Ceylon - Lift Irrigation Project

Report No. 5634 The World Bank Impact Evaluation Report Sri Lanka Lift Irrigation Project (Credit 121-CE) May 8,1985 Operations Evaluation Department FOR OFFICIAL USE ONLY Document of the World Bank 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. FOR OFF ALUS ONLY THE WORLD BANK Washington. D.C. 20433 US.A. Offie of Direcor-Cuwral Operaion Evakuitn May 8, 1985 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Impact Evaluation Report - Sri Lanka Lift Irrigation Project (Credit 121-CE) Attached, for information, is a copy of a report entitled "Impact Evaluation Report - Sri Lanka Lift Irrigation Project (Credit 121-CE)" prepared by the Operations Evaluation Department. Attachment AThis document has a restricted distribution and ay be used by recipients only in the Verformtanc of their official duties. Its contents may not otherwise be disclosed without World Bank authori;ation. IMPACT EVALUATION REPORT SRI LANKA LIFT IRRIGATION PROJECT (CREDIT 121-CE) TABLE OF CONTENTS Page No. Preface ............................. Lessons of Experience ........................................ iii I. INTRODUCTION .....:..................................... 1 . Purpose of the Study .................................. 1 The Project .. .......... o ......................... 1 Major Conclusions of the PPAR .......................... 4 II. IMPACT ON PROJECT FARMERS ............................... 6 General .. ....... o... . ........... *... ........ 6 Diversity of Project Sites ............................. 8 Farmer Responses ...... . ......... ................. 12 Project Assumptions and Outcomes ....................... 13 Farmer Incomes .................. .....................0. 14 Impact on Employment ... .............................. 17 Conclusions ............ ........0..............0.*.... 18 III. TECHNOLOGICAL IMPACT ...........................e. ..... 19 Modern Input Use ........... ..... ......-.............. 19 Husbandry and Water Management .......-...o............ 21 Conclusions .............. ... ..... .......... ..... ...... 22 IV. IMPACT ON LOCAL INSTITUTIONS .................o........ 22 Water Users Associations ...... .............. 22 Government Institutions ........................ ........ 24 Conclusions .... .* ....... ........ 26 V. PRODUCTION IMPACT ... ............ .*.....*...... 26 Cropping Intensity .................................... 26 Productivity - Yields .................................. 28 Prices and Costs .............o........0.......... .... 28 Water Charges ................................ .... 29 Economic Justification ............................... 30 Conclusions .......................................... 30 Annex I - Economic Rate of Return Calculation Cost Benefit Stream .................................... 33 IMPACT EVALUATION REPORT SRI LANKA LIFT IRRIGATION PROJECT (CREDIT 121-CE) PREFACE This is an impact evaluation of the Lift Irrigation Project in Sri Lanka for which Credit 121-CE was approved in June 1968 in the sum of US$2.0 million. The credit was closed, after a cancellation of US$65,000, in December 1977. An OED mission visited Sri Lanka in August 1979, and a PPAR No. 2801 was issued on December 28, 1979. The report concluded that the project had so far failed to achieve its main objectives, and due to uncer- tainties about the future production of high value crops, a positive return to the economy was doubtful. It discussed the difficulties encountered in securing adequate water supplies, coordinating the work of the various agen- cies and organizations involved in project implementation, collecting water charges to safeguard proper maintenance and replacement of pumps/engines, as well as the farmers' reluctance to fully make use of project-provided facili- ties. The impact evaluation has been undertaken about eight years after the physical completion of most of the project works and five years after the performance audit. Special interest is attached to this project because of its pilot nature in Sri Lanka as the first irrigation project supporting high value crops. An OED mission visited Sri Lanka in September 1984. The mission discussed the impact of the project with Sri Lanka authorities. Basic infor- mation, through a farm sample survey, that covered all sub-project areas and comprised interviews with 300 farmers representing 5% of project partici- pants, was provided by a Sri Lanka consulting firm. Because of time and resource limitations, this study did not cover the secondary effects of the project. A copy of the draft report was sent to the Borrower on December 7, 1984 for comments, but none were received. The mission gratefully acknowledges the valuable assistance of the Irrigation Department, the Research and Extension Department and all other contacted Government offices, their staffs, and the participating farmers met during the field trips. IMPACT EVALUATION REPORT SRI LANKA LIFT IRRIGATION PROJECT (CREDIT 121-CE) LESSONS OF EXPERIENCE Conclusions 1. The successes and failures of the Lift Irrigation Schemes (LIS) program demonstrate that lift irrigation can increase farmers' incomes sub- stantially under certain circumstances, and increase the supply of high value crops, in this case chillies and onions. The basic concept of providing irrigation to the uplands was soand. Most itwortant, however, was the demon- stration that most of the conditions under which the project might have been viable were not met during or after project implementation. It is now doubt- ful whether appropriate provisions will be made to sustain the program, even in areas where it has had a positive impact and continues to show promise. 2. Overall, the LIS was a failure. It was conceived as a crash pro- gram, prepared without an adequate data base about farmers, conditions or institutions involved. It was rushed into implementation without reference to the Sri Lanka experience of lift irrigation schemes already in existence, and without a proper appreciation for the role of lift irrigation within the context of settlement schemes. Most importantly, it was only partially implemented; it failed to do all it was supposed to do. 3. First, and most critically, the LIS failed to provide irrigation water adequately, dependably or equitably. Systems were underdesigned; water supply was inadequate or irregular; canal deliveries were not synchronized with the requirements of lift systems; pumps broke down and took too long to repair; and water distribution was poor. These shortcomings were due to design and implementation deficiencies, some of which were overcome in time, but not until many farmers had justifiably lost confidence in the schemes. 4. Although water was the prime resource that the project was designed to supply, very little had been planned to ensure its economic use. The water management components of projects should receive at least as much attention as the implementation components and certainly deserve more atten- tion than is evident in this project. 5. No serious attempt has been made to comply with one of the cove- nants of the Credit Agreement that farmers should ultimately shoulder the cost of operation and maintenance of the water supply and cost of replacement of pumps and engines. At present, participants are obliged to bear only the cost of fuel used for pumping and no funds are available to replace the pumps and engines which have at best one or two years left before becoming unser- viceable. 6. The high degree of support from GOSL institutions expected for the successful implementation of the project did not materialize. The legitimate function of the implementing agency, the Irrigation Department, was only the construction and operation of the water delivery system; other agencies with equally'prominent roles to play in a primarily agricultural project were not closely involved in project formulation or implementation. At project level, consultation with district officials, who would be responsible for the direct administration of the schemes, was insufficient. Even after the need for closer coordination at district level had been pointed out at audit, no mean- ingful steps have been taken in the setting up of such coordinating units. 7. Agricultural extension services were less than adequate and GOSL has not met its commitments to the project in this regard. More active and properly supervised extension services would have improved both the extent and the productivity of these crops in a shorter time except at Vavunikulam, where the technology of chillie and onion cultivation was well known. 8. The project design had made no provision for possible ineligibility of participants for institutional credit as a result of earlier defaults. This ineligibility in many cases excluded farmers from participating in the LIS. Establishment of farmer organizations which might have overcome this shortcoming did not materialize. In some schemes project activities have been maintained by recourse to Government funds - a remedy which in the long run would tend to increase farmers' dependence on Government support. 9. Since the appraisal of the LIS, both external and internal changes have created conditions adverse to the project. The escalation of fuel prices, input ccats and labor costs have reduced project benefits, while the development of consumer interests has kept the prices of both chillies and onions below appraisal expectations. 10. As against these adverse factors, the technology of chillie and onion cultivation has begun to spread within the project areas and limited export markets have begun to develop for green chillies. Thus, although the prospects for this project cannot be said to be bright, lift irrigation in general will continue to be used by farmers for the cultivation of chillies and onions and other high-value crops. In such cases irrigation water is likely to be used to supplement rainfall or gravity irrigation, and lifting mechanisms are likely to be individually controlled. 11. Farmers participating in the project enjoy incomes which are dis- tinctly above those of non-participating smallholders. The project also suc- ceeded in introducing modern technology to the project farmers specializing in chilli-onion production. Due to modern input use chilli yields reached an average of 1000 kg/ac which compares favorably with the appraisal assumption of only 500 kg/ac at full development. These results were obtained when water supplies were reliable, where farmers were properly trained, markets existed or were developed and where capital or credit were available. If these conditions are met, benefits are tangible. - iv - 12. In summary, the most salient problems in the LIS program have been technical and administrative, rather than social. Where reasonable condi- tions prevail, which provide clear opportunities to farmers and expose them to minimal risks, farmers have responded positively. Recommendations 13. Prior to site selection, an assessment should be made of capabili- ties and constraints of farmers and local institutions and consideration needs to be given to the settlement history of prospective target groups. This can be done by a skilled investigator through a series of brief reconnaissance visits to prospective sites; it need not require a full-blown socio-economic survey. Based on this assessment, each project site could be evaluated for its appropriateness for inclusion in the project. Such minimal data gathering and analysis should lead to the elimination of inappropriate LIS sites. 14. Institutional responsibilities and relationships should be speci- fied clearly and reasonably during appraisal, including manpower and resource commitments, based on an understanding of local bureaucratic patterns. This may not solve all problems, but it could eliminate many unnecessary and self- defeating calls for cooperation, coordination, and so on, at a later stage when there is little hope of recommendations being accepted and implemented. 15. Project components need to be reviewed more carefully to ascertain their appropriateness for the intended target groups. In irrigation systems, the design should ensure that prospective recipients will get enough water without having unrealistic efficiencies imposed on them while they are learning to use irrigation water. Design parameters should respond to immediate needs, not to theoretical future use when water management might be fully mastered. Clearly many LIS problems resulted from inappropriate assumptions which could have been scrutinized during appraisal for their social as well as technical feasibility. 16. The choice of a lead agency has profound implications for the implementation of a project and its ultimate impact. The LIS project was an upland agricultural development project, not merely an irrigation project. However, it became a defective irrigation project in implementation because the Irrigation Department tended to treat the project as a residual responsibility which conflicted with other departmental priorities. This is most clearly seen in discussions which continue to occur regarding LIS schedules vis-a-vis maintenance schedules for gravity canals. It is also seen in other elements of field implementation, such as the down-playing of the role of allied agencies. The choice of a different implementing agency could have led to an emphasis on the broader objectives of project as a whole, not merely on one aspect, such as construction. 17. Supervision recommendations mst be clear, reflecting government's agreement or disagreement, and specifying steps that are to be taken to achieve them. If action is not forthcoming on the part of responsible authorities, the recommendations should be rescinded or further steps taken to enforce them. LIS supervision reports are replete with vain demands for extension assistance, provision of credit, and so on, but little indication of concrete responses. If recommendations are important, they should be followed up. If not, they should be dropped. 18. Consistent with the above, some thought =ust be given to establish- ing procedures to make supervision more effective. The LIS fell apart during implementation; many necessary programs never materialized, they were late in coming, or were ineffective. The supervision process missed opportunities to see the project as a whole and to help develop mechanisms to achieve desired results. 19. With regard to the project, Government should reduce its responsi- bility for continued support for the individual schemes by increasing its efforts to promote suitable farmer organizations in each scheme. These organizations, which might require some initial support from the respective District Development Councils, should be geared to take over initially all responsibility including the replacement of pumps and engines. IMPACT EVALUATION REPORT SRI LANKA LIFT IRRIGATION PROJECT (CREDIT 121-CE) I. INTRODUCTION Purpose of the Study 1.01 This study tries to assess the major impact of the Lift Irrigation Project on the farmers who participated in the project, especially on their incomes, their technology improvement and the employment situation, as well as on the institutions responsible for project implementation. Insufficient information is available to undertake a meaningful assessment of the secondary effects generated by expenditures of incremental farmer incomes. Special attention has been given to the project's economic impact and sustainability. The Project 1.02 The project was IDA's first and the Bank Group's fifth lending operation in Sri Lanka (then Ceylon). It was identified by a FAO Cooperative Program (CP) Mission which visited Sri Lanka in October/November 1965. The Government prepared the project in early 1966. The project proposal was then reviewed by another CP mission in June/July 1966 which found that further studies were required. The final project proposal and the request for financing were submitted to the Bank in June 1967. 1.03 The project proposed envisaged the provision of lift irrigation and agricultural development for 13,500 acres, of which 6,100 acres of uplands in existing settlement schemes of formerly landless laborers, and 7,400 acres alotted for settlement of unemployed, educated single youths. A Bank mission appraised the project in September/October 1967. The mission recommended a reduction of the youth settlement component to a 400 acre pilot scheme. Credit negotiations for the reduced project took place in April/May 1968 and Credit 121-CE was approved by the Board on June 4, 1968, signed on June 9, 1968 and became effective on August 5, 1968. 1.04 The project was to construct irrigation channels and provide pumps, pumphouses, and water delivery structures to irrigate about 6,500 acres of upland in four non-contiguous areas in the so-called "dry zone" of Sri Lanka. Water was to be supplied to individual farms by pumping from existing gravity irrigation canals which were already supplying paddy fields. The estimated cost of the project was US$3.3 million equivalent, for which an IDA credit of US$2.0 million was approved. The Irrigation Department (ID) was responsible for project implementation. The project was expected to be completed in mid-1972. -2- 1.05 The main objective of the project was to support the Government's policy to attain self-sufficiency in food crops. It was designed to increase production of upland high-value cash crops, particularly chillies and onions which play a signficant part in Sri Lankan diets, and which at the time of appraisal were mainly imported at an annual cost of over US$11 million equiv- alent. At the same time it aimed to improve the incomes of the small farmers in the project area. 1.06 The project was to be located in four of Sri Lanka's major coloni- zation schemes, some of which had been in operation since the 1940's. In each of these schemes, colonists had been given land units of approximately 5 acres consisting of 3 acres of irrigated lowland for rice cultivation and 2 acres uplands. The lift irrigation project was to provide irrigation water to one acre of the high land for the cultivation of supplementary food crops (SFC). Included were also two more recent settlements, Youth Settlement Schemes, which did not provide irrigated lowlands; in these schemes each youth was to receive water from the lift irrigation project for cultivation of 2 acres of upland. 1.07 In the early stages of implementation, one of the schemes (Nahakan- darawa, 1,000 acres) was replaced by a more recently established colonization scheme (Nagadeepa, 1,200 acres) and part of the area coming under the Polon- naruwa scheme was also abandoned. To compensate for these changes the area in Rajangane was increased by 332 acres, thereby maintaining the total proj- ect acreage of 6,500 acres. 1.08 Project cost was estimated at appraisal at Rs 19.6 million (US$3.3 million equivalent at the 1969 rate of exchange of Rs 5.90 to the U.S. dollar). The breakdown of costs was as follows: Foreign Local Exchange Total Costs Component Costs (USS'000 Equivalent) Pump Procurement and installation 160 450 610 Irrigation System 620 285 905 Land Development 650 30 680 Housing for Youth Settlement 60 - 60 Vehicles and Tractors - 200 200 Contingencies 20% 300 195 A05 Engineering, Administration and Agricultural Specialist 110 50 160 Operation and Maintenance during initial four years of project life 120 70 190 Total 2,020 1,280 3,300 -3- 1.09 The IDA credit (121-CE) of US$2.0 million (then Rs 11.9 million), covered approximately 60% of estimated total project cost. The balance of the cost, totalling Rs 7.7 million, was to be met by GOSL. 1.10 Project implementation was scheduled to start in February 1969. However, neither the start-up nor the subsequent phases of implementation were able to match original expectations. In fact by June 1973, the original closing date of the project, only about 55% of the credit had been disbursed and the project was 2k years behind schedule. 1.11 Among the several initial setbacks which hampered implementation, the most prominent were the following: (i) the Irrigation Department was initially not able to allocate sufficient supervisory personnel and vehi- cles to the work sites; (ii) there were delays in ordering, delivering, and assembling equipment due to the general unrest before and during the elections of May 1970; (iii) with the change of Government after the elections, personnel were transferred and responsibilities for projects were reallocated between Government depart- ments. This resulted in a general slowing down of the work; (iv) the insurrection of April 1971 seriously dislocated work on the project; (v) incomplete land levelling and limited impact of agricultural extension slowed down the pace of agri- cultural development work in the project. 1.12 Following the completion of all works, land utilization lagged behind expectation. The principal reasons for this were: (i) an agronomist was not appointed to the project as agreed at appraisal; (ii) persistent droughts between 1973 and 1976 led to a short supply in irrigation water and the farmers wanted the available water for paddy cultivation; (iii) initially pumps broke down frequently, repairs were slow and therefore water supplies were not reliable; (iv) agricultural extension services were ineffective and poorly coordinated; (v) production credit was not made available for most beneficiaries because they had not repaid previous loans obtained for paddy cultivation; -4- (vi) there was inadequate coordination between the Irri- gation Department which managed the irrigation sys- ten in the Project and the Department of Agriculture which advised on agronomy; (vii) except in the Vavanikulam scheme, the settlers had experience and expertise only in rice and slash and burn cultivation; (viii) at appraisal, the crop intensity for areas irrigated by the Project was expected to be 133%. At project completion, cropping intensity was only 28% even after excluding Nagadeepa. 1.13 All civil works were completed in 1976 and the IDA credit was finally =losed in December 1977, almost 5 years behind schedule. As a result of cost escalation, the Government's contribution increased from US$1.3 mil- lion to US$3.1 million. 1.14 The cost increase was due both to price escalation during the extended implementation period and to changes in construction, especially the need to enlarge the lined channels as well as to increase the number of pumps; the cost of the water delivery system turned out to be three times the original estimate. The overall project cost increased by 79% from the Rs 19.6 million estimated at appraisal to Rs 35 million. Major Conclusions of the PPAR 1.15 The project audit at the end of 1979 noted th-t project performance had fallen behind appraisal assumptions, but optimistically envisaged improvements in project activities in the future. Part of the likely improvements was to come from the project farmers themselves, such as better water management and expanded cultivation, while at the same time performance of the implementing agencies was also expected to improve. The PPAR also expressed the hope that certain natural conditions such as poor ra-nfall as during the 1973-1976 period, and initial technical problems, such as frequent breakdowns of pumps, would not continue to inhibit the project in the future. 1.16 Specifically the audit criticized the fact that the actual field water requirements for chillies and onions had far exceeded appraisal esti- mates and additional pumps had to be installed to meet this demand. Under improved water management conditions (which were expected to result from the IDA-supported Tank Irrigation Modernization Project), it was envisaged that irrigation requirements would be close to appraisal estimates and that the extra pumps supplied would no longer be required for the project. 1.17 It was also noted that by 1978 only about half the target acreage of 6,500 had been brought under irrigation, and of this, less than 1,500 acres were being cropped in chillies and onions. Thus, against a planned cropping intensity of 133%, the intensity achieved in the 1977 dry season, excluding the Nagadeepa project, had been only 28%. -5- 1.18 Further comments were made on the inadequacy of agricultural exten- sion which affected adversely project performance, and the hope expressed that the National Agricultural Extension Project to be establishedl/ would be able to provide all areas with the necessary technical advice. 1.19 The PPAR referred to the assurance obtained from GOSL that an annual water charge would be levied on all project farmers and that this would be gradually increased so that by the 5th year after completion of con- struction, all operational and maintenance costs, and pump and engine replacement costs would be collected from individual farmers. The initial levy was estimated in 1968 at Rs 135 per acre and by 1972, this estimate had been revised with the Irrigation Department (ID) concurrence to Rs 300. At the time of audit in 1979 the ID had raised this estimate to Rs 700; together with the replacement component the full water charge was estimated at Rs 1,000/acre annually. 1.20 The PPAR recognized the need to modify the original project design - which catered exclusively to the production of chillies and onions - and to extend it to the cultivation of other SFCs, as well as to tree crops. It also reduced the production target of chillies from 65,000 to 54,000 cwt and of onions from 200,000 to 45,000 cwt. The acreage under chillies was reduced at the time of audit from 6,500 to 4,650 and that under onions from 2,200 to 445 acres. 1.21 Highlighted was the lack of institutional coordination which had not been envisaged at appraisal and which had been seriously detrimental to farmers' activities. A coordinating committee had been set up but no mean- ingful progress had been made up to the time of audit in resolving inter- agency problems. At project level the PPAR had recommended the establishment of a Task Force in each district consisting of the Government Agent and Dis- trict heads of the Irrigation and Agricultural Departments which was to be responsible for the promoting and implementing of irrigated agricultural development. 1.22 The audit gave prominence to the fact that an IDA mission had obtained several assurances from GOSL in June 1979: (i) In the Bakamuna Youth Scheme (Polonnaruwa project) it was agreed that the balance of 36 settlers would be selected in time to commence cultivation in January 1980. This agreement has been fulfilled. (ii) At Nagadeepa, it had been agreed that LIS water could be used to grow tree crops, particularly banana, papaya and citrus. It was hoped that the remaining acreage would be taken up with tree crops in a period of about 2 years. 1/ Agricultural Extension Project, Credit 931-SL, approved on June 12, 1979. Date of effectiveness October 4, 1979. -6- (iii) In Polonnaruwa, it had been agreed that water charges at least to cover O&M costs would be levied from the beginning of 1980 - the charges being levi- ed from all farmers who had access to water irre- spective of whether they used it. 1.23 Relying on these assurances and indications of growing farmer interest, the PCR assumed that full development would be reached until 1989 (paras. 4.06, 4.07). This optimism was not shared by the audit (para. 21) which based its reservaticas on the prospects of full development on: (i) farmers' reluctance to participate in the past years; (ii) considerable reduction in the producer price of chillies and onions; and (iii) lack of adequate supporting services. 1.24 In the following section the socio-economic aspects of the project approach will be highlighted to explain the disappointing farmer participa- tion in the lift irrigation scheme. IV. IMPACT ON PROJECT FARMERS General 2.01 Three of the four project areas were visited during the mission: Polonnaruwa, Rajangana and Nagadeepa. Vavunikulam was not visited because of recent civil disturbances in the general area. The field visits revealed very clearly that conditions vary considerably within sites, as well as be- tween sites; problems and responses also vary accordingly. However, the most significant feature common to all the systems is their status as settlement areas. Each Lift Irrigation Scheme (LIS) site is located on a government settlement scheme which was intended to provide land for landless farmers and to halt chena (shifting, slash and burn) cultivation. 2.02 Settlement programs have been carried out in Sri Lanka for over 30 years. Originially, farmers were moved from the densely populated wet zone to the dry areas of the island. More recent settlers are relocated from nearby parts of the dry zone. Often they are the landless offspring of orig- inal settlers or squatters who moved to the periphery of the first settle- ments. The project sites include both new settlements established about the same time as the LIS (Vavunikulam, the Youth Schemes in Polonnaruwa, and Nagadeepa) and more established sites (Polonnaruwa and Rajangana). Although the length of settlement has implications for the differential response of farmers to LIS, as indicated below, target farmers share a number of influ- ences by virtue of their participation in the settlement schemes: -7- (1) Government inolvement is pervasive in many aspects of their lives. Settlers are under the jurisdiction of colonization and rural development officers, among others, toward whom they establish dependency relationships which are rarely broken. Responsible officers may be feared, resented or admired, but they cannot be ignored because their directives con- tinue to affect settlers long after the schemes are established. (ii) Settlers within a given scheme each receive a simi- lar allotment of land (a house plot and upland and/ or paddy land) and a similar array of services. However, their actual endowments vary greatly in terms of the quality of the land, location of plots, access to irrigation and drainage, and so on. More- over, the arrangement is fixed at the time of set- tlement and rarely can be changed. Thus basic ine- qualities are built into the new communities and continue to flourish. (iii) With the possible exception of Vavunikulam, most of the settlers are landless chena cultivators who must learn through experience to become good irrigation- based settled farmers. In early stages, they can be expected to be neither skilled nor efficient in the use of irrigation water. Thus the limits of a water delivery system will become immediately apparent in head/tail discrepancies - a fact that is rarely taken into consideration in system design. (iv) Depending on the amount and quality of site prepara- tion already undertaken before their arrival, set- tlers have a lot of work to do before they can actually become farmers. Consequently, the ability of farmers to respond to innovations which cost money or require significant labor inputs may be limited by labor demands for clearing or opening land, by low incomes due to incomplete land utiliza- tion or by low productivity as their farming systems are being established. Depending on the administra- tion of the system, the use of food and income sup- plements during the establishment period may just as easily retard the process as speed it up. (v) Even though some settlers may share some social bonds prior to reaching the site, it takes time to establish social networks and exchange patterns which form the base for patterns of collective -8- behavior on a community level. The degree of homo- geneity and the nature of prior collective experi- ences of the settlers greatly affects community for- mation, but it is also affected by the magnitude of demands made on settlers to secure the survival of their indivIdual families. Until families are really established, new community bonds are hard to form and collective action is rare. 2.03 It is essential to identify a few general points which relate to all of the schemes and significantly influence the choices of farmers. First, all of the systems have suffered from uncertain or very limited water supplies either because of shortages at the source, pump breakdowns, schedul- ing problems, poor cooperation between farmers, wasteful water use, or com- binations thereof. The frequency and magnitude of the shortages varied from one system to another, but none has been immune. Moreover, some of the worst shortages occurred at early stages of the project, when farmers were most vulnerable to the effects of crop failure. Second, intensive chilli and onion cultivation is an expensive proposition. The crops require substantial capital and labor inputs which are beyond the capability of small subsistence farm households, especially in early stages of settlement. And it is poten- tially disastrously expensive to such households when irrigation water sup- plies are uncertain. Third, the costs of production have gone up and chilli prices have gone down, since appraisal and audit. Consequently, although chillies and onions are still among the most lucrative crops farmers can grown, they are not as lucrative as they were several years ago. Diversity of Project Sites 2.04 The economic and social impact of the project varied greatly from one site to another. Implementation details have been presented and analyzed in the PPAR, thus this discussion confines itself primarily to the period since the project audit, except where necessary. 2.05 Vavunikulam (500 acres). This has been the only scheme which developed as envisaged in the Staff Appraisal Report (SAR). Farmers came from Tamil areas in the north, where intensive vegetable cultivation is traditional, and where local lift irrigation is not uncommon. They responded immediately to the availability of LIS water by planting chillies on most of the irrigated land. 2.06 When the farmers were compelled to purchase fuel for LIS pumps, starting in 1981, onion production started to increase markedly, because of the combination of increased input costs and decreased market prices for chillies. The farm sample survey revealed that Vavunikulam farmers are con- cerned about input costs, market prices, and market demand in descending order of priority. Of all systems, the survey in its interviews found the pronounced awareness among Vavunikulam farmers of the existence of farmers' organizations prior to project implementation, as well as afterwards. In -9- addition, they were most familiar with agricultural extension efforts and had the highest incidence of input use on all crops. 2.07 Without any possibility of a site visit, it is difficult to deter- mine whether the decline in LIS utilization in Vavunikulam is a temporary or permanent phenomenon. Likewise, it is difficult to assess to what extent the relatively active involvement of farmers in LIS is due to their prior experi- ences, ethnic factors, community cohesiveness, to the clear presence of extension and other support personnel, or to combinations of the above. 2.08 Rajangana :Left Bank = 1870 ac; Right Bank - 1830 ac). This is the largest LIS scheme and the most interesting for a number of reasons. It con- tains a mixture of settlers, some are locals, others are from more distant areas; most are legitimate settlers, but a number are squatters who settled on peripheral areas or took over plots abandoned by the original settlers. Many of the local settlers continued to work on paddy plots in their original villages, in addition to their new holdings, thus they had no time or inter- est in subsidiary food cropping until chilli production was well established on other farms. - 2.09 Rajangana consists of two very different units. Farmers on the Left Bank are utilizing the LIS, but the Right Bank is barely operational, although system limitations have been overcome since the inception of the project. During the early years, water supplies were too limited and unreli- able. In later years, however, the Mahaweli Project diverted water to the Rajangana Tank. LIS use has subsequently increased, especially since 1980. 2.10 Chilli production is now well established on LIS plots on the Left Bank, and there are indications that interest is spreading to the Right Bank. Some of the most visible, entrepreneurial chilli growers on the Left Bank are squatters or hidden tenants who manage the lands of other settlers. The availability of credit, markets and fairly dependable water supplies fully have had a cumulative effect. 2.11 In contrast, farmers in the two functioning pump units on the Right Bank are only now cautiously starting to get into chilli production, primari- ly in a mixed vegetable/fruit regime. Without access to credit for fuel, they must expand their operations a little at a time as they try to accumu- late capital from previous years' harvests. 2.12 The commercial success of the export program on the Left Bank has prompted many chilli growers to depend on hired labor, primarily female day labor, for much of the work. This has brought new employment opportunities in the area. No employment impact is evident on the Right Bank however. Families without either capital or available labor essentially do not engage in chilli production, and most of the pump units cannot begin to function until channels are repaired either by Right Bank farmers themselves or the irrigation department. - 10 - 2.13 Polonnaruwa (Youth Schemes - 400 ac; Colonist Schemes - 450 ac). Polonnaruwa also contains-a mixture of varied elements, but prospects for utilization of the LIS are not as promising as in Rajangana. There are essentially three types of settlement schemes included in Polonnurawa: new youth schemes, and new and old colonist schemes. The different characteristics of each will be discussed in turn. 2.14 Youth Schemes. The youth schemes are the only functioning LIS units in Polonnaruwa. Unlike other settlers, people on the youth schemes were given 2 acres of irrigated upland, but no paddy land. Because these farmers depended on the LIS for basic production, rather than supplemental income, they took a more active interest in the system. The LIS youth settlements were completely undertaken within the project. Settlers themselves cleared the land and established their farms, thus the settlement process was a bit protracted, and, the last 20% of the settlers could not be selected until after completion of Credit disbursement. 2.15 Until 1981, youth settlers were provided with food supplements and other incentives to remain on their farms. In 1981, however, most of these programs were terminated concurrently with the imposition of the requirement that farmers assume fuel costs for the LIS pumps. Because of their inability to pay for fuel, farmers stopped using the system for two years, starting again in 1983. During the interim most of the settlers survived by working as farm or construction laborers. Once it was clear that the government policy was not going to be changed, and once farmers either saved enough money for a season's fuel or obtained fuel credit from the Bank of Ceylon, the settlers returned to work on their farmsteads. With the help of extension and colonization officers, they improved chilli production and increased water utilization efficiently. 2.16 New Colonists. Giritale was a new settlement which was covered by the LIS. Located on marginal land, the scheme was essentially abandoned after better lands became available under the Mahaweli Program. Pumphouses and some channels remain as evidence of the LIS, but nothing else. Homesteads have disappeared since many people moved away around 1980. Poor lands and uncertain water supplies forced settlers to move to other locations. This system was simply not viable from the beginning. 2.17 Old Colonists. In contrast to the new settlers, people on the older settlements included in the LIS had productive paddy lands so that they showed little interest in upland cultivation. This situation was described in project supervision reports and is supported by the income data generated by the consultant's survey. The mission found farmers in one older settlement who did not use the LIS because they could not muster up enough of an organization to collect fuel money. They had made fuel contributions to the Cultivation Committee in 1981, but the funds never reached the Irrigation Department and no subsequent attempts were made to establish a responsible, responsive farmers' association. - 11 - 2.18 Nagadeepa (1200 ac). It was a major mistake to include Nagadeepa in the LIS program. This is a new settlement on gently rolling terrain which was chosen to substitute for another scheme that was found infeasible after the onset of the project. Settlers arrived in Nagadeepa in the belief that they would cultivate two paddy crops per year. However, due to insufficient irrigation water and high soil percolation rates, water supplies have been chronically short. Two paddy crops have been obtained only once in fifteen years (1983/1984), and yields are so low that many farmers are compelled to leave during the dry season to find supplemental employment. 2.19 The Nagadeepa LIS was completed near the end of the project period, and approximately 18% of the potential command area received water during one season. In subsequent years, farmers were asked if they wished to utilize some of the canal water for upland areas, but they declined for one or both of two reasons. First, many farmers had already planted fruit trees on the uplands, which they did not want to irrigate. Second, in the absence of prior experience and extension assistance in intensive chilli cultivation, farmers tended to use their uplands for mixed subsistence crops, as they had done under the chena system,2!/ and they did not want to divert water to these lower priority lands at the cost of reducing already uncertain supplies to their paddy lands. This failure to make use of the LIS was interpreted officially as a lack of interest on the part of farmers. However, given their circumstances, it is more correct to say that the system itself was faulty, and farmers could not afford the luxury of utilizing it. In early 1981, all pumps were removed from the scheme and the scheme was deemed a failure. 2.20 It is interesting to observe that a 1971 supervision report rejected the request of the Government Agent in the area to undertake another scheme instead of Nagadeepa because of its unsuitable terrain. Ironically, one of the last supervision reports directed attention to the excellence of the Nagadeepa LIS design. In retrospect, it is difficult to give much credence to the aesthetic qualities of an irrigation system design if the system has no water to deliver. Had the earlier suggestion been explored as seriously as the Government's other recommendations regarding water requirement modifications, the money spent on the Nagadeepa scheme would not have been wasted. Instead of an LIS scheme, farmers in Nagadeepa would have benefitted much more substantially from efforts to help them improve water management or an agricultural extension program directed to help them grow subsidiary food crops on their paddy lands during the dry season, thereby reducing total water requirements. However, neither occurred, thus it is clear that the only tangible benefits resulting from the Nagadeepa scheme were the employment opportunities during construction and thereafter in guarding pump houses. 2/ Slash and burn. - 12 - Farmer Responses 2.21 There are many reasons why farmers were less than enthusiastic about the LIS. For example, the PCR identifies "lack of motivation" as an important factor, meaning that farmers did not want additional income or the additional workload of intensive upland cultivation. Given objective conditions in the schemes, especially problems with water availability, however, such a contention seems questionable. 2.22 The PPAR claims that paddy and upland labor peaks coincide, which discourage farmers from upland cultivation. Interestingly, irrigation officials share this view, but farmers do not. Neither the consultant's survey nor the mission's field interviews obtained data to substantiate the claim that upland cultivation has been constrained by labor shortages, except in Nagadeepa. In the latter case, it is likely that labor shortages resulted from seasonal out-migration rather than paddy/upland labor peaks. Indeed, the experience at Rajangana indicates that effective chilli cultivation generates enough income to warrant the use of hired labor. 2.23 Three factors would seem to have an important bearing on farmer responses, however. First, most farmers in the LIS schemes are still basically subsistence farmers who have to follow risk aversion strategies in order to survive. Given the high capital requirements necessary to start chilli production, especially compared to most other crops, and given uncertainties in water supply and marketing, it is not surprising that farmers are uninterested in LIS cultivation, or accorded it low priority. The risks were too great for most farmers to assume, especially given their need to establish their farmsteads. Lacking governmental and community support networks, farmers had to guarantee their subsistence to the extent possible before taking risks. The schemes in which upland cropping incurs the lowest risks-Vavunikulam and Rajangana Left Bank-have the highest rates of LIS utilization. 2.24 Second, experience and skill can lower risk levels, as in Vavunikulam, where upland horticulture is part of local tradition. In contrast, intensive chilli cultivation was new to most farmers on the other schemes, who could not depend on extension or irrigation personnel to help them improve their cultivation or water management skills, and thereby reduce their exposure. The extension people posted on LIS sites seemed to know less than the farmers themselves, and no agency concerned itself with water management at all. 2.25 Third, farmers can reduce their risks if they can gain some element of control over important variables. In Rajangana Left Bank, the group credit schemes gave farmers some control over working capital, and government intervention enabled them to establish a dependable export marketing program. Both of these activities helped Rajangana farmers to reduce the risks inherent in chilli cultivation and thereby encouraged utilization of the LIS. - 13 - 2.26 Adding to the risks, theft is a problem in chilli production. Farmers in Vavunikulam and the Polonnaruwa youth schemes were fortunate to have their upland plots encircling their homes, from which they could guard their chillies without disrupting their family life. In contrast, farmers in Rajangana have to erect temporary shelters on their fields in order to guard their crops. 2.27 Until they started paying for fuel, LIS farmers had virtually no control over water distribution or water use. People at the heads wasted water and those at the tails never got enough. However, this situation has started to change in the most active LIS schemes, by individual farmers receiving water on requested days and with the establishment of water user groups. 2.28 On balance, it can be said that LIS farmers have behaved rationally and appropriately, given their circumstances, even though such behavior was not consistent with the SAR projections. Project Assumptions and Outcomes 2.29 What caused the major discrepancies between anticipated and actual benefits, between anticipated farmer responses and actual responses? Were the discrepancies due to faults in the conceptualization, design or implementation of the LIS project? Do they reflect technical, adminisrative or social problems? 2.30 In the mission's view, the basic idea of the project was sound, but detailed preparation was inadequate to provide an appropriate blueprint for successful implementation. Moreover, actual implementation focused almost exclusively on construction problems, ignoring very important institutional questions or glossing over them. The project was conceptualized as an integrated, irrigation-based agricultural program. As such, it indicated a sensitivity to issues like the proper sequencing of project activities, its institutional matrix, and the array of services required for successful implementation. However, the design went little beyond that point. It made many assumptions about national subsidiary food crop needs, institutional relationships and capabilities, and the priorities, abilities and universe of choice of farmers, among other things. 2.31 However, the SAR did not spell out relationships or the allocation of responsibilities in sufficient detail to provide operational guidelines; nor did it contain provisions for implementing agencies to verify assumptions and modify the project accordingly, during implementation. Thus the design included basic elements necessary to ensure the viability of the project, but the framework was too sketchy to mobilize those elements once the project got underway and many basic assumptions proved to be false. That is, what appeared to be a good project concept fell apart in implementation because it was in fact only partially implemented. This can be seen by reviewing some project assumptions and following their fate during implementation. -14 - 2.32 The SAR made two fundamental assumptions of a social nature. First, it assumed that farmers in the target areas could easily choose to shift from subsistence to commercial production. This assumption is predicated on the prior assumption that physical and institutional elements of the project are in place. Where the project elements were established (Vavanikulam, Rajangana Left Bank and Poloanaruwa youth schemes) the assumption proved to be reasonable. Where they were not established, the onus of failure was placed on farmers, rather than project administration or design. 2.33 Second, it was assumed that farmers would eventually be given responsibility for the operation and maintenance of the LIS facilities. This assumption has both bureaucratic and social aspects. It assumes that irrigation authorities are willing to relinquish control over facilities and are able to prepare farmers to assume control. It further assum;-s that farmers are willing and able to accept the responsibility. In practice, the bureaucratic element of the assumption proved to be false and the social element proved to be true, rather by default than design. That is, irrigation authorities wish to relinquish responsibility for the LIS schemes, but not badly enough to establish a systematic pattern of transmittal. On the other hand, farmers have shown that they can assume administrative and regulatory responsibilities, if given them. Nonetheless, they have yet to be given the opportunity to handle the mechanical and other technical aspects of pump management, so it is still premature to say whether it is feasible to break completely the dependency relationship which still ties farmers to officials. Farmer Incomes 2.34 One of the principal objectives of the project, and predictably one of the key factors in project justification, was the increase of farmer incomes resulting from the project. The anticipated net return in year 10 of the project was Rs 1335 to the one-acre colonist and Rs 1815 to the two-acre Youth Settlement farmer at 1963-1966 import prices of chillies and onions. At the somewhat higher Guaranteed Price Scheme (GPS) prices, these figures would rise to Rs 2335 and Es 4115 respectively. At project completion it was noted that although incomes of participating farmers had increased substantially, the project had "so far failed to create the expected impact, namely Improving incomes of the participating small farmers...." 2.35 Direct comparison between current (Yala 1983) performance and 1968 expectations has not been possible since income expectations had been based on the cultivation of high-value chillies and onions, and at present onion is being cultivated only at Vavanikulam. Several other crops are being cultivated to a smaller extent with LIS water in the other schemes, these being, however, relatively low-value crops. 2.36 Although it has not been possible to separate incomes received by farmers from irrigated supplementary food crops (SFC) and non-irrigated SFC, figures of mean SFC incomes suggest a close relationship between these incomes and the percentage of SFC cultivated under LIS. - 15 - Project Proportion of Mean SFC Income SFC under LIS () (Ra) Nagadeepa 0 450 Rajangana Right Bank 23 1040 Polonnaruwa 73 3410 Rajangana Left Bank 81 5260 Vavunikulam 100 6110 2.37 The link between SFC incomes and the percentage of LIS acreages is obviously not a direct one, since both the extent and mix of crops vary from scheme to scheme. However, the distinct income advantages enjoyed by farmers who make greater use of LIS is evident from both the above figures and from the income percentages below, which show SFC incomes as a proportion of total agricultural incomes. The table shows that approximately half of all incomes in the two schemes which make high use of LIS (Vavunikulam and Rajangana Left Bank) comes from SFC. At Rajangana Right Bank (23% LIS use) and at Nagadeepa (no LIS use) these percentages are 18 and 5 respectively. The relevant percentage at Polonnaruwa is distorted by very high rice incomes. Ratio of SFC Income to Mean Agricultural Income Project LIS use Mean agric. Mean SFC Proportion of (%) income (Rs) income (Rs) SFC income (Z) Nagadeepa 0 9,140 450 5 Rajangana RB 23 5,830 1,040 18 Polonnaruwa 73 24,220 3,410 14 Rajangana LB 81 10,900 5,260 48 Vavunikulam 100 12,740 6,110 48 2.38 Since project farmers depend on farming as their principal source of income - accounting for over 75% of total family income - it is to be expected that their physical quality of life will reflect to a considerable extent, their success in farming. In the absence of baseline evidence relating to the project areas or to strictly comparable areas, besides comparisons among themselves, some comparisons are made to Lower Uva where recent data on physical quality of life of rural subsistence farmers (who did not have the benefits of LIS) is available. It is to be noted, however, that the determinant is total income from agriculture, of which LIS outputs constitute only a minor component (less than 20%) except at Vavunikulam and Rajangana Left Bank, where the relevant percentage is around 50%. - 16 - 2.39 The following table shows the distribution of farmers enjoying average and above-average grades of living conditions in the different schemes and in the Lower Uva region - signifying the average f or rural Sri Lanka. Proportion of Farmers Showing Average and Above-average Living Conditions (Percentages) Living Vavuni- Polon- Rajangana Rajangana Naga- Lower Condition kulam naruwa RB LB deepa UVA Housing 40 77 28 24 34 - Household Furniture 18 79 35 45 32 26 Household Appliances - 70 14 29 14 8 Food 28 93 68 97 88 27 Clothing 60 88 71 81 92 43 Non-essential expenditure - 76 43 62 24 10 Source of drinking water 90 39 20 27 20 12 Latrine facilities 40 91 69 71 92 34 Disease prevalence rates (Z) 1.0 5.1 2.0 5.1 8.6 6.7 Death rate (per 1000) 1.0 3.4 3.0 1.3 3.2 4.3 Literacy 92 87 85 92 77 71 School attendance - 5 - 9 yrs M 71 75 42 68 60 76 F 82 90 71 76 71 75 10 - 14 yrs M 100 100 90 97 92 80 F 93 88 96 100 77 76 15 - 19 yrs H 100 92 95 97 89 41 F 93 100 95 100 70 41 - 17 - 2.40 In respect of housing conditions, quality of household furniture and grade of household appliances, Polonnaruwa farmers with their high agri- cultural incomes fare better than farmers in othel- projects. Rajangana Left Bank farmers, at about half this level of agricultural income, are also rela- tively well equipped with household appliances. On the other hand, Vavuni- kulam farmers who enjoy very high incomes and take the maximum advantage of LIS, give themselves very little by way of these amenities except in the case of housing; their standards in respect of furniture and household appliances being even worse than that of subsistence farmers in Lower Uva. These condi- tions are characteristic of the high degree of austerity practiced by farmers in the Northern part of the country. 2.41 In the Polonnaruwa and Rajangana Left Bank projects the grade of food consumed is far superior to that in the other projects. With the excep- tion of Vavunikulam, farmers in all schemes do far better than rural farmers in Lower Uva. Expenditure on non-essential items among Polonnaruwa farmers * and, to a lesser extent, among Rajangana Left Bank farmers reflect strongly their higher incomes; predictably the Vavunikulam farmer maintains non-essen- tial expenditure at a very low level - in fact even lower than the subsis- - tence farmers of Lower Uva. 2.42 The Vavunikulam farmer appears to invest his higher agricultural incomes in other areas than in conspicuous consumption. He alone among proj- ect farmers seems to be sensitive to the need for improved drinking water. In respect of above-average grade latrine facilities, he is surpassed only by the Polonnaruwa farmer . Not surprisingly the disease prevalence rate is lowest among Vavunikulam farmers (1%) compared to the corresponding figure of 5% in other areas. The death rate at Vavunikulam is 1.7 per 1000 compared to 3 per 1000 in the other projects and to the national figure of 6.8 per 1000. 2.43 Literacy rates are significantly higher than the national rates at Vavunikulam, Rajangana Left Bank and Polonnaruwa - the highest being at Vavunikulam. The literacy rate among Nagadeepa farmers is significantly lower than the national rate. School attendance is significantly higher among project families than the national average - the highest attendance being recorded at Polonnaruwa, Rajangana Left Bank and Vavunikulam, and the lowest at Nagadeepa. 2.44 Thus viewed from any aspect of living conditions and social stan- dards, farmers in. schemes where LIS is practiced to a higher degree are significantly better off. The only exception is the Vavunikulam farmer whose increased income resulting from LIS will not necessarily be reflected in his level of livi-&g because of his characteristic sense of austerity. Impact on Employment 2.45 The employment rates (as a percentage of the total population) in the different schemes constituting the project are compared inter se as well as with the relevant figures for Lower Uva and for the country. - 18 - Area 10-24 years 25-59 years Vavunikulam 8.0 43.0 Polonnaruwa 4.0 44.0 Rajangana Right Bank 2.0 48.0 Rajangana Left Bank 3.0 45.0 Nagadeepa 3.0 49.0 Lower Uva 16.8 54.50 Rural Sri Lanka 18.31 53.22 2.46 The degree of participation in the LIS does not seem to have any bearing on the employment situation in the different schemes. Scae employment opportunities, for females during peak weeding and harvesting periods, have arisen especially on farms with small families. Conclusions 2.47 The project did not consider adequately the constraints under which farmers worked. Most of the target farmers were still trying to establish themselves on their settlement sites, trying to meet their subsistence needs. Some farmers worked their former holdings, as well as new ones. Farmers in some sites had an easy time adjusting; others found harsh and uncertain conditions that were difficult to adjust to. LIS sites were selected without considering such isenes, particularly without determining whether or not settlers had reached the stage where they could afford the time and capital requirements of intensive upland cultivation. Moreover, sites were selected without adequate data about the physical limitations of the sites, especially water limitations. As a result, a number of sites were chosen which were totally inappropriate for LIS, or for which LIS was premature considering the development stage of the settlement. 2.48 The project did not involve farmers in project implementation, except as laborers, and made no provisions to transfer responsibility for the schemes to farmers. The farmers were expected to be enthusiastic, but passive participants in a major development process. Not surprisingly, enthusiasm appeared only when farmers started to assume some responsibility for the system, however marginal. 2.49 Although the absolute extent of chilli and onion production is far less than anticipated in the SAR, production is now well established in certain areas and may be expanding. Moreover, farmers in LIS areas have even purchased their own small pumps to supplement LIS deliveries, clearly indicating the viability of lift irrigation under reasonable conditions. Indeed, the greatest danger facing the schemes at present is not that farmers will abandon upland cultivation, but rather than the pumps may fail and the system collapse, leaving farmers in the lurch. 2.50 Despite these problems, some aspects of the LIS are working. Where a number of elements converged, where water was available, where farmers were experienced or trained, where markets existed or were developed, and where labor and capital, or credit, were available, chilli and onion production has been established. The cost of the program was high compared to the magnitude of benefits obtained so far, or even potential benefits, but the benefits are tangible and promise to continue to be made manifest. III. TECHNOLOGICAL IMPACT 3.01 Project farmers were basically rice farmers with limited experience in some extensive wet season crops. The traditional non-irrigated cropping with its dependency on adequate rainfall restricted input use to a minimum and entailed low level cultural practices. Due to the introduction of small centrifugal pumps to lift water from existing channels, as well as some diversion of gravity irrigation from paddy tracts, some knowledge of modern input use and improved cultural practices were established prior to the start of the project. On the other hand it needs to be pointed out that traditional chillie cultivation is on a very extensive basis without any use of modern inputs. The project thus succeeded in introducing modern technology to project farmers. Modern Input Use 3.02 Based on available research work several varieties of chillies have been recommended for cultivation to farmers. All farmers interviewed by the mission have been growing either the MI One or MI Two variety. Both varieties produce good yields of quality chillies which can be dried or which find good export markets in the form of green chillies. A negative characteristic of these varieties is their increasing susceptibility to virus infection and lack of resistance to collar rot (phytophthora palmivora). 3.03 The response of chilli and onion yields to fertilizer application is well accepted and, according to the sample farm survey, almost all farmers are applying the recommended dosage of 300 kg/ac, mainly in the form of compound chilli/onion fertilizers. It is worth noting that fertilizer appli- cation does not show any difference in project areas where farmers claim good extension support as opposed to the schemes where farmers consider extension to be insufficient. - 20 - 3.04 Equally widespread is the use of pesticides in chillies despite relatively high costs of about Rs 2,000/ac and season. A total of 6 sprayings are recommended per season but variations are possible depending on the occurrence of aphids, the main vector in transmitting virus diseases in chillies. No information could be obtained on plant control measures practiced in the only onion-growing scheme at Vavunikulam. 3.05 Land preparation for chillies and onions requires plowing which is best done by tractors. While tractors have been popular in the northern parts of the country, they are sparsely used in the other regions except for some of the project chilli growers. The survey undertaken shows the following distribution of modern technology adoption: Proportion of Farmers Adopting Modern Technology Z using % using Z using Z using fertilizer insecticides weedicides tractor 6 fungicides Chillies 100 100 9 9 Polonnaruwa Cowpea - 33 - 22 Greengram 30 100 - - Vavunikulam Onions 100 98 98 96 Chillies 100 100 80 100 Cowpea 7 83 21 3 Rajangana Chillies 71 86 19 13 Right Bank Maize - 50 - - Blackgram - 100 - - Chillies 98 100 17 22 Rajangana Cowpea 5 34 - - Left Bank Bengal gram - 100 - 25 Greengram - 33 - - Maize - - - 13 Nagadeepa Chillies 22 33 - 11 Cowpea - 17 17 - Greengram 50 50 - 50 Generally it can be concluded that the technology adopted by the farmers participating in the project is on a higher level compared to their counterparts engaged in rainfed cultivation. - 21 - Husbandry and Water Management 3.06 The recognition by farmers of the high profitability of chilli/onion production has led to the introduction of a monoculture with all its negative side effects. Since no other crops can assure equal or at least similar income levels, farmers tend to cultivate the high value crops year after year on the same plot. The small size of the project farms does not permit the introduction of a crop rotation without, from the farmers' point of view, serious consequences, i.e., loss of income. The absence of any crop rotation has produced an increasing incidence of collar rot against which no economic control has been found and Sri Lanka research has so far been unable to breed resistant varieties. The farm sample survey confirms the absence of rotational practices. Cropping Pattern (acres) 1976 1977 1978 1979 1980 1981 1982 1983 Vavunikulam Chillies 353 436 436 408 400 200 100 28 Onions 70 80 80 - - 100 150 210 Other - - - - - - - - Rajangana Chillies 220 294 294 295 939 413 1025 1367 Onions 10 12 12 10 2 - - - Other - - - 288 575 184 178 94 Polonnaruwa Chillies - 265 265 230 230 - - 220 Onions - 5 5 5 5 - - - Other - - - 30 30 - - 26 The appraisal had assumed a restricted form of rotation: chillies to be followed by onions on about one-third of the acreage. This assumption was based on a rotation calendar of chilli growing from January to July and onions from July to September. However, farmers prefer to plant in March, thus harvesting in August/September, having no chance of following the chilli crop with onions. 3.07 The mission was unable to ascertain if the adopted planting schedule is a consequence of farmers' more recently recognizing the need for proper water management. By planting in March/April farmers can still use water and soil moisture from the October to April rainy season without any - 22 - irrigation. Improved water management has resulted from the introduction of charging farmers the cost of fuel and lubricants in 1982. Officials of the Irrigation Department stated that prior to farmers paying the fuel costs, they had received continuous complaints about insufficient irrigation water supplies. Now farmers take maximum advantage of any rainfall available during the cultivation period to reduce irrigation. 3.08 However, there is further scope for improved water management. The appraisal had assumed that "land levelling and grading would be carried out over the entire project area" (AR, para. 4.03). The mission found no signs that any land levelling had ever taken place. Indications are that substantial quantities of irrigation water must be lost due to undulating fields. Conclusions 3.09 The technology of chilli and onion cultivation among farmers has improved considerably over the last ten years, and the research data for supporting chilli and onion cultivation has also expanded a great deal; although some disease problems remain unresolved. Farmers make good use of modern inputs but dangers of monoculture have become apparent. 3.10 Two basic technical assumptions proved to be problematical. First, the water requirement quantities used in the design of the LIS underestimated crop requirements and overestimated the efficiency with which farmers would use the water. This led to the selection of inappropriate sites, as well as uncertainties in water delivery. Design was modified during implementation, but many farmers had already become disenchanted with the LIS by then, and others never received enough water to learn how to utilize it efficiently. 3.11 The SAR assumed that the Irrigation Department would assist farmers in establishing efficient water management techniques. In practice, the department has been concerned only with the construction, operation and maintenance of physical structures. It has no experience in teaching and has yet to manifest concretely the commitment to improve water management which was identified in the PPAR. IV. IMPACT ON LOCAL INSTITUTIONS Water Users Associations 4.01 According to the appraisal report farmers were organized into cooperatives with the exception of the Youth Scheme which was, however, expected to be organized along similar lines. The SAR stated that: "It is Government's intention to eventually turn over responsibility for the opera- tion and maintenance to these cooperatives."3/ 3/ SAR, p. 9, para. 6.02. - 23 - 4.02 Due to the Government's reluctance to introduce water charges, there was never a transfer of Operation and Maintenance (O&M) responsibili- ties to farmer/water user associations. 4.03 There is evidence that at the commencement of the project an effort had been made to organize farmer groups in some of the schemes. The LIS Survey revealed that farmer groups had been extremely active in Vavunikulam from the commencement of this project and continue to provide support to individual farmers to this day. Similar efforts had been made at Polonnaruwa though with only moderate success. 4.04 Only after 1981 when farmers assumed responsibility for supplying fuel, they have started to form nascent water user associations. The first impetus for this process came from the Irrigation Department's directive that LIS pumps not be operated unless farmers themselves establish a procedure for supplying fuel as of 1982. Some groups began to organize in 1982; others took one or more years to do so. Some groups started pooling funds to buy - fuel collectively, but too many of the collection committees abscounded with the funds, so individual purchase and delivery soon became the norm. In sub- sequent stages, most of the groups have become water management groups which . serve to ensure that water is delivered on time and that delivery quantities correspond to the amount of fuel supplied by individual farmers. In other words, once they acquired clear vested interests in the delivery system, and were given responsibility for it, farmers began to be concerned with equity as well as water use efficiency. 4.05 At the Rajangana Left Bank scheme a small degree of group activity had existed throughout and this activity has been successfully revived in recent times. This group has now been incorporated into a limited liability company for the specific purpose of exporting green chillies to Middle East markets. The initiatives to this end had come from the Ministry of Trade and Shipping and its export firm, the Export Development Board. Under the aegis of this company, elected farmer committees recover the costs of fuel expenses for pump operation and also allocate production quotas to individual farmers to meet export orders for green chillies. At present the company has more than 350 shareholders and its savings deposits are of the order of Rs 800,000. 4.06 The success of this corporate enterprise has apparently encouraged farmers in the Rajangana Right Bank to form a society for the purpose of exporting fresh vegetables to Middle East markets. A limited liability con- pany on the lines of the Left Bank company is being contemplated. 4.07 In Polonnaruwa the Bakamuna Youth Settlement Scheme which restarted cultivation in 1983 after a two year shutdown is being managed wholly by project participants. The fledgling organization set up by these youths looks after the supply of fuel, the operation, and the maintenance and repair of the pumps in the scheme. The existence of a great deal of social homoge- neity among these youths, who are all children of colonists, augurs well for the development of this farmer organization. - 24 - 4.08 The slow growth of farmer organizations in these schemes, though limited in scope, is nevertheless encouraging because Sri Lanka has not had a tradition of self-reliant farmer organizations to support agricultural activities in the rural sector. The assumption in the project formulation that such organizations already existed or could be developed within a very short time was, however, erroneous. Government Institutions 4.09 The institutional support expected for the project from GOSL had been clearly identified at project appraisal and the more serious shortcom- ings disclosed at audit had been equally clearly indicated. 4.10 The PCR pointed out the lack of effective coordination between the Irrigation Department (ID) which was responsible for the construction and maintenance of pumps on the one side; and the Department of Agriculture on the other. There had been no serious attempt to achieve effective coordina- tion even after project completion. As far as the ID was concerned, civil works and the strictly technical aspects of irrigation certainly receive more attention than other aspects. Management of pumps, the question of recover- ing fuel costs from farmers, the question of payment of electricity charges, payment to pump operators, attempts to control cannibalization of defective pumps to maintain others - these, naturally, had been the main pre-occupation of the ID staff in the districts: although, admittedly, the senior staff at the Head Office in Colombo were certainly mindful of the economic and agri- cultural aspects as well. However, since the main concern of the implement- ing department was irrigation, other aspects tended to take a second place. 4.11 The Task Forces which had been established in each District compris- ing the General Administrator, the Chief Irrigation Engineer and the District Agricultural Extension Officer might have served to promote the schemes to some extent, had they been effectively followed up. Even after the need for closer coordination at district level had been pointed out at audit, no meaningful steps have been taken to improve coordination. 4.12 Thp farm survey confirmed that agricultural extension services were less than adequate and that GOSL had not met its commitments to the project in this regard. The appraisal had assumed that farmers would need to learn intensive chilli and onion production techniques and that the agricultural extension service was able to provide appropriate instruction and committed to do so. In practice, it proved to be difficult to mobilize extension people at all, and those who were mobilized had very little to offer the farmers. The department lacked both the commitment and the substantial message which had been assumed. 4.13 The same survey reveals that except in Vavunikulam and to some extent in the Ra_angana Left Bank, many farmers claim that Government officials were inactive. At Vavunikulam, all farmers had confirmed that Government officials were active in their area, and in Rajangana Left Bank - 25 - 50% questioned reported they were active; in the other locations, only 20% felt they were active. At Nagadeepa all those questioned reported that Government officials were not in appearance. 4.14 The audit had pointed out that although subject matter specialists were available only at Vavuniya and Rajangana, useful technical know-how was available at Maha Illupalama research station. It was hoped that action might be taken (particularly in the light of the assurance given in 1979 to an IDA Mission that extension services at Nagadeepa and Polonnaruwa would be improved) to improve* extension services in the weak areas. However, no evidence was found of steps having been taken towards tangible improvements. 4.15 A farmer deciding on producing chillies would require working capital of up to Rs 1000/ac, mainly to finance land preparation, hired labor and for the purchase of inputs. During appraisal it was assumed that credit facilities would be made available to farmers. In practice, this was not the case because many farmers had already defaulted on loans. The SAR had not adequately explored this question, and the project itself developed no means to overcome this significant gap except on an ad hoc basis. To date many farmers, especially in Rajangana Right Bank, lack the required capital to take up chilli production. 4.16 The general consideration given to the LIS is best reflected in an aide memoire dated June 30, 1982 summarizing the main findings and agreements reached between an IDA mission4/ and concerned Sri Lanka Government offi- cials, where it says ..."It is noteworthy that both IDA and GOSL are not placing a priority on lift irrigation projects...." The aide memoire, of course, "en passant" remarked that the views expressed therein were subject to final approval by IDA Headquarters in Washington, D.C. However, this comment, at least as far as the attitude of the Government officials in the relevant Departments was concerned, aptly describes the "ennui" that characterized official attitudes vis-a-vis the project, after it was declared closed. 4.17 Questions can be raised about this attitude since without the replacement of the pumps/engines, sunk costs in channels, pipes, pumphouses, with a life expectancy of at least another 15 to 20 years, would be written off. The Bank and GOSL have entered discussions on an irrigation rehabilita- tion project. In the mission's view the prospects of rehabilitating the LIS in areas where farmers are making good use of existing facilities should not be neglected. Financing pump/engine replacements through a credit scheme may merit consideration. 4/ The Region in its comments draws special attention to the fact that it continued up to 1982 in assisting Government in making use of project facilities and in introducing necessary adjustments in order to achieve project objectives. - 26 - Conclusions 4.18 The project did not provide the institutional support which the SAR had identified as necessary for the success of the project. Extension, credit, water management and marketing support were needed, but rarely forth- coming because the implementing agency could not obtain the cooperation of relevant agencies. The SAR assumed institutional commitments which did not materialize, and probably would not have been assumed if the question had been explored adequately during project preparation. Moreover, supervision reports indicate a powerless awareness to generate such commitments during implementation. 4.19 The project, even after its closure, might have been rendered a limited success if in the post-audit period, most of these shortcomings had been rectified. The lacuna in every respect appears to have been the non-existence, in the apex statutory body viz. the Ministry of Lands and Land Development, of a mechanism which could have governed, managed and directed the IDA-financed LIS on a constantly monitored basis. 4.20 In a report to the Ministry of Lands and Land Development on the Nagadeepa scheme dated December 1980, the Agrarian Research and Training Institute remarked: ..."There appears to have been a serious dearth of coor- dinated supervision of lift irrigation...little coordina- tion among the three most important agencies...this is probably a major reason for the failure of this proj- ect...... It was the diffusion of responsibility and the concurrent lack of a specific and defined focus of responsibility that rendered Government institutional support ineffective. In the post-project period, the impact of the project could have been greater, had these factors been corrected. V. PRODUCTION IMPACT Cropping Intensity 5.01 At appraisal the cropping intensity for the project areas was expected to be 133% - an annual cropping of approximately 8,450 acres. The virtual abandonment of irrigated cultivation in the Nagadeepa scheme after only one season saw the total irrigable area decline to 5,150 acres. During the 1977 season the cropping intensity - excluding Nagadeepa - reached only 28%, representing a cropped area of 1,440 acres. This represents only 17% of the original anticipated cropping. Official figures indicate the following areas under irrigation: Cropped Area from 1978 to 1983 (acres) 1978 1979 1980 1981 1982 1963 Chilllee nions Other Chillies O61on Other Cllies Omoson Other Chillies Wioas Other i=lligs oon Other Chili1se tons Other WaTualkwlas 436 80 - 406 - - 400 - - 200 100 - 100 150 - 28 210 - Majampais 294 12 - 295 10 288 939 2 575 413 - 184 1025 - 178 * 1367 94 foleasarwa 265 5 - 230 5 30 230 5 30 - - * - - - 220 - 26 WaWeep& - - - - - - - . . . . . - . - - - Total 995 97 - 933 15 318 1569 7 605 613 100 184 1125 150 178 1615 210 120 oran total 1092 1266 2181 897 1453 1945 - 28 - While the PCR had assumed that cropping intensity would reach 40% by 1983, the actual percentage stands at 22%. The highest ever reached was in 1980 when 34% were irrigated. The cropping intensities in the individual schemes for 1983 were: Vavunikulam 49%; Rajangana Right Bank 16%; Rajangana Left Bank 62%; Polonnaruwa 26%. Performance of the Vavunikulam scheme in 1983 was well below capacity since its cropping intensities had exceeded 100% in previous seasons. The poor performance last year was due to the low water level in the Vavunikulam reservoir - a consequence of inadequate rainfall during the early part of 1983. 5.02 It should be noted that chillies are grown in the same areas without the LIS. The chilli areas outside the project - using own pumps or no irrigation water at all - are quite significant as can be seen from the following table. Chilli Acreage in Project Areas Scheme Under LIS Outside LIS LIS as % of total (ac) (ac) chilli acreage Vavuniku.Lam 238 0 100 Polonnaruwa 242 86 74 Rajangana LB 1164 272 81 Rajangana RB 297 990 23 Nagadeepa 0 1016 0 1941 2364 82 Productivity - Yields 5.03 The cultivation of the same crops under LIS and outside LIS in several schemes provides an interesting opportunity for the assessment of incremental productivity under LIS. The following figures, derived from the LIS Survey, demonstrate the high level of productivity for all crops in all schemes where LIS is practiced. The only exception is the greengram crop under LIS at Polonnaruwa, which seems to have failed for reasons other than irrigation - the high productivit, of greengram under LIS being demonstrated at Rajangana Left Bank. It should be noted that 1983 yields were below average due to low rainfall which in turn curtailed irrigation water availability during the growing season. Chilli yields in other years and also in 1984 average 1,000 kg/ac. This yield compares favorably with the appraisal assumption of 500 kg/ac at full development between the fifth and tenth year. Onion yields are in line with appraisal estimates. - 29 - 1983 Yields (kg/acre) Crop Polon- Nagadeepa Rajangana Rajangana Vavuni- naruwa RB LB kulam LIS 903 - 339 749 633 Chillies Non-LIS - 68 94 567 - LIS - - - - 4518 Onions Non-LIS - - - - - LIS 384 - 410 493 - Cowpea Non-LIS 25 166 153 215 - LIS 200 - - 600 - Greengram Non-LIS 270 430 - 240 - Source: Consultant Survey Prices and Costs 5.04 The appraisal asumed that the average price under the Guaranteed Price Scheme (GPS) would be Rs 225/cwt (Rs 4.5/kg) for chillies and Rs 27/cwt (Rs 54/kg) for onions (1968 constant value). Using the official price deflators, these prices at current rates would amount to Rs 18.81/kg for chillies and Rs 2.25/kg for onions. Actual prices received are Rs 22/kg chillies and Rs 5.13/kg onions. However, free market prices during the appraisal year - at 1983 values, had been Rs 26.1/kg chillies and Rs 2.34/kg onions. While there is now a better price obtainable for onions, chilli producers arv faced with chilli prices that are lower in real terms compared with 1968. 5.03 On the other hand operating and maintenance costs have increased dramatically. The appraisal had assumed total O&M costs to reach Rs 564/ac (1983 current value). However, fuel costs alone, paid by the farmers, average Rs 2500/ac and total O& costs have not reached Rs 3480/ac (1983). Of course nobody could have anticipated the 1973 and 1979 oil price increases back in 1968, but even without fuel costs, appraisal assumptions were too optimistic. Despite the worsening terms of trade chilli production under irrigation remains an attractive and profitable undertaking. WaL r Charges 5.06 With the exception of the Vavunikulam scheme where a Rs 300/ac fee had been levied since 1975, no water charges have been collected since - 30 - project inception. Although final assurances had been obtained by the Bank that water charges would be introduced in 19795/ no action was ever taken by the Government. Only in 1981 it was decided that farmers had to supply fuel and lubricants for the working of the pumps, which leaves Government to carry the full cost of maintenance and operating staff. 5.07 The absence of any charges for pump/engine replacements will make it impossible to replace the existing equipment now becoming increasingly obsolete and prone to breakdowns. The Irrigation Department has no funds available to procure required replacements. From its point of view already too much money had been invested to assist a very small number of farmers. Project beneficiaries, when questioned by the mission about the need to provide funds for new pumps/engines, seem to believe that Government will nevertheless step in again and continue to provide irrigation facilities. A few farmers have acquired small motor driven pumps but irrigation with these is only feasible on fields in close proximity to the main irrigation channels. Economic Justification 5.08 In 1978/79, when the PCR was prepared, some indications were found that gave credence to the assumption that full development of the project's irrigable area could be reached. The Government was in agreement with Bank suggested steps to improve extension activities, to finally introduce water charges which would permit the establishment of a sinking fund for pump/ engine replacements. Even in areas like Nagadeepa, where unreliable water supplies hampered chilli production, some farmers had considered production of high value fruit. All these factors contributed to the audit not comp- letely ruling out the the possibility of a positive ERR up to 16%. 5.09 Unfortunately the PCR assumptions did not materialize. The Nagadeepa scheme was dismantled in 1980; water charges have not been intro- duced. Due to the absence of any funds for pump/engine replacements the LIS is expected by the Irrigation Department to come to an end in 1985-86 when the mechanical parts will reach the end of their useful life. Because of delays in construction and implementation and due to a much lower cropping intensity, the recalculated ERR (see Annex 1) stands now at 1.3% - positive only because chilli yields are about 100% higher than envisaged at appraisal. Conclusions 5.10 The project contributed up to 40,000 cwt dried chilli equivalent to the economy. At appraisal it was estimated that about 65,000 cwt of dried chillies would be produced, leading to annual savings of US$2.3 million for imports. Despite its poor performance - acreage wise - higher yields contri- buted to 60% of the production targets for chillies being reached. 51 See PPAR, OED Report No. 2801, dated December 28, 1979: Aide Memoire of June 18, 1979 reproduced at Appendix 1. - 31 - 5.11 The unresolved issue of water charges illustrates the difficulties encountered by Governments in reaching correct decisions. There was a need for providing incentives to farmers. They needed encouragement to invest substantial funds in chilli cultivation - not properly identified at appraisal - and there was justifiable fear that introducing water charges immediately would frighten them away. In the mission's view, a credit package including the water charges for the initial years would have merited consideration. 5.12 The project has to be classed as "non-sustainable." There are no funds to provide for pump/engine replacements in the next two years when their economic life comes to an end. The lined channels, pump houses, pipes, etc., which would have another 15 years of useful life, will become redundant. Based on this shortened life the project's Economic Rate of Return has been recalculated at 1.3%. � � И � � �4 � � � � С�, � N г 33 - Annex 1 Economic Rate of Return Calculation Cost Benefit Stream Cost Stream Benefit Stream Net Stream 2,000 - - (2,000) 2,157 - - (2,157) 3,094 - (3,094) 5,467 - (5,467) 8,023 - (8,023) 12,355 - (12,355) 9,672 - (9,672) 2,373 1,260 (1,113) 532 3,125 2,593 604 4,806 4,202 703 4,003 3,300 845 7,499 6,654 807 6,475 5,668 807 5,038 4,231 807 8,127 7,320 807 8,300 7.493 807 8,500 7,693 ERR 1.3% SI LANKA * LIFT IRRIGATION PROJECTSCHEMES LOCATION MAP IBRD 2151 R (PPA Roads Demarcation of Wet ••••••and Dry Zones JAFN - . MANNA .l DRY ZONE. VAVUNIKULM4 -TRINCOMALEE ANURADHAPURA . POLONNARUWA PUTÝALA PO1.ONNARUWA- RAJANGANA BATTICALOA CHILAW0 NAGADEEPA KANDY • GO COLOMBO E T z E KALUTARA GALLE HAMåANTOTA MATARA 7M ss be is IBRD 2151 R. Fowar~ 1. 2 Dmen~ad~cefg h~a b~ msa0de o 24 43 ffit oms leMsiQaN.M L ES MARCH 198