provision of water supplies for livestock and domestic use. Benefits to the Economy and Economic Analysis 7.05 The benefits from the reiabilitation works to be carried out bt the DI have been compared with the tota investment costs of the project. On this basis the overall economic rate of return (ERR) would be about 20%. The economic cash flow for the total project is provided in Annex 5, Table 7 and the following outlines the principal basis for the economic analysis: (a) Project Life and Phasing. The project life has been taken to be 25 years, although adequate maintenance would ensure a longer working life. The project's phasing is given in Table 3.1. Since rehabilitation for each scheme will be completed within one year, benefits will start in 1982 and full production levels will be reached in 1991. (b) Prices. For internationally-traded goods, prices were derived from projected world market prices expressed in mid-1980 currency values. Appropriate adjustments were made for freight, handling, processing and quality differences. Local costs and locally valued benefits were expressed in border rupees, through application of appropriate conversion factors, the principal ones being a Standard Conversion Factor (SCF) of 0.90 1/ and a Construction Conversion Factor (CCF) of 0.92 (Annex 5, Table 3). To adjust for the seasonality of labor demand, a 30% discount has been made for off-peak labor, with economic wage rates converted from financial wages by a 1/ Official Exchange Rate (OER) = Rs 18.0/US$. Shadow Exchange Rate (SER) = Rs 20.0/US$. SCF = OER/SER = 0.90.  - 46 - conversion factor of 0.9 in the peak season and 0.8 in the off-peak season. Year-specific prices have been used for inputs and outputs over the 1980-90 period, after which 1990 prices were assumed to remain unchanged. (c) Benefits. Incremental crop benefits are based on the assumptions presented in paras 6.01-6.04. Some 75% of all schemes are assumed to lie in the dry zone. Increased crop security is reflected in an increase in the proportion of the sown area that is harvested. The increase in the cropped area is assumed to be obtained in the ratio 36% in the first year, 64% in the second, 84% in the third and 96% in the fourth, with a corresponding build-up schedule for yields applied to the increase in cropped area. Yields without the project are assumed to rise at a constant rate over a ten year period. (d) Investment Costs. Average rehabilitation costs per scheme (Rs 300,000 approximately at 1980 prices) were based on an analysis of preliminary DI estimates for 375 schemes (Annex 2, Table 2), including an additional 10% physical contingency allowance. A construction conversion factor of 0.92 was applied to this cost, reflecting the assumed weighted average share of unskilled labor (41%), traded items, excluding fuel (17%), fuel (8%) and non-traded items (34%) (Annex 2, Tables 3). The CCF for the field channel distribution system and for the incremental O&M of completed schemes (Rs 250/ha at financial prices) were converted at a conversion factor of 0.87 to represent the labor intensive operations likely to be involved (Annex 5, Table 3). Other project costs were valued at their finan- cial values, with local costs converted at the SCF. (e) Other benefits are not susceptible to normal techniques of cost-benefit analysis but include: (i) the creation of a systematic approach to water management which could be extended beyond the schemes rehabilitated under the project; (ii) the residual value of equipment and the creation of a regional quality control organization in the DI; (iii) the use of tanks for domestic water supply, livestock, fisheries and perhaps other uses; and (iv) a contribution to the over- all stabilization of agriculture in the dry zone (para 1.14). Project Risks and Sensitivity Tests 7.06 The overall scope and phasing of the proposed VIR project has been adjusted in line with the Government's allocations for the project established in January 1981. Nevertheless, it is conceivable that financial stringency will entail further adjustments to the project allocations in the future. The divisibility of the project is such that this could be readily accommodated through a reduction in the number of schemes rehabilitated in any one year. Since this would reduce costs and benefits virtually pro rata (given the  - 47 - weight of the civil works component in total costs), it would lead to a corresponding reduction in the project's NPV but would leave the ERR virtually untouched. The economic viability of the project would therefore remain secure, although the contribution of the project to net incremental value added to the economy might be delayed. 7.07 In common with similar rehabilitation projects elsewhere, the invest- ment costs are small relative to the total flows of gross benefits and recur- rent costs. The project is therefore relatively insensitive to the costs of rehabilitation, but relatively sensitive to the assumptions concerning the difference between the with and without project cases. The latter is demon- strated in Annex 5, Table 9. For instance, if average cropping intensities increase from 82.5% (including rainfed paddy) to 108% rather than 116%, then the ERR falls from 19.8% to 14.7%. The variability of conditions inevitably involves the risk that some sub-optimal schemes will be selected despite meeting the agreed criteria, while uncertainties are also attached to such factors as the effectiveness of the proposed water management programs, externalities in relation to the impact on neighboring tanks and the optimum timing of rehabilitation for any particular scheme. Nevertheless, it has been argued that if the water management programs prove successful then the benefits projected for the project could be greatly exceeded, while the cost limits set in the criteria for scheme selection (Annex 4) provide an important safety factor to ensure that most schemes selected remain economically viable. Switching values for individual and total cost and benefit streams are provided in Annex 5, Table 8 and the table also includes PWs (Present Worths) for total cost and benefit streams with their breakdown into relevant categories. The table shows that any particular minor scheme could accept a very substantial increase in the costs of rehabilitation (+66%), provided the expected average increase in agricultural production is achieved while still earning an accept- able economic return. 7.08 The project's economic viability was also tested for its sensi- tivity to a number of other key assumptions, including prices, yields and totJl command area. The results are tabulated as follows, together with the corresponding impact on the ERR: Change ERR