Sri Lanka - Tree Crop Diversification (Tea) Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 5041 PROJECT PERFORMANCE AUDIT REPORT SRI IANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CREDIT 819-CE) April 13, 1984 Operations Evaluation Department 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. ABBREVIATIONS CF - Cardamom Farms DA - Department of Agriculture DCA - Development Credit Agreement EPF - Estate Workers Provident Fund FAO - Food and Agriculture Organization FD - Forestry Department GDP - Gross Domestic Product GM - General Manager GOSL - Government of Sri Lanka GRC - Gurugoda-Ritigaha River Catchment Area HA1T - Higher Altitude Mixed Tree Farms IDA - International Development Association JEDB - Janatha Estates Development Board LANT - Lower Altitude Mixed Tree Farms LRC - Land Reform Commission MAL - Ministry of Agriculture and Lands MADR - Ministry of Agricultural Development and Research MKC - Mahaoya-Kudaoya River Catchment Area L4P - Member of Parliament MPCS - Multipurpose Cooperative Societies MPEA - Ministry of Planning and Economic Affairs HRID - Ministry of Rural Industrial Development NAC - Nilambe-Atabage River Catchment Area NADSA - National Agricultural Diversification and Settlement Authority PA - Project Agreement PCR - Project Completion Report PTU - Project Technical Unit, National Livestock Board SCD - Soil Conservation Division SLFP - Sri Lankan Freedom Party SPC - State Plantations Corporation UNDP - United Nations Development Program UNP - United National Party FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CREDIT 819-CE) TABLE OF CONTENTS Page No. Preface *********** *********** *********** *********** *********** 0 - . . o Basic Data Sheet o ...........................* *...........*** Highlights ........ ................................... 111 PROJECT PERFORMANCE AUDIT MEMORANDUM I. PROJECT SUMMARY ........ ...... 1 II. AUDIT FINDINGS ........ ............ 3 A. General ..oo........ ..............o .......... 3 B. Project Preparation and Design ..................... 5 C. Borrower Support and Borrower Agencies' Performance.. 7 D. IDA Performance After Appraisal -........ooo ...... 9 Table 1 - Sri Lanka Tree Crop Diversification (Tea) Project - Disbursements ........ -oo... o.o......... o...o.. 12 Annex I - Communications Received from the Ministry of Plantation Industries 13 PROJECT COMPLETION REPORT I. Introducti6n ............ 17 III. Project Implementation and Performance ........... 26 IV. Rates of Return .... .. ... . ...o......... 41 V. Institutional Performance and Socio-Political Developments o..................... ......... oo...o. 41 VI. IDA Performance ..... ..................... . ..... ........ 47 VII. Conclusions .... o........ oo...... oo..................... 52 ap- IBRD 13383 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. PROJECT PERFORMANCE AUDIT REPORT SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT . (CREDIT 819-CE) PREFACE This is a performance audit of the Tree Crop Diversification (Tea) Project in Sri Lanka, for which Credit 819-CE in the amount of US$4.5 million was approved in June 1978. The last disbursement was made on February 16, 1982, and the Credit was closed in August 1982 instead of June 1983 as planned. The undisbursed balance of US$0.28 million was cancelled. The Government of Sri Lanka agreed to prepay the disbursed portion of the Credit in three installments between October 1982 and October 1984. The audit report consists of an audit memorandum prepared by the Operations Evaluation Department (OED) and a project completion report (PCR) dated July 18, 1983. The PCR was prepared by the Bank's South Asia Regional Office. The audit memorandum is based on interviews with Bank staff associated with the project, and on a review of the PCR, the Appraisal Report (No. 1936a-CE) dated April 3, 1978, the President's Report (No. P-2330-CE) dated May 19, 1978, the Credit Agreement of July 12, 1978, correspondence with the Borrower, and internal Bank memoranda on project issues as contained in relevant Bank files. A copy of the draft report was sent to the Borrower for comments on December 30, 1983. Communications received from the Borrower's Ministry of Plantation Industries are included as Annex I. The audit finds the PCR comprehensive and accurate with respect to the project's principal achievements and shortcomings and has no reason to question its conclusions. The audit memorandum deals with project design aspects, Borrower support during implementation, and IDA supervision performance because of these issues' relevance for explaining the final outcome of this project and their wider applicability for similar other projects. PROJECT PERFOKMANCE AUDIT REPORT SRI LANKA TREK CROP DIVERSIFICATIOU (TEA) PROJECT (CRET 815-CE) BASIC DATA SHEET KEY PROJECT DATA Appraisal Actual or Actual as Z of Ices Estimte Estimated Actual Appraisal Estimate Total Project Costs CUSS million) 6.5 11.2 172 Credit Amount CUSS million) 4.5 4.5 100 Disbursed 4.5 4.2 93 Cancelled - 0.3 Repaid (as of 10/31183) - 3.0 Outstanding (as of 10/31183) - 1.2/a Date Board Approval - 06f011s Credit Agreement Date - 07112f78 Date Effectiveness 10112/75 12/15/78 150 lb Date Physical Components Completed 07131/84 C Proportton Then Completed (M) 100 35 Closing Date 06f30/83 0/20182 82 /b Economic Rate of Return (2) 16 Not calculated but assumed negative Institutional Performance - MLasd Agronomic Performance - No record as yet Nmber of Direct Beneficiaries 4.500 890/d 20 CUULATIE DISBURSEMENTS Fy79 F80 FT81 FY82 FT83 Appraisal estimate %Mss million) 0.3 1.6 3.2 4.1 4.5 Actual CUSS million) 0.3 1.7 3.3 4.2 Actual as 2 of estimate 100 106 103 102 Date of final disbursement February 16. 1982 MISSION DATA Date No. of Manueeks Specializations Performance Types of mission (mlr) Persons in Field Represented /a Rating /f Trend L Problems /b Identification 09-10/76 2 6.8 Preparation 03/75-02/76 7 31.0 A.B Appraisal /1 10177-11/77 6 30.0 A,.C T,K.A.F Subtotal 15 67.8 Supervision I 06/78 1 1.4 C 2 2 M Supervision 2 09/78 1 1.4 B 2 1 P Supervision 3 03/79 2 2.8 B 2 1 P,0 Supervision 4 09/79 2 2.8 B 2 3 F,P Supervision 5 03/80 2 2.8 a 2 1 F,P Supervision 6 07/80 1 0.8 2 2 F,P Supervision 7Lj (07/81) - 0 - 2 2 F Supervision 8 10/81 2 1.6 A.3 3 2 F.P Supervision 9 02/82 1 0.2 A 3 3 P.F,O Supervision 10 05/82 2 2.8 A,B 3 3 F,0.P Subtotal 14 16.6 Total 29 84.4 OTHER PROJECT DATA Borrower Government of Sri Lanka Executing Agency National Agricultural Diversification and Settlement Authority (NADS) Fiscal Tear January 1 - December 31 Name of Currency (abbreviation) Sri Lanka Rupees (SLRs) Currency Exchange Rate: Appraisal Year Average USS.00 - SLRa 16 Intervening Years Average USS1.00 - SLRs 17.5 Completion Year Average US$1.00 - SLRs 20.5 Follow-on Project: NOne Is The Credit is to be prepaid in three installments, the first one having been paid by 10/01/82 (USS1.5 million), the second by 10f01/83 (US$1.5 million) and the remaining falling due on 10/01/84 (USS1.2 million). / Calculated in term of months from the date of Board approval. 7F The date of completion is uncertain. As of April 1983, the weighted average of physical com- pletion for the various project components m- about 352 of appraisal targets, although these targets and the physical specifications of some components (e.g. housing) have changed since appraisal. Id Goverament's revised target for settlemant on project lands is 5,430 families, planned to be achieved by the end of 1983. le A - Agriculturist; B - Agricultural Economist; C - Financial Analyst. I - Problem-free or minor problems: 2 - Moderate problems; and 3 - Major problema. 1 - Improving; 2 - Stationary; and 3 - Deteriorating. F - Financial; N - Hanagerial; T - Technical; P - Political; K - Marketing; A - Agricultural: and 0 - Other. /I Second appraisal; appraisal of 05/76 rejected the project. ZI Annual update - no field visit. - iii - PROJECT PERFORMANCE AUDIT REPORT SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CREDIT 819-CE) HIGHLIGHTS The Tree Crop Diversification (Tea) Project was the first IDA-assisted project in the tree crop sector in Sri Lanka. The project was prepared simultaneously with a Tea Rehabilitation Project. In 1967, facing the need to improve overall tea production and yields, the Government of Sri Lanka (GOSL) appointed a commission to review the effects of a replanting progra& to replace old seedling tea with high--yielding clones. This study identified the need to diversify some tea land to other crops, and in 1969, GOSL requested UNDP/FA0 assistance in determining a stra-egy for optimal use of ted land. A five-year UNDP project followed to study this aspect and, subsequently, GOSL requested FAO/CP help in preparing a project for IDA financing. The project had to be appraised twice. The main objective of the project was to diversify -unsuitable mid-country estate tea lands into mixed tree crops and to settle farmers on one-ha farms, each with a 0.2 ha homestead consisting of vegetable garden and a simple temporary house. The project started well after some delay in credit effectiveness. Procurement picked up quickly, physical works progressed impressively, and disbursements were well up to appraisal targets. However, several institu- tional aspects involving the settler selection and land allocation process did not proceed as planned. These were considered to be only minor impediments at first but continued unresolved and ultimately led to the project failing. The settler selection and land allocation problems resulted in unanticipated financial burdens on the implementing agency (NADSA) and intense social and political opposition to the project. More specifically, the main problems were threefold. First, NADSA was forced to be responsible for factory operations, as well as estate staff and labor payrolls on project lands handed over by the government-owned estates to settlement. This was contrary to agreements between GOSL and IDA. Second, delays in transferring labor from the project area to other government-owned estates as agreed resulted in strong opposition from labor unions to the project. Finally, cost overruns and the additional expendi- tures imposed on NADSA resulted in a shortage of funds. These problems continued to plague the project throughout implementation, and IDA should have applied pressure to rectify the situation. In spite of the forces working against it, NADSA achieved remarkable physical progress. But the delays in settler selection and land allocation meant that established farmsteads remained unprotected and subject to vandalism from those opposed to the project. As a consequence, much of the farmstead development investment was lost. -iv - Once the political opposition to the project reached its peak, GOSL ordered the retransfer of parts of project lands from NADSA control back to the control of the Land Reform Commission (LRC), so that the land could be returned to tea cultivation. This action violated basic covenants contained in the Credit Agreement. On the basis of these events, IDA suggested to GOSL a need to disassociate itself from the project. As a consequence, GOSL proposed to prepay the disbursed amount of the credit. Thus, the credit was closed in August 1982 after cancellation of the undisbursed portion of about US$0.28 million, and repayment is to take place in three installments during the period 1982-84. Other points of interest are: - there was a long pre-project history with involvement of external agencies (PPAM, paras. 13-15; PCR, para. 2.03); - appraisal and project design suffered from inadequate coverage of socio-political aspects (PPAM, paras. 16-17; PCR, para. 7.03); - some procedures and responsibilities were inadequately specified in advance (PPAM, paras. 18-19; PCR, para. 6.09); - the Borrower made design changes and decisions during implementa- tion, some without consulting IDA, which adversely affected project progress and IDA support (PPAM, paras. 20-21; PCR, paras. 3.04, 3.20 and 3.12-3.13); - the implementing agency, while performing well with physical imple- mentation initially, was unable to develop a suitable approach for resolving problems with an uncooperative local population, and to communicate effectively (PPAM, para. 22); and - IDA supervision was inadequate in terms of frequency, type of skill used and response to persisting problems (PPAM, paras. 23-27; PCR, paras. 6.19 and 7.03). PROJECT PERFORMANCE AUDIT MEMORANDUM SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CREDIT 819-CE) I. PROJECT SUMMARY 1/ 1. The tree crops subsector, comprising primarily tea, rubber and coconuts, is Sri Lanka's main foreign exchange earner and employs over one million people. Tea has been grown in the country for nearly one hundred years and is concentrated in estates, although private smallholders are also involved and produce about one-third of total production. The estate work force is predominantly Tamil, who live in closely-built clusters and comprise citizens as well as non-citizens of Sri Lanka. Historically, tea cultivation has been extended into marginal areas, where production has tended to become unprofitable in view of the increasingly competitive international market situation. In some of these areas, tea bushes have been deteriorating, subjecting estate lands to the threat of erosion and general degradation. 2. In 1967, a Government commission reviewing the tea situation recom- mended a replanting program in suitable areas and diversification out of tea for marginal areas. A five-year UNDP project was implemented in the early 1970s to assist in the diversification of uneconomic tea and rubber areas. Subsequently, a diversification project was prepared by FAO/CP, which was rejected by an IDA appraisal mission in May 1976 on grounds that there were too many technical, economic and organizational uncertainties. Revised project proposals were reappraised in October 1977, and eventually accepted. The estimated project cost was US$6.5 million. IDA approved a Credit (819-CE) in the amount of US$4.5 million in June 1978. The Credit, which became effective in December 1978, was closed in August 1982 after cancella- tion of the undisbursed balance of about US$280,000. Because of Borrower action contravening provisions in the Credit Agreement, IDA decided to disassociate itself from the project and agreed with a Borrower proposal to prepay the disbursed amount in three installments during the period 1982-84. In accordance with that agreement, the Borrower had promptly repaid US$3.0 million up to October 1983; the balance of US$1.2 million is due in October 1984. 3. The objective of this four-year project was to diversify 4,600 ha out of tea, 4,900 ha out of scrub land and 1,800 ha out of degraded forests into more appropriate land use. In areas suitable for settlement, about 4,500 farms of 0.8 to 1.6 ha each were to be developed as mixed tree, tea or cardamom farms. Forest plantations and conservation forest were to be established in the remaining areas. The project provided for the construc- tion of simple settler houses, domestic water supplies, and access roads. The National Agricultural Diversification and Settlement Authority (NADSA) L/ Adapted from the PCR. - 2 - was the implementing agftncy. Designed as a pilot operation, the project was expected to be followed by a series of diversification projects. 4. Physical implementation, after some initial delays, proceeded rapidly, and disbursements remained slightly above appraisal targets throughout. However, signficant factors in the favorable disbursement picture were the vastly increased expenses for the construction of settler houses due to Government's decision to upgrade construction standards, and greater general cost escalation than anticipated at appraisal. Infrastruc- ture development, i.e., conservation measures, afforestation and road construction, was impressive. Farmstead and homestead development also had a good start and over 2,700 homesteads and 1,090 houses were completed as of April 1983. The shortfall against project targets resulted from the external forces that eventually brought the project to a halt and not from physical constraints on the part of the implementing agency. 5. Major adverse factors plagued the project from the beginning. Operationally and financially, NADSA was handicapped by its unforeseen assumption of responsibility for the existing estate labor force and for the tea factories when the 61 project estates were transferred under its authority. It had originally been anticipated that excess labor would be transferred to other public estates, and factory management assumed by other public entities. An immediate, and gradually growing, effect was the financial burden, which effectively stopped physical project activities in the end. The tea factories tended to be unprofitable, and the estate labor demanded legally justified gratuity payments as a condition for moving, for which NADSA had not received any funding. Labor relocation was also discouraged by the lack of accommodation at the intended recipient estates. / Another insurmountable obstacle turned out to be settler selection, for which inadequate procedures were established at appraisal. When selection began, it became apparent that Tamil participation was negligible. The virtual inability to move and the reluctance to participate as settlers galvanized project estate residents into a potent opposition with increasingly political overtones. 2/ OPS comments that the delays in transferring labor from the project area to other government estates were only the symptom of the more basic problem, namely that this resettlement scheme was poorly designed and implemented from the beginning. The design, according to OPS, was flawed by errors of judgment and lack of concern for its social conse- quences and implications. The Region points out, in contrast, that it had been agreed with the Sri Lankan government that the process of labor transfer would take place outside the project and not be internalized, as it subsequently was, through the involvement in this process of NADSA. Furthermore, concern for improvement of estate labor housing was the reason for including an investment component for this purpose under Credit 818-CE, which was appraised in conjunction with this project but encountered large implementation delays. -3- 6. Because of continual postponement of Government action on these crucial issues, project activities moved out of phase with each other, with implementable activities proceeding as planned while the others remained stalled. In this way, settler houses and farm plots were completed -before prospective settlers could be designated, i.e., in reverse order from appraisal expectations. The result was vandalism and extensive damage on the vacant properties. On the political side, the increasingly vocal opposition precipitated an enquiry by the Ministry of Agricultural Development and Research into the activities and performance of NADSA. The results of that enquiry were critical of NADSA and included recommendations to dissolve NADSA and transfer project lands to the Land Reform Commission (LRC) to assure their continued use as estate lands. 7. The return of land designated for diversification to estate cultivation was counter to the objectives of the project. When it was learned that six project estates had actually been transferred to LRC in February 1982, IDA reacted to this violation of the Credit Agreement by sending a mission to assess possibilities for completing the project under the new circumstances. However, these prospects were deemed unattractive in light of the sharply reduced economic viability and the existing local opposition to project activities. The fact that project lands with substan- tial infrastructure investments were unilaterally removed from the project's authority, with severely detrimental economic implications, led IDA to disassociate itself from the project. It suggested and obtained a request from the Government of Sri Lanka (GOSL) for terminating IDA's participation in the project; the undisbursed balance was cancelled and the disbursed amount to be repaid over a three-year period. 8. The benefits derived from the project are uncertain. The Govern- ment has indicated its intention to continue implementation on a modified basis, i.e., by settling a larger number of families on smaller farmsteads than envisaged. An economic rate of return on investments to date has not been calculated because of lack of data. The project did not materialize as was envisaged at appraisal. A meaningful evaluation would have presupposed an extensive information-gathering effort, but such an effort would have been premature in light of the delayed settlement and agricultural production schedule. However, the last review mission prior to IDA withdrawal estimated that completion of a truncated version of the project would have yielded marginal returns at best, and that the overall returns, i.e., including investments on the lands to be used for non-project purposes, would have been negative. II. AUDIT FINDINGS A. General 9. This project attempted to improve the productive structure of Sri Lanka's agriculture by shifting some of the marginal estate tea areas into smallholder cultivation of a variety of different crops. Under this endeavor, the existing inhabitants on the estates covered by the project were to be either settled as smallholders, repatriated to India or relocated, largely to other estates where some demand for additional labor existed. The objective of the project was sound.3/ The approach to diversification, i.e., smallholder settlement, remains uncertain, and the methods applied under the project proved unsuitable. 10. Experience with this project includes a number of positive elements. The project served to draw attention to the need for major adjust- ments in this area of the agricultural sector and helped to further identify the major difficulties as well as opportunities involved. It also showed that an implementing agency like NADSA can perform impressively in civil works construction and related activities. If good cooperation with other agencies can be established, as was the case with the Forestry Department, implementation problems can be minimized. The project also offers useful conclusions with respect to project design and organization. 11. However, these partial accomplishments and favorable aspects cannot compensate for the fact that overall project objectives as specified at appraisal were not achieved. Moreover, some project investments were lost when completed farmsteads and houses were left unoccupied for too long, and some crucial technical and financial assumptions with respect to smallholder enterprises could not be tested. There also remain a number of unanswered questions which render the feasibility of similar future investments doubtful. Major issues concerning this project are highlighted in the following sections.4/ 3/ OPS questions whether, in retrospect, the objective was sound, as it proved unrealistic. The audit maintains that the immediate objective of improving the area's productive structure (para. 3) was economically desirable. As to the impact on the welfare of the project area population, it was conceded at the appraisal stage that some people would be adversely affected although the overall net welfare gain was expected to be positive. The approach to diversification remains uncertain because, while crop diversification can take place under smallholder as well as estate conditions, the project demonstrated neither superiority of one or the other, nor under what conditions the establishment of a smallholder structure would be feasible and preferable. A basic precondition of the latter certainly would be a better understanding and accordance of more appropriate recognition of the human factors involved than was the case under this project. 41 Regional staff point out that additional issues, not fully covered in the PPAM, pertaining to a full understanding of the project's evolution and an explanation of the project's design, relate to the social f:ictions between the Tamils and Sinhalese and to the dichotomy caused by differing political philosophies of governments in power during project preparation and implementation. -5- B. Project Preparation and Design 12. The root causes of the failure of this project have to do with factors relating to the population in the project area. The audit suggests that these factors have not been given sufficient weight and therefore have been inappropriately dealt with all along. Among these factors are the living and settlement patterns of the local inhabitants, their attitude towards the project in general and their willingness either to participate in the project's settlement activities or to relocate elsewhere, and the political realities encompassing the interplay with local politicians and labor unions. The overwhelmingly unfavorable experience with this project is especially aggravating in light of the long pre-project history and early Bank association. 13. Involvement of external agencies began in 1969/70 when GOSL requested technical assistance from UNDP/FAO to develop a strategy to ensure optimal land use for marginal tea estates. A five-year UNDP project, "Agri- cultural Diversification of Uneconomic Tea and Rubber Areas", was initiated in 1970. From 1973 this project focussed on diversification out of tea and settlement of such land in a region which included what later became the project area for this credit. The UNDP/FAO project produced some 80 publications, including a terminal report (AG:DP/SRL/70/522) in 1977. 14. Bank Group involvement started when in June/July 1974 an Agricul- tural Policy and Program Review Mission reviewed the work of the UNDP/FAO project. This review was followed by an FAO/World Bank CP reconnaissance mission in March 1975 which led to an official communication by IDA later that year, informing GOSL of IDA's intention to assist in the preparation of an investment project through FAO/World Bank CP. As reported in the PCR (paras. 2.04-2.05), the project was appraised, first unsuccessfully, in May 1976 and again in November 1977. Board approval of the credit occurred in June 1978. The period from reconnaissance to Board approval thus spanned 39 months, and a very large amount of Bank staff time was spent on the project during that time. 15. Several features of the project's long pre-Board history are note- worthy. First, the UNDP/FAO project produced only superficial sociological insights as it had focussed mostly on technical and economic aspects. When this project was found unsuitable by the first appraisal mission, it was for organizational, policy, technical and economic reasons; interestingly, sociological aspects did not surface. For essentially the same reasons the project was designated as a pilot operation by the second appraisal mission. Second, between September 1976 and September 1977, when GOSL was engaged in preparing a revised proposal after the unsuccessful first appraisal, very little progress was actually made. This unimpressive performance casts some doubt on the commitment to the project of some of the agencies involved. Third, despite the poor state of preparation reported as late as September 1977, field appraisal was launched in October 1977. While additional preparation time might not have resulted in a better prepared project under the then existing arrangements, it is conceivable that some of the short- comings of the appraisal reflect deficiencies in preparation. Finally, - 6- irrespective of the quality of preparation, the appraisal suffered from the composition of the mission. There was no socio-political expert on the eight-member mission, which also appraised the Tea Rehabilitation Project, although instructions were to pay particular attention to the -social acceptability of the proposed pattern of development.- 16. Relatively little attention was paid to crucial questions relating to the posture of the project area population throughout the eight-year period prior to final appraisal. It is not surprising, therefore, to find t-is aspect inadequately treated in the second issues paper (dated 11/30/77) which addresses it as an equity issue rather than the socio-political one which it was. It is stated there that -non-citizen Tamils are not eligible for settlement and would be put out of employment faster than they would be through continued estate deterioration without the project.- It was con- cluded subsequently that hardship imposed on a segment of the population by the project was an acceptable social cost (PCR, para. 5.11). 17. In contrast to the formulation in the issues paper, this complex aspect went far beyond non-citizen Tamils and encompassed the whole project area population. The real issues were: (a) to what extent and under what conditions local people were willing to participate in the settlement program; and (b) how, where and under what conditions the excess population could be relocated. Experience ultimately showed that it was unacceptable to Tamils to be settled in a pattern which would disperse them throughout the project area. This aspect and a biased method of settler selection discouraged Tamil participa- tion. It also became apparent that relocation of non-settler families was stifled by unanticipated problems relating to gratuity payments, and by a lack of suitable housing at the recipient estates and of the required resolve and organization on the part of the authorities. 18. The major difficulties that were later encountered are clearly traceable to the way the project was designed: minority interests were not adequately protected, details and timing of labor transfer were not deter- mined (PCR, para. 6.09), and the responsibilities of participating agencies (e.g. the Janatha Estates Development Board and the State Plantations Corpo- ration) were not sufficiently specified and agreed in advance. Furthermore, NADSA's seven-member Board of Directors was ill-equipped to resolve conflicts with the estate population because it did not include representatives of all key parties and agencies concerned (PCR, para. 5.07). Some of the inadequate design features are apparent in the Credit Agreement, reflecting primarily the incomplete or inaccurate assessment at appraisal of these features rather than deficiencies in the design of covenants. Schedule 3 of the Agreement (Settler Selection and Land Allocation Procedures) does not reflect the Tamil preference for living in clusters separate from those of other ethnic groups, or the need for other provisions to attract minority participation (e.g. bilingual settler applications and instructions). Section 3.02 of the - 7 - Agreement posed the most serious implementation problems. It provided that The Borrower shall take all measures possible to provide, either through relocation to other estates or through accelerated repatriation, relief to displaced estate workers in the Project Area not eligible for settlement under the Project." 19. Another provision in the Agreement also proved inadequate, or in retrospect unsuitable for the intended purpose, and shows that it is risky not to resolve crucial organizational questions at the time of appraisal. Section 3.03 states that "The Borrower shall by November 1, 1978 make the necessary arrangements with NADSA to ensure the transport and processing of green leaf from the Project Area." As it turned out, in addition to inherit- ing responsibility for all existing labor on the estates transferred under its control, NADSA also, against expectations, had to take over the manage- ment of the tea factories and arrange for transport of tea. The managerial, and especially financial, strain on NADSA posed serious problems and dis- tracted attention, both within NADSA and IDA, from the socio-political issues. In the audit's view, a more in-depth treatment of these issues at appraisal would either have resulted in a better project design and more suitable implementation procedures, or conceivably would have brought to light that an acceptable solution to these problems actually was not at hand. C. Borrower Support and Borrower Agencies' Performance 20. After slow and unenthusiastic preparation of Government's project proposal, implementation started with a handicap with the assignment of responsibility for tea processing to NADSA. Another Government decision, which adversely affected the project's economic viability and NADSA's finan- cial situation, was the upgrading of construction standards for settler houses. 3ut the most serious problem was the inability to move the excess population out of the project area or to devise other means to defuse local opposition to the project. A lack of coordinated support among participating Government entities also hindered implementation. The consequences of the above factors were aggravated by Government's inadequate financial support for NADSA, given the cost increases related to the tea factory operations, the large labor payroll, and the extensive use by NADSA of overdraft facil- ities. 21. Interventions by local Members of Parliament in late 1981 led to a step by Government which contributed greatly to the withdrawal of IDA support for the project. An enquiry into NADSA's activities and performance was undertaken in January 1982 by the Ministry of Agricultural Development and Research (MADR). IDA was neither promptly informed of Government's inten- tions in this respect nor of the enquiry's findings and recommendations. The enquiry produced two recommendations which contravened provisions in the Credit Agreement: to dissolve NADSA, and to transfer control over project lands to other public entities. To dissolve NADSA was a reason to suspend the Borrower's right to make withdrawals from the IDA credit account under Section 4.01 (d) of the Credit Agreement, which calls for suspension in the - 8 - event "The Borrower or any other authority having jurisdiction shall have taken any action for the dissolution or disestablishment of NADSA or for the suspension of its operation." According to Section 5.01 (b) of the Credit Agreement "the transfer of exclusive control over the lands included in the Project to NADSA" was a condition precedent to effectiveness of the Credit Agreement, and any action to remove control thus was in conflict with the provisions of this Section. Furthermore, as sizeable investments had already been made on the lands designated to be transferred, this recommendation violated Section 2.03 (b) of the Project Agreement which stipulates that "Except as the Association may otherwise agree, NADSA shall cause all goods and services financed out of the proceeds of the Credit made available to it by the Borrower to be used exclusively for the Project". The nature and legal implications of these recommendations point to a desire by members of the enquiry to terminate the project in its then existing form. The instruc- tions of February 2, 1982 by the Secretary of MADR to transfer six estates to the LRC, issued again without prior consultation of IDA, made the violation of the Credit Agreement an accomplished, and for political reasons irrevers- ible, fact. In the audit's view, extensive consultation between the Borrower and IDA should have taken place preceding the above actions. Such consulta- tions could have been used to search for alternative solutions which might have included a modification of the project. 22. As far as NADSA is concerned, IDA staff repeatedly expressed satis- faction with its performance during the initial years of implementation. Physical implementation proceeded rapidly and disbursement was on target, although the latter was largely due to a change to a more costly housing design and to GOSL's application of unconventional reimbursement procedures (PCR, para. 3.25). However, in some respects NADSA management was overly accommodating and unrealistic. In retrospect, the initial assumption of responsibility for tea processing and the prolonged tolerance of the unre-. solved estate labor situation not only adversely affected the pruject in the end, but also cast an unfavorable light on NADSA when some of its improvised solutions were found to be irregular or unpopular. What recourse was actually available to NADSA considering that it was an agency of MADR, other than to solicit IDA intervention, is difficult to ascertain at this late stage, however. Concerning progress on settlement, communications with IDA tended to be too optimistic and not conducive to move IDA into taking appro- priate action. There are also indications that NADSA's relations with the local community, especially the labor unions and village leaders, were strained in part because of poor communications. But equally seriously, NADSA was in breach of Section 2.05 (b) of the Project Agreement for not informing IDA of the events of December 1981-February 1982 (para. 21). That Section specifies: "NADSA shall promptly inform the Association of any con- dition which affects or threatens to affect the progress of the Project, the accomplishment of the purpose of the Credit, or the performance by NADSA of its obligations under this Agreement." While NADSA undoubtedly accomplished much under the circumstances, it also contributed to the unfavorable outcome by trying to work around the serious problems it was facing rather than tackle them decisively at an early stage, and by its failure to appropriately communicate locally and with IDA. - 9- D. IDA Performance After Appraisal 23. IDA supervision began very timely in June 1978 and immediately identified aspects related to the growing social problems. These aspects continued to be observed by all supervision missions, though their emphasis was reduced at times, especially when NADSA's financial problems were partic- ularly acute or the missions obtained reassuring statements from high- ranking local officials. On all occasions, Government was informed of super- vision missions' findings through an aide-memoire left behind by the mis- sions, and a supervision letter subsequent to the completion of supervision reports. The settlement and labor transfer issues were repeatedly addressed in supervision follow-up letters, leaving no question of IDA and GOSL not being aware of the existence of these issues. It is not clear, however, whether their seriousness and intractability were correctly assessed either by Sri Lankan authorities or by IDA. 24. In retrospect, the effectiveness of project supervision could have been improved in two ways. First, the spacing of missions should have been more appropriate. After two missions each in 1978 and 1979, a full super- vision mission visited the project in March 1980. The next mission after that was primarily concerned with the preparation of a follow-on project. In connection with a review of the preparation status of that project, the mission also visited the ongoing project from July 3 to 7, 1980 and took the opportunity to update the IDA Supervision Summary (Form 590). Subsequently, a mission was not undertaken until October 1981. This supervision gap of 15 months (19 months if the updating mission is not counted) occurred during a crucial period and in all likelihood was most detrimental for the project. Except for disbursement matters, there was virtually no project communication between IDA and the Borrower between October 1980 and June 1981. Yet there were four covenants in the Credit Agreement requiring Borrower action by July 1, 1981. Progress in achieving compliance with these covenants should have been closely monitored and followed up on by IDA. The October 1981 mission found the project in disarray and deteriorating and was unable to forestall the events leading to Government's enquiry and resulting action shortly thereafter. 25. A second aspect related to supervision effectiveness was the treatment of and response to crucial issues. Supervision missions prior to October 1981 were generally satisfied with the project's physical accomplish- ments and, of necessity, focussed heavily on problems related to NADSA's financial situation. The issues of settler selection and relocation of estate labor, though consistently mentioned, were never explored in-depth by IDA staff. The fact that NADSA communicated poorly with some segments of the community was not discovered until the May 1982 review mission. Assuring information provided by GOSL and NADSA was a contributing factor to an inade- quate IDA response to a deteriorating situation. In retrospect, a more con- structive approach and affirmative action by IDA would have been called for. Instead of continuing to remind GOSL and the implementing agency, with little effect, of their obligations with respect to settler selection and labor relocation, a thorough review of these problems should have been undertaken by IDA at the latest after the third supervision mission in March 1979. Such - 10 - a review should have involved discussions with all interested parties, i.e., including the affected local population, labor unions and politicians. The objective would have been to identify measures and devise procedures for overcoming these problems. If agreement on appropriate modifications could not have been reached disbursements should have been suspended. Even using suspension without such a review, as alluded to in the PCR (para. 6.06), would have been preferable to the non-action course followed by IDA until February 1982. 26. On some occasions, a more forceful IDA intervention would have been beneficial for the project. For example, a more critical attitude was called for when GOSL decided to upgrade the housing construction standards. Lower standards, more consistent with appraisal specifications, would have reduced costs and increased the economic viability of investments, aided NADSA's financial situation and preserved IDA credit funds. Similarly, if transfer to NADSA of responsibility for tea processing and for the welfare of estate labor could have be-n prevented through timely IDA intervention, the serious adverse consequences for the project would not have occurred, or at least would have been mitigated. 27. Overall, IDA's actions during implementation were not entirely consistent with what could be expected with respect to a project which was "...of a pilot or test nature that hopefully [would] be the first in a series of diversification projects." (SAR, para. 2.06). Supervision was not outstanding either with respect to intensity, frequency, or resourcefulness in resolving special problems. There are indications that IDA was relatively concerned about and promoted the preparation of a follow-on diversification project. Such a project was mentioned to the Government after the March 1980 supervision mission, and the status of preparation was reviewed by IDA staff visiting the country in July and October 1980. With hindsight, these efforts were misdirected. Solutions to the key problems of this project had not yet been found, nor had the project generated the kind of data for which it was designated a pilot operation. In the audit's view, the project was super- vised more to, or even below, the norms of an ordinary project, although, as a pilot project it would have benefitted from a more intensive and construe- tive approach. A significant shortcoming was the failure to engage the needed socio-political expertise after it became apparent that this area was inadequately covered in the pre-project phase and during appraisal. 28. In conclusion, the audit basically agrees with the findings of the PCR as summarized in para. 7.03 (a)-(i), except for the format implied under point (c); the effectiveness of a sociological support unit is doubtful, and sociological expertise should preferably have been made available in a flexible but timely manner at all stages of the project cycle. Experience with this project also shows that key issues such as settler selection, labor transfer and cooperation with other agencies should have received more attention at the preparation/appraisal stage. Borrower assurances that necessary and appropriate action related to these issues would be taken during implementation were shown to be unrealistic. Consequently, the agreed actions and procedures, i.e., those formalized as covenants in the Credit Agreement, should have been reviewed by supervision missions as soon as - 11 - Borrower compliance became doubtful in order to either introduce modifications if necessary or devise supplementary measures in support of project implementation. Sections 3.02, 3.03 and 3.08 of the Credit Agreement were experiencing Borrower compliance problems and did require remedial action. This project also clearly demonstrates that every effort should be made to maintain good communications between IDA and the Borrower. Lack of consultation and unilateral action by the Borrower led to the withdrawal of IDA support for this project. 29. Overall, despite the fact that some lessons have been learned, the project failed to demonstrate how the crucial problems affecting this project could be solved as, in fact, they remained unresolved at the time of IDA withdrawal. The actual conditions and workable procedures for local partici- pation in settlement, and the complexities of relocating large numbers of families who cannot be accommodated as settlers are still without proven answers in Sri Lanka. Time was also too short to test key technical project assumptions, i.e., the smallholder cropping patterns, which had sparked controversy during preparation/appraisal. Consequently, the approach to diversify estate lands into multi-crop smallholder units must be considered as remaining at best experimental and risky. In this vein, the audit does not see any basis for the PCR's statements about possible salutary effects of IDA withdrawal on continuing project activities (PCR, para. 7.04). Although GOSL intentions in this regard are not questioned, the feasibility of their execution within the envisaged timespan is unsubstantiated. One drawback of IDA's complete withdrawal from this practical experiment to diversify agricultural production through smallholder settlement is that it cannot continue to monitor developments and evaluate the final outcome. Potentially valuable experience remains to be assessed and documented. - 12 - TABLE 1 PROJECT PERFORMANCE AUDIT REPORT SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CREDIT 819-CE) DISBURSEMENTS Category Credit Agreement Final Schedule 1 (07/12/78) Disbursement /a (US$ millions) (US$ millions) 1. Civil works and local development 1.00 3.42 2. Equipment, vehicles, animals and planting materials 2.30 0.25 3. Staff and operating costs 0.80 0.55 4. Consultants' services 0.10 /b 5. Unallocated 0.30 - Total disbursed 4.50 4.22 /c Amount cancelled - 0.28 /a As of August 20, 1982, the date of IDA's letter agreeing to the cancellation. /b Actual disbursement was US$3,089.09. /c As requested by the Borrower and in accordance with IDA's letter dated August 20, 1982, the amount disbursed will be prepaid. The prepayment is to take place within a three-year period, the first payment of US$1.5 million having been made on 10/01/82, and the balance to be paid on 10/01/83 and 10/01/84. The amount disbursed will, therefore, be fully prepaid prior to the first repayment date as originally indicated in the Credit Agreement. - 13 - ANNEX I My No. S/PI/29 (T) Ministry of Plantation Industries 349, Galle Road Colombo 3, SRI LANKA 26th Jan. 1984 Mr. Shiv S. Kapur Director Operations Evaluation Department World Bank 1818 H Street N.W. Washington, D. C. 20433 U.S.A. Dear Mr. Kapur, PROJECT PERFORMANCE AUDIT REPORT ON SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CREDIT 819-CE) I thank you for your letter dated 30th December, 1983, forwarding the Project Performance Audit Report and the Project Completion Report in respect of the Tree Crop Diversification (Tea) Project. I have read these two documents with considerable interest. I must, however, inform you that none of the organizations involved in this project now come under the Ministry of Plantation Industries. You are no doubt aware that from February 1980, several months prior to my assumption of duties as Secretary of this Ministry, the JEDB and the SPC were relocated under separate Ministries of Janatha Estates Development and State Plantation respectively under His Excellency the President. No documentation in regard to this Project is now available in this Ministry and this Ministry has had no functions whatsoever to discharge in respect of this project from February, 1980. In these circumstances, I am sure you will appreciate the fact that it is not appropriate for me to make any comments on these reports sent by you. Yours sincerely, I.O.K.G. Fernando Secretary Ministry of Plantation Industries 一／7一 戶么＇渺。右戶〞戶 15 CONFIDENTIAL PROJECT COMPLETION REPORT SRI IANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (Credit 819-CE) July 18, 1983 ---/c - 17 - PROJECT COMPLETION REPORT SRI LANKA TREE CROP DIVERSIFICATION (TEA) PROJECT (CR. 819-CE) I. INTRODUCTION 1.01 This is a Project Completion Report (PCR) on the Tree Crop Diver- sification (Tea) Project in Sri Lanka, for which Credit 819-CE for US$4.5 million was approved in June 1978. This PCR has been prepared based on findings of supervision missions which visited the project area shortly before the Credit was suspended in 1982. This information is supplemented by data collected by other missions which discussed the project at ministerial level subsequent to credit suspension. Although Government of Sri Lanka (GOSL) has provided all requested information, no direct input from GOSL is contained in the preparation of this report. 1.02 The project was intended to be implemented over six years; four years of field work and two years of consolidation of settlement activities. The project started well but after four years ran into insurmountable problems which brought implementation to a standstill. As a consequence, GOSL agreed to repay the US$4.2 million which had been disbursed. The Credit therefore closed in August 1982. Nevertheless, GOSL decided to continue project implementation without external financial assistance and introduced some changes in project design. It now expects to complete implementation by the end of 1984. A. Background 1.03 Economic growth in Sri Lanka depends largely upon performance in the agricultural sector, which accounts for 37% of GDP, 55% of total employment and about 80% of export earnings. The tree crops sub-sector, comprising primarily tea, rubber, and coconuts, dominates the agricultural sector and is the main foreign exchange earner and alone employs over one million people, more than a quarter of the total national labor force. 1.04 Tea, rubber, and coconuts are cultivated in Sri Lanka by both private smallholders and public estates. About two-thirds of all tea production in Sri Lanka originates from public estates, and no less than 80% of all tea produced is processed on public estates. An understanding of the organiza- tion of the estate sector and of the background to the social structure in the tea growing areas is crucial to understanding the outcome of the project. - 18 - B. The Estate Sector 1.05 Tea was planted in Sri Lanka in the late 19th century by individual estate owners desperately attempting to salvage something from the devasta- tions of coffee rust disease. Virtually every available square yard of. estate land was planted to seedling tea, including considerable areas of marginal tea land and which today continue to contribute little to overall production. 1.06 The growth of the estate sector in Sri Lanka from 1845 onwards led to the development of a largely imported labor force. This was a result of tea production being highly labor intensive and demanding regimentation and discipline from the work force. In the early days of tea growing, Sinhalese villagers were reluctant to submit to these requirements and indentured Tamil labor from South India was therefore introduced. The estate work force, headed by the superintendent, himself a foreigner, consisted mostly of Tamils with only a few local staff. This Tamil labor force had its own language, culture, and religion and contact with the surrounding Sinhalese villagers was minimal. In fact, considerable tension existed between these two social groups. As a consequence, the Tamils, being a minority, have always felt a greater sense of security by living in close-knit clusters and working as wage laborers on the estates. Many present day laborers are third or fourth generation estate workers whose whole life has been centered on cultivation. Employment usually includes the whole family, with the husband, wife, and frequently one or more children being employed on the estate. Wage rates are low when compared with industrial labor but fringe benefits such as free housing, medical attention, fuel, water, child-minding facilities and even a standard tea ration augment the family living requirements. Although Tamil labourers maintain kitchen gardens near their barrack-style accommodation, they have traditionally not been farmers. Instead, they have become tea specialists. 1.07 In areas less suitable for tea growing, old tea bushes are dying leaving exposed land and, consequently, soil erosion is increasing. This problem was exacerbated by the private estate owners, who, during the period before the enforcement of the Land Reform Laws, felt threatened by potential nationalization and proceeded to extract the maximum from their estates with minimal reinvestment. During the 1960s, low world tea prices caused a reduc- tion in estate profitability and further degradation of neglected tea lands. The effects of neglect and advanced soil erosion became particularly apparent in many areas of the mid-country. 1/ The project, therefore, found a weak international market and a deteriorating national tea industry employing mostly Tamil labor in estate enclaves. I/ The "mid-country" is variously defined as that part of the wet-zone having an average annual rainfall of 2,000-3,700 mm and lying between elevations 300 and 1,000 m. This covers about 1.5 million ha or, for tea industry statistical purposes, those tea estates lying between 600 and 1,200 m. - 19 - II. PROJECT FORMULATION A. Project Genesis 2.01 In 1967, facing the need to improve overall tea production and yields, OSL appointed a Commission to review the effects of a replanting replanting program to replace old seedling tea with high-yielding clonal material. The Commission concluded 2/ that the replanting program was neces- sary if the Sri Lanka tea industry was to remain competitive in the world market; as a corollary, it recommended the diversification of some marginal tea land to other crops, to prevent excessive supply and for the depression. of the world tea market. 2.02 GOSL determined that the region best suited for diversification was the mid-country sin4e (a) yields and quality of tea there were on average lower than elsewhere; and (b) many mid-country estates were therefore neglected and, consequently, soil erosion was accelerating so that these lands would soon be virtually useless unless remedial action was taken in this densely populated area. 2.03 In 1969-70 GOSL requested UNDP/FAO technical assistance to develop a strategy to ensure optimal land use to marginal tea land, based on technical and economic feasibility studies of alternative crops. A five year UNDP project, "Agricultural Diversification of Uneconomic Tea and Rubber Areas," was initiated during the 1970s to study the problems and make recommenda- tions. From 1973 this project focused on crop diversification on tea estates held by the Land Reform Commission (LRC) and possible settlement of such land by smallholders in three river catchments areas, Nilambe-Atabage (NAC), Gurugoda-Ritigaha (GRC), and Mahaoya-Kudaoya (MKC). The project produced no less than 80 publications but unfortunately these focused mostly on tech- nical and economic aspects and devoted little attention to in-depth sociological studies. Any such studies undertaken concentrated on demographic evaluation rather than attitude analysis. 2.04 Following conclusion of the project and at the request of GOSL, the FAO/World Bank Cooperative Program helped in the preparation of a project for Bank group financing and, in May 1976, an IDA mission appraised the project. The project as proposed was found to be unsuitable for IDA financing. The key problems were: 2/ Sessional Paper No. XVIII (1968). - 20 - (a) Uncertain yield performance of important crops; (b) Concern over the size of the small holding proposed; (c) Unknown response of settlers; (d) Incomplete identification of land available for settlement; (e) Uncertain market prospec :s for the diversification crops, mainly spices; (f) Absence of cost recovery proposals; (g) Unsatisfactory organizational arrangements; and (h) The inclusion of social and physical infrastructure proposals (hospi- tals, schools, markets, roads) that could not be justified economi- cally. 2.05 Following this appraisal, considerable pressure was maintained by the Bank on GOSL to produce revised proposals. The revised proposals resolved each of the outstanding issues, with the exception that yield data remained scarce and the response of settlers was a priori of course difficult to resolve. As a result, a modified project was appraised in November 1977. B. Project Objectives 2.06 The major objectives of the project were to: (a) Develop an institution which could implement a program of rehabilita- tion, diversification, and settlement by landless families on degraded tea lands, mainly from nationalized tea estates; (b) Halt erosion in uneconomic tea areas by introducing more appropriate farming and land use systems (perennial tree crops, pasture and forests) that would (c) Increase national agricultural production and income. 2.07 GOSL has recognized that, with appropriate use of improved technol- ogy, all the tea which Sri Lanka can profitably sell could be produced on a much reduced acreage. Over the years, the poorer tea areas, the obvious candidates for diversification, have suffered accelerating deterioration. Initially, this stemmed from reduced unkeep resulting from uncertainty over GOSL intentions regarding land reform. Later, and following land reform, it was the result of poor management. 2.08 The project was to be an initial attempt to diversify, rehabilitate, and reset tle about 11,500 ha of marginal tea land and other wasteland over a - 21 - four-year period. About 4,500 families were to be settled on one ha farm- steads on which they would grow mostly perennial crops such as pepper, cof- fee, cardamom, cloves, etc. Forests for fuelwood and timber would also be established and dairy cows would be provided to some farms. In addition, the entire area settled would be protected by soil conservation measures. The project was also to develop an institutional capability for executing larger, more comprehensive projects of this type in the future. 2.09 The project was of a pilot nature that hopefully would become the first in a series of diversification projects. The project was expected to generate data on the practical aspects of modifying land use and settling former estate workers and landless people. It was also expected to provide technical data on farming systems and crop/livestock mixes and which reflected the existing weak data base on yields of crops other than tea. C. Project Area 2.10 The project was located in the densely populated mid-country near Kandy, 90 km northeast of Colombo C1ap 13383). It encompassed parts of three catchments, NAC, GRC, and MKC in Kandy and Kegalle Districts. The total cultivable area was estimated to be 42,000 ha and project works were to be carried out on 61 scattered tea estates totalling about 11,500 ha of which 9,000 ha were cultivable. Rainfall and temperature were considered not to be limiting to crop production. The main features of the catchments are the steeply dissected and mountainous terrain and the wide range in elevation from 500 m to 2,000 m.a.s.l. Most slopes exceed 30%, many exceed 70% and some are precipitous. Rock outcrops are common and the soils vary greatly in depth and which produces variable crop growth. Under natural cover the soils have a pH of about 5.5 but leaching and extensive use of sulphate of amonia on tea have reduced pH to 4.5 or less. Soil Erosion 2.11 Following many years of poor financial returns and consequent neglect, most mid-country estate tea has many gaps. This, coupled with neglect of soil conservation, has tended to accelerate erosion with the development of many gulleys and extensive removal of top soil. The resultant silting up of rivers and reservoires is therefore serious. Social Setting 2.12 The total population living on the 61 estates was about 16,000 people, or about 22% of the total population in the project area. These residents were almost exclusively of Tamil origin but about half had opted for Sri Lankan citizenship, whilst the remainder had elected to be included in a repatriation program under an agreement between Sri Lanka and India. 1/ 11 In which for approximately every Tamil offered Sri Lankan Citizenship, two would return to India. - 22 - The traditional smallholder sector is dominated by Sinhalese living in small scattered villages between the estates. Many Sinhalese (particularly women) find casual work on nearby tea estates during peak work periods. The unemployment rate was 30% and of the 5,000 joining the labor force yearly few could find work. The number of days work per month offered to labor, a significant indicator of estate viability, had decreased and many families were existing on less than US$150 per year. The average family income of estate workers in 1978 was estimated to be SLRs 3,000 (then US$187.5) per year, a figure used to calculate Government support to settlers during the crop establishment period. Land Tenure 2.13 About 40% of the project area consisted of village lands and tradi-. tional settlements and the rest was mostly tea land owned by the Land Reform Commission and administered as tea estates by the public sector State Planta- tions Corporation (SPC) and Janatha Estates Development Board (JEDB). Individual private smallholdings in village areas are small, with 50% being less than 0.5 ha. The size of the 61 estates ranged between 30 ha and 400 ha. Diversification Crops 2.14 The project relied heavily on the possibilities of establishing a wide range of crops on degraded tea lands. Mixed forest gardens (the renowned "Kandy" gardens) are a feature of the area but occupy distinct pockets around homesteads. These gardens may contain 20 or more crop species ranging from vegetables to large trees, haphazardly arranged in a garden area averaging about 0.5 ha. These gardens are poorly managed with no regard to efficient production, receive little or no fertilizer other than household waste, yet frequently show impressive vegetative growth. Crop yields are poor relative to the potential for individual crops but total production satisfies farmers' needs for food and marketable surplus with little investment and with little risk of crop failure. Contrasting with the important estate crops, research on improving crop mixtures and even efforts to collect yield information from the component crops have been minimal. Some yield informa- tion was collected by FAO consultants but no trials work under controlled conditions had been carried out. Subsequent to project establishment a field trial together with availability of planting material was carried out by the Minor Export Crops Research Institute at Delpitiya in the project area. Experience on this trial together with availability of planting material resulted in modification to the project crop models and the first authenti- cated yield data for most crops were collected after project appraisal. The pilot project was therefore established on uncertain production estimates. The modifications to the crop model for a Low Altitude Mixed Tree (LANT) farm are reflected in the table below. - 23 - Table 2.1: LANT FARM CROP MDDEL Species Number of Plants Appraisal (0.8-1.0 ha) Modified (0.7 ha) Banana 16 50 Papaya 12 0 Coconut 22 25 Avocado 4 0 Coffee 30 156 Arecanut 40 0 Bread Fruit 4 0 Jak 8 0 Gliricidia 40 380 Albezia 4 0 Clove 9 12 Nutmeg 4 0 Other Trees 20 18 Pepper 80 380 Ginger 0.8 ha 0 D. Project Description 2.15 The project was to diversify 4,800 ha out of degraded tea, 4,900 ha out of scrub land and 1,800 ha out of degraded forest into more appropriate land use. In areas where the land was suitable for settlement the project would develop two types of mixed tree farms, 1/ tea farms (based on pockets of tea still in economic production), and cardamom farms; in areas unsuitable to tree crops, forest plantations and conservation forest would be estab- lished. About 4,500 farms of 0.8 to 1.6 hectare would be demarcated, planted with appropriate perennial crops and allocated to landless families from the project catchments. Simple settler houses would be built in clusters, for which water supplies and, where necessary, access roads would be constructed. The project would be carried out by the National Agricultural Diversification and Settlement Authority (NADSA), which was set up by GOSL to carry out this and future projects. NADSA would implement all project works using the prospective settlers as its labor force. 1/ High Altitude Mixed Tree Farms (HANT farms) and Low Altitude Mixed Tree (LAMT) farms. - 24 - 2.16 The project would comprise: (a) Development of Mixed Tree Farms: (i) 2,550 LAMT farms below 750 m, on 2,800 ba; (ii) 1,300 RANT farms with pasture and dairying between 750-1,000 a, on 2,000 ha; (iii) 390 cardamom farms (CF) on 700 ha; and (iv) 230 tea smallholder farms on 180 ha. (b) Social Infrastructure. Construction of simple houses located in clusters, access roads, and village water supplies. (c) Forestry. Establishment of about 1,800 ha of pines, 1,900 ha of eucalyptus and 2,100 ha of conservation forest. 2.17 The project would change the existing land use as follows: Table 2.2: EXISTING AND PROJECTED LAND USE (ha) Existing With Project Tea 4,800 2,800 LAMT Farms Scrub Land 4,900 2,000 HAMT Farms Degraded Forest 1,800 700 Cardamom Farms 180 Tea Smallholder Farms 3,700 Commercial Forest 2,100 Conservation Forest All land presently under tea would be used for LAMT farms, HAMT farms and tea smallholder farms. Scrub land would be converted into HAMT farms (230 ha) and the balance into commercial and conservation forest. Degraded forest lands would be used for establishing cardamom farms (700 ha) and for commer- cial forest. The Preparation Report proposed a 5-year project with a total cost of SLRs 77 million. At appraisal these plans were modified to a four-year pilot project to be followed by two years of consolidation during which GOSL would continue supporting the settlers. The cost of the appraised project was SLRs 104 million and contained no major changes from the prepara- tion document. The cost difference largely resulted from changes in the exchange rate, from greater support for NADSA and an estimated need for higher contingencies. - 25 - 2.18 Extensive manpower resources were devoted to the identification, preparation, and appraisal of the project. This was a result of the complete change of agricultural policy in the area, which was associated with the need to diversify tea estates, as well as from the al.ost complete lack of techni- cal, demographic and economic data, and from the belief that the pilot project would develop ways of overcoming the social and technical problems of the mid-country. E. Negotiations and Effectiveness 2.19 Negotiations were held in Washington in April 1978 and attended by the Chairman and General Manager of NADSA together with the Additional Direc- tor, External Resources Division of the Ministry of Finance and Planning. No special difficulties arose. In addition to normal documents two matters important to the project were agreed upon (a) the transfer of exclusive control over Project area lands to NADSA; and (b) the Recovery of Project Costs. Land Transfer 2.20 Detailed Procedures for land transfer were set out in a side letter which established that "exclusive control of land covered by the Project area shall be transferred to NADSA as a condition of effectiveness". It was, however, stated that "management of these lands shall be transferred to NADSA according to a transition plan developed by the concerned agencies" (later specified as SPC, JEDB, Government Agents of Kandy and Kegalle Districts, the Director Land Alienation of the Land Reform Commission, and the Chairman of NADSA). It was specified that "Until the management of the lands in the project area is turned over to NADSA under this plan, those agencies cur- rently managing the lands will continue to do so...". A schedule for trans- fer of management of project lands was laid down and completion of the trans- fer was expected within 6 months. The dates set out in the schedule were to be "used by each concerned agency as a basis on which to make, as appropriate, necessary arrangements to provide for alternative employment for estate labor ineligible for settlement, and estate clerical and management staff." 2.21 A crucial point which had a direct bearing on future problems in the project concerned the operation of estates. It was stated in the side letter agreed at negotiations that the committee of concerned agencies would "estab- lish which tea factories in the Project Area would remain in operation and which ones would be phased out of operation. While factories continued to operate, they would remain under the management of their current operators." Furthermore, factories would continue to provide greenleaf transportation to NADSA and tea smallholders under current procedures and conditions. These various specifications amounted to a plan whereby NADSA would have total - 26 - control over the lands, SPC and JEDB would remove all excess labor and staff but would continue to transport and process all tea grown in the area. This plan did not succeed. III. PROJECT IMPLEMENTATION AND PERFORMANCE A. An Overview from Supervision Missions Perspective 3.01 The project became effective on December 15, 1978. Even before effectiveness, in June 1978, a supervision of the project start-up activities revealed several issues. Staff of the estates which were designated for transfer to NADSA, although assured of continued employment for one year, were concerned about their employment status thereafter. As a consequence, many began bringing pressures to bear to have their estate taken off the list of those to be diversified. Subsequently, SPC began arguing for the withdrawal of some estates from the program because the last two years had shown tea yield increases to the point where these estates were no longer eligible to be diversified. However, a review of estate records revealed quite the contrary: that yields had steadily been declining. 3.02 Thus began a period of uncertainty with respect to which estates would or would not be included in the project. This uncertainty delayed the process of prequalification of settlers. The supervision mission identified this problem and warned that the situation could lead to labor unrest. It also warned that excess labor, resulting from discontinued replanting activities and cultivation activities in the project area, should be moved out of the project area in full consultation with the labor union, to avoid any confrontations. This mission also identified managerial frictions between the General Manager of NADSA and the Minister concerned. It should be reemphasized that these problems were identified before credit effective- ness. 3.03 A second mission visited the project in September 1978. At this point, the transfer of land to NADSA, the only condition of effectiveness, had not yet been finalized. The credit effectiveness date had now been extended from the original July 1, 1978 target date to November 30, 1978. It did not become effective until December 15, 1978. The mission sought and received assurances from the Minister that the settler selection process would be apolitical and that the criteria agreed upon with IDA would be followed. At this point, JEDB and SPC were said to be in the process of working out an agreed time-phased plan to transfer non-citizen labor from project estates (para 5.12) to other estates. NADSA had already been actively recruiting staff, and had also begun project procurement. 3.04 The mission discussed the problem facing NADSA of having been forced to take over the management of tea factories on those estates transferred to NADSA, as well as the estate labor force which, contrary to agreement, had not been transferred to other estates. This, in spite of the fact that it was - 27 - explicitly agreed to earlier that SPC and JEDB would continue to be respon- sible for these factory operations and would absorb the excess labor (para 2.21). However, SPC and JEDB refused to continue management responsibilities on these estates. The heavy financial and managerial burden on NADSA of running these factories was apparent. The mission sought and received. assurances from the Minister that special funding would be arranged for NADSA to cover these unwarranted expenditures. However, no such funding actually took place. 3.05 The mission clearly had identified the emerging socio-political characteristics which would ultimately shape the project into a form unaccep- table to IDA. To quote the observations of that mission in its supervision report: "The project area is in the densely populated mid-country near Kandy, where land pressure is severe and where the plantation sector with a mostly Tamil labor fcrcc and the smallholder sector of Sinhalese coexist. In the past and up to the present pressure by local politicians (Members of Parliament) has been brought to bear on NADSA to revert some of the project lands to their previous owners. Equally, pressure in the hiring of NADSA staff is applied and is likely to be applied in the settler selection process. To ensure that NADSA is given enough national political backing to resist these pressures, the mission met with the Minister of Agriculture Development and Research and raised these points (see Aide Memoire to the Minister, Annex 3). In a frank discussion with the Minister the mission received his assurance that he will do all in his power to protect NADSA from these local political interferences." 3.06 A third supervision mission was fielded in February/March 1979. Surprisingly, this mission report focused more on the physical progress in project implementation which was already beginning to pick up impressively. Even though the problem of NADSA having to manage the tea factories remained, no mention was made about the financial impact of these operations on NADSA's budget. The mission report did state, however, that "Now with 500 homesteads (plus 1/4 ac vegetable plots) nearly complete, the most critical phase of the project (a) transfer of non-citizen labor to other estates; and (b) settler selection and allocation of houses, is about to begin. Given that the project area is in the densely populated mid-country near Kandy where land pressure is severe and where there are frictions between the plantation sector with a mostly Tamil labor force and the smallholder sector of Sinhalese, this phase of the project is bound to be difficult and will probably have political overtones." - 28 - 3.07 The mission also noted the emerging problem of transferring excess labor to other estates. It stated that in principle JEDB and SPC had agreed to absorb the non-citizen excess labor, but no detailed, time-phased plan to implement this had been worked out. In addition, two other problems affect- ing this aspect had surfaced. Gratuity payments for past services due to the laborers had now become a NADSA liability (with the forced take over of the factory responsibilities and the estate labor force) without corresponding funding provisions to NADSA. And even if NADSA had the money, estate employ- ment records were not available to calculate these payments for individual workers. In addition, it was becoming apparent that JEDB and SPC did not have adequate accommodation to absorb the excess labor as originally believed. Again, GOSL assured the mission that these issues would be dealt with satisfactorily. 3.08 In August of 1979, a fourth supervision mission visited the project. The mission reported that all physical works, except housing, were on target. It stated that "the main problem at present is the greatly increased project costs and NADSA's difficulty in meeting these expenses without reducing the pace of project implementation." The mission was assured that a revised NADSA budget proposal would be submitted to the Cabinet by the end of September 1979. The report also stated that the "settlement aspects (settler selec- tion, transfer of noncitizen labor to other estates, and land allocation) had just started". It further reported: "Firm agreement has been reached between NADSA's management and the supervision mission that by the end of this year, these activities should be finalized and settlers identified with their homesteads, even if houses have not yet been constructed." 3.09 The fifth supervision mission visited the project in February/March 1980. It reported that settler selection had begun and was then said to be scheduled for completion in July of 1980. In comparing assurances on this aspect given previous missions, to progress actually made, it was clear that the settler selection process was beginning to run into serious problems. The movement of non-citizen labor out of the settlement area was also progressing slowly. Thus, the forces of political confrontation were begin- ning to intensify. 3.10 Another mission in July 1980 reported that the two major obstacles to project success were now inadequate funding for NADSA in light of cost overruns and resolution of the outstanding labor issues. It is interesting to note that settler selection and land allocation activities had progressed little since the last supervision and that even though the August 1979 mis- sion had received assurances that the settler selection process would be finalized by 1979, the finalization of this process was not yet in sight even by July 1980. 3.11 The next supervision mission to visit the project was not until September/October of 1981; more than a year later. This mission found that the settler selection process had been totally stagnant for the past year and that the project continued to be plagued by the familiar three problems that had been haunting it almost from the outset: - 29 - (a) a shortage of funds, (b) political opposition to settlement at electorate level, and (c) continued responsibilities forced upon NADSA for activities not originally part of the project (management of tea factories, etc). But the problem with settler selection and political involvement now had brought about another problem. Because the work on farmstead and homestead development had continued, without the corresponding settlers taking over and tending and protecting the established settlements, vandalism and natural degradation had resulted in loss of much of the project fieldwork achieved thus far. The mission warned that the project had reached a critical stage and that implementation of recommendations made in its report had to be verified within six months to ascertain whether the project would have a chance of success. 3.12 The NADSA Chairman at that time assured the mission that settlements would be fully allocated by the end of 1981. To back their statement, NADSA management produced copies of a Presidential directive supporting the project and the settlement program and which had been issued shortly before. Never- theless, local political opposition continued and local MPs apparently received strong indications from their constituencies to continue to inter- fere in the process. Also, labor unions representing the estate workers continued their opposition to the project. Thus, the President received strong complaints from these MPs, as well as the Labor Union Presidents, 1/ and, as a consequence, he instructed the Ministry of Agricultural Development and Research to undertake an enquiry into the activities and performance of NADSA. This enquiry, which took place in mid-January 1982, criticized NADSA for not keeping the original targets set at appraisal, which had, therefore, defeated the objectives of diversification and settlement. The report fur- ther recommended the dissolution of XADSA and the immediate handing back to the LRC of 17 of the 61 project estates. 3.13 A brief mission in February of 1982 discovered that, on the basis of the recommendations of the ministerial enquiry, the Secretary MADR issued instructions on February 2, 1982 to transfer six estates within the project areas to the LRC in a blatant violation of the DCA. The implication of this action was that considerable project expenditures incurred on these estates and funded with project/credit funds had now been transferred outside the project area, directly violating basic principles of the Development Credit Agreement. In addition, NADSA management was indicating that no frv*h-mr land allocation would be possible in the then existing political climate. The mission made several unsuccessful attempts to meet with the Minister to discuss these developments. 1/ The president of one of the major labor unions also harpened to be the Minister of Rural Industrial Development. - 30 - 3.14 Subsequently, a mission visited the project in April/May 1982 to make an assessment of the possibilities of completing the project in a way accept- able to IDA. Recognizing IDA's serious concern about this project, GOSL expressed a commitment to complete a truncated version of the project, excluding the recently transferred areas. The proposal comprised a target for 3,150 settlement units to be completed by December 31, 1982. The mission agreed with the concept that, politically, the areas transferred to the LRC could not be returned to NADSA. It also suggested that NADSA be given an opportunity to undertake the proposed truncated program. However, IDA management did not feel that this program was achievable in light of past experience and existing circumstances. In addition, the transfer of land out of the project area and lost project expenditures through degradation remained unacceptable and had contributed to the project having become uneconomic. Consequently, IDA suggested and obtained a GOSL request for terminating IDA's participation in the project and for an accelerated repay- ment by GOSL of the disbursed credit funds. B. Project Components Performance 3.15 The following paragraphs review the performance of the individual project components compared to project objectives. Whenever applicable, the revised program (which GOSL indicated it intended to pursue) is also indi- cated. Farmstead and Homestead Development 3.16 The project objective was to establish 4,500 farmsteads and homes- teads over the four-year project period. Farmsteads would consist of mixed tree crops, cardamom, or tea. Homesteads would be planted with garden vegetables or paddy when the terrain allowed. Each HAM farm would be provided with one cow. These farms would have 0.3 ha of pasture land. The project envisaged providing 1,650 dairy heifers by the beginning of project year three. During the June 1981 visit to the project by the Secretary, MADR, he reported that there were 254 animals at the Kirimetiya Estate and that this heifer raising unit had already cost SLRs 1.163 million, and recom- mended that the unit be closed down and the animals distributed among set- tlers. This recommendation was subsequently adopted and no further distribu- tion of heifers has taken place. In the early stages, farmstead and homes- tead development was taking place at a very accelerated pace. Availability and quality of planting material was reported to be good. Already in March 1979, less than three months after credit effectiveness, the supervision mission reported that cluster selection and homestead and farm demarcation was nearly complete and that soil conservation work was well under way. By September 1979, all physical works, except house construction, was reported to be on target. The progress in farmstead and homestead development and the related soil conservation works over this period was remarkable. 3.17 The intention was to complete the settler selection process and then have the selected settlers form the labor force which would develop their own - 31 - farm and homesteads. This would serve both as a means to maintain a sense of connection between the settler and the work carried out, as well as to provide a temporary form of income before they were settled. Failure to implement the settler selection process was a major impediment to successful implementation of the project. Because of the uncertainties related to the settler selection process, there was little or no identification of the labor force with the settlements being developed. As a consequence, there was little incentive to protect works once completed. As of April 1983, one and a half year after farmstead and homestead development should have been com- pleted and allocated, only 890 farmsteads and 1,648 homesteads are complete and occupied by settlers (see Table 3.1). 3.18 GSL now plans to simply finalize settler selection, allocate demar- cated farm and homestead lands, and then supply free planting material for settlers to plant up their own areas. In addition, the average farmstead size will be reduced to about one acre, rather than the originally proposed two acres, to accommodate a greater number of settlers. Also, homes are now supposed to be built on the farmstead itself rather than in homestead clusters, as originally intended in conformity with traditional practice. 3.19 Table 3.1 below summarizes the physical status and proposed program for completion. It indicates that the new target is for a total of 5,430 farmsteads over 7,286 acres and that 890 farmsteads have already been estab- lished based on the old model. The remaining 4,540 farmsteads will have an average size of about one acre. Some 2,715 homesteads have been established, of which 1,648 have been occupied. The remaining 2,715 will be established through payment of a subsidy for self-help development. - 32 - Table 3.1: STATUS OF NADSA PROGRAM AS OF APRIL 1983 (net of any transfer of land and works) a/ Occupied by To be Completed Completed Settler and Occupied Total Acres No. Acres No. Acres No. Acres No. HAM & LAM Farms 5,263) Cardamon Farms 685) 3,034 2,858 890 4,428 4,540 b/ 7,286 5,430 Tea Farms 658) Homesteads 680 2,715 412 1,648 267 1,067 680 2,715 Forestry 4,638 - - - - - 4,638 - Houses Constructed - 1,090 - 987 - 534 - 1,624 c/ Water Supply to be completed in 1984 ---- a/ The following was handed over to LRC: b/ 1/3 ac units 72 HAM & LAM farms - 721 ac 3/4 ac units - 1,873 Tea Farms = 836 ac 1 ac unit = 2,074 Forestry - 2,573 ac 1-1/4 ac units - 6 1-1/2 ac units = 96 1-3/4 ac units = 73 2 ac units- 13 4,540 c/ An additional 3,734 self aided houses to be built at SLRs 5,000 subsidy per unit. Social Infrastructure 3.20 The project originally aimed to provide basic shelter to settlers in the form of semi-permanent houses costing about Rs 1,100 or (US$69) and built on a 0.2 ha homestead. However, soon after project inception, GOSL expressed a desire to change the hcusing design to a more permanent model. The result was a four and a half fold real increase in the housing unit cost. GOSL gave assurances that since the decision to build more elaborate houses was a government one, NADSA would be provided adequate funds to cover the increased costs. Additional funding of SLR 24 million was provided to cover these costs. But delays in construction, due mainly to the more elaborate design requiring more professional contractors, along with rapidly rising construc- tion costs, put additional pressure on project funds. 3.21 The project anticipated finishing construction of 4,500 units in 1979. By April 1983, only 1,090 units bad been completed, of'which 987 were occupied by settlers. An additional 534 units remain to be completed. Under - 33 - its revised programs, GOSL now anticipates providing financial support, up to SLRs 5,000 per unit, for construction of self-aided houses by settlers for the remaining 3,734 units. The water supply system to homesteads was to consist of 540 wells and 180 water distribution points through diversion of small streams. Only about 20 wells and 40 distribution points have been completed. The revised program calls for another 250 wells. It is to be completed in 1984. GOSL has given no indication as to how it plans to provide settlers with water in the intervening period. Roads were built mostly on schedule, but the cost of maintenance turned out to be greater than anticipated at appraisal. Because of that, upkeep has been below what is necessary to keep the roads in usable condition and, consequently, many of the access roads built under the project have degraded considerably. Forestry 3.22 The forestry component envisaged planting a total of 3,700 ha. As of April 1983, 2,920 ha have been planted. This represents 79% of project target. No further planting is planned under the project. As envisaged at appraisal, the forestry component was sub-contracted to the Forestry Depart- ment and was implemented with little problem. As with all other components, funding became the major constraint, particular in later years. Since the implementation of the forestry component did not depend on the settlement program, it took place in a semi-independent setting. It was not subject to vandalism. Consequently, final overall success rate was much higher than for the other field work activities. C. Project Costs 3.23 Actual project costs are summarized in Table 3.2. Cost figures are denominated in 1978 constant rupees to compare with appraisal estimates. It is seen that actual base costs exceed appraisal estimates by 57%. Farm development and social infrastructure are the main areas of cost overrun. In farm development, the cost overrun was caused by unanticipated upkeep of established farmsteads and duplication of farmstead development activities. Delays in settling farmers on the established areas meant that NADSA had to continue incurring costs for upkeep of these areas and in some cases repair the damages of vandalism and natural erosion which often involved having to reestablish the whole farmstead. The main reason for cost overruns in social infrastructure was GOSL's insistence, after the pr-ject had begun, on a drastic change in the design of the houses to be constructed. What was originally conceived as a SLRs 1,500 temporary cottage, became a SLRs 6,800 semi-permanent house. The reason actual total housing costs are not four and a half time appraisal estimates is that the nousinx program itself was dras- tically reduced. Total project costs, inclusive o.. contingencies, are more than twice the appraisal estimates in rupee terms, while in terms of US dollars they exceed appraisal estimates by about 72. Details of actual costs by year are presented in Table 3.3. - 34 - Table 3.2: PROJECT COSTS Estimated Appraisal Estimates Actual Actual as Local For- Total Local For- Total Percentage eign I eign b/ of Appraisal Estimate NADSA Capital Costs 3.2 5.1 8.3 3.7 6.0 9.7 118 Operating Costs 8.1 6.7 14.8 a/ 12.2 5.4 17.5 118 Farm Development 26.3 14.0 40.3 88.8 22.7 111.6 277 Forestry 7.5 2.2 9.7 8.2 1.5 9.7 100 Social Infrastructure (Housing, etc.) 11.0 0.5 11.5 24.9 14.4 39.2 341 Other Labor Costs (EPF, etc.) - - - 10.1 - 10.1 - Settler Allowances - - - 16.3 - 16.4 - Labor Learning Incentives - - - 0.5 - 0.5 - Total Base Costs c/ 56.1 28.5 84.6 164.7 50.0 214.7 254 Contingencies 15.9 3.5 19.4 n.a. U.a. n.a. .a. Total Project Costs 72.0 32.0 104.0 164.7 50.0 214.7 205 (US$ Million) 4.5 2.0 6.5 8.6 2.6 11.2 172 al Includes settler allowances. b/ Estimates. Precise figures not available. cy/ Base costs for appraisal estimates in constant 1978 SLRs million. - 35 - Table 3.3: ESTIMATED ACTUAL EXPENDITURES BY YEAR (SLRs million) Projected 1978 1979 1980 1981 1982 1983 1984 Total Capital and Operating Costs 1.6 4.0 5.4 5.8 4.2 2.2 2.2 25.4 Farm Development 6.5 23.8 18.3 17.5 24.5 9.5 - 100.1 Forestry 0.4 1.4 4.2 1.9 0.4 - - - 8.3 Social Infrastructure - 0.2 6.6 8.9 8.6 8.3 12.2 44.8 Other Labor Costs - - - - - 12.3 1.1 13.4 Settler Allowances - - - - 1.8 9.8 10.6 22.2 Labor Leaving Incentives -- - 0.3 0.1 0.1 0.5 Total 8.5 29.4 34.5 34.1 39.8 42.2 26.2 214.7 D. Cost Recovery 3.24 In order to recover a part of the costs of the project, GOSL was expected to complete, by July 1, 1981, all measures to enable the Borrower to assess, impose and collect charges satisfactory to the Association with due regard to the settlers ability to pay such charges (DCA-Section 3.07). The Agreed Minutes of Negotiations set out calculations for determining a "project rent" which took into account management and risk costs, and valuing family labor assuming that "two members from the family could work elsewhere at Borrower's minimum wage for about 265 days in each year." The heavy finan- cial burden on settlers in the early years was recognized at appraisal and three ways of overcoming this were proposed: (a) including fast growing crops in farm budgets; (b) providing farm inputs from NADSA with costs to be recovered from the rent; and (c) employing prospective farmers as paid laborers during the first six years. No system for cost recovery was actually prepared during project implementa- tion. Because of this, none of the few settlers allocated a farmsead ever received ownership titles to their land. As was planned, instead, they were given temporary lease agreements, until an acceptable cost recovery mechanism could be worked out. As of April 1983, no such mechanism had yet been prepared. - 36 - E. Procurement and Disbursements 3.25 Procurement early in project implementation was slow. This was caused by the then recently introduced changes in GOSL procurement policies requiring procurements of above SLRs 1 million to be referred to the Cabinet. Nevertheless, procurement accelerated rapidly during the latter half of the first year of the project. By the end of the first project year, most procurement had been finalized. Thus, except for some delays early on, procurement performance was good. It was projected at appraisal that disbur- sements would take place over five years beginning in the second half of IDA FY79. Actual disbursements were rapid as compared to average IDA projects and remained very close to appraisal estimates. After the first two years, US$1.7 million or 38% of the total credit had already been disbursed, which was slightly above the target disbursement level projected at appraisal (see Table 3.4). By the end of FY82, shortly before GOSL and IDA agreed to suspension of disbursements and accelerated repayment of the credit by GOSL, US$4.2 million had been disbursed (or 93% of the credit). The reasons for the rapid rate of disbursements were a dynamic onset of the project with respect to field development, housing activities, and procurement combined with costly design changes of the houses to be constructed and other faster than anticipated cost increases. But also important was the arrangement adopted by GOSL, contrary to the DCA, for reimbursements of project funds in which the executing agency (NADSA) obtained reimbursement withdrawals directly into its own account. Its budgetary grant allocation consisted only of the GOSL component. Inherent in this arrangement is the incentive mechanism which assures that more funds are available to the executing agency the quicker it processes its disbursement withdrawal applications. This contrasts sharply with other projects in Sri Lanka where disbursements are paid to the Treasury and where funding to the project executing agency is in the form of budgetary allocations, which are generally independent of the agencies disbursement performance. - 37 - Table 3.4: DISBURSEMENTS BY YEARS (US$ Million) IDA Appraisal Estimates Actual Cumulative Actual FY Period Cumulative Period Cumulative as X of Estimates 79 0.3 0.3 0.3 0.3 100 80 1.:! 1.6 1.4 1.7 106 81 1.6 3.2 1.6 3.3 103 82 0.9 4.1 0.9 4.2 102 83 0.4 4.5 F. COMPLIANCE WITH COVENANTS 3.26 The following paragraphs discuss the extent to which dated and specific Development Credit Agreement (DCA) and Project Agreement (PA) covenants were adhered to in the process of project implementation. Number Covenant Comments DCA, Section 3.01(b) "The Borrower shall make This was not complied proceeds of the Credit with. See para 3.25. available on a grant basis to NADSA under arrangements to be entered into between the Borrower and NADSA upon terms and conditions which shall have been approved by the Association." DCA, Section 3.02 "The Borrower shall take Efforts to relocate all measures possible to estate workers proved provide, either through ineffective. IDA relocation to other estates missions were repeatedly or through accelerated assured that relocation, repatriation, relief to though delayed somewhat, displaced estate workers would proceed with little not eligible for settle- problem, being properly ment under the Project." coordinated between relevant ministries. - 38 - DCA, Section 3.03 "The Borrower shall by Contrary to understandings November 1, 1978, make reached at negotiations, the necessary NADSA was forced to continue arrangements with NADSA running tea factories and, to ensure the transport therefore, was providing and processing of green transport for itself. This leaf tea from the burden was of great Project Area." significance in hampering NADSA from concentrating its attention to what was its primary task, that of developing farmsteads and social infrastructure. In addition, it implied a great financial burden in terms of staff payroll and unprofitable tea factory operations. DCA, Section 3.04 "The Borrower shall, These cooperatives were before dairy animals never established. It was are sold to farmers claimed that there was not in the Project Area, sufficient business at the but in any event by time. Instead, settlers July 1, 1981, would sell milk through establish an authorized private appropriate number of collectors who collect milk dairy cooperative for the M[ilk Board. This societies with system appeared to function by-laws acceptable to satisfactorily in the the Association." limited scale of its operation. DCA, Section 3.05 "The Borrower shall, if Never became applicable yearly incomes of the since practically no settler families in the settlers were settled. Project Area are found by NADSA to be below SLRs 3,600 in constant 1978 Rupees, promptly take necessary measures, acceptable to the Association, to ensure that settlers have the financial ability to continue to develop their farms." - 39 - DCA, Section 3.06 The Borrower shall CF were registered as institute by July 1, 1981, they were established and thereafter maintain a but none had been registration scheme for allocated to settlers. cardamon growers in the The issue of available Project Area, and ensure processing channels was that cardamom estates owned not yet applicable. by the Borrower in the Project Area provide a processing and marketing channel for all such cardamom growers. DCA, Section 3.07 "In order to recover a Because of delayed part of the costs of the implementation, Project, in accordance as of April with the Borrower's 1983, no final cost resource mobilization recovery mechanism had program and except as been established by GOSL. the Borrower and the Association shall otherwise agree, the Borrower shall complete, by July 1, 1981, all measures adequate to enable the Borrower to assess, impose, and collect charges satisfactory to the Association with due regard to the settler's ability to pay such charges." DCA, Section 3.08 "The Borrower shall, by As discussed in this July 1, 1981, complete the report, this was not allocation of land to complied with. Even as of beneficiaries under the April 1983, only 890 Project on the basis of the farmsteads had been criteria set forth in allocated. Schedule 3 to this Agreement." DCA, Section 3.09 "The Borrower shall ensure The project did obtain that FD, PTU and SCD are technical and provided with sufficient implementation support and qualified staff to from other units. assist NADSA on a full time basis in carrying out the Project." - 40 - DCA, Section 3.10 The Borrower shall ensure Not verifiable at this that measures acceptable stage. to the Association shall be taken to safeguard water supply systems provided under the Project against waterborne diseases. DCA, Section 5.01(b) "The following events are The lands were transferred specified as additional but NADSA was forced to conditions to the take over responsibility effectiveness of the for factory operations and Development Credit estate labor and staff, Agreement within the contrary to mutual meaning of Section understanding. 1/ 12.01 (b) of the General Conditions: eb) the transfer of exclusive control over the lands included in the Project to NADSA. PA, Section 2.06 "In order to assist NADSA NADSA did employ in carrying out the consultants which were Project and especially in satisfactory. variety selection and the processing and marketing of the most important spices grown in the Project Area, NADSA shall employ consultants whose qualifications, experience, and terms and conditions of employment shall be satisfactory to the Association." 1/ This was specified in agreed minutes of negotiations. - 41 - PA, Section 2.07 "NADSA shall collect data This was never relevant in relating to the income of view of the delayed settlers in the Project settlement program. Area at least twice a year in order to determine the income level of such settlers. In addition to the above, NADSA had not submitted its audited accounts as was required by Section 4.02 of the Project Agreement. IV. RATES OF RETURN 4.01 Available data do not permit a meaningful evaluation of benefits and project rate of return. The project did not materialize as was envisaged at appraisal. Although project implementation began in line with the appraisal design, substantial changes took place which drastically %,a.nged the benefit stream and farm budget models. The size and crop composition of High and Low Altitude Mixed Farms were changed, affecting both the financial viability of a farm unit, as well as the aggregate benefit stream. In addition, erosion and changes in the project area delayed and reduced the project benefit stream profile. Thus, no attempt was made to recalculate the rate of return of the project as not only the lack of data but also the substantial changes which occurred in the nature and scope of the original project would have made such an exercise rather difficult if not meaningless. V. INSTITUTIONAL PERFORMANCE AND SOCIO-POLITICAL DEVELOPMENTS A. Government of Sri Lanka 5.01 GGSL established NADSA in early 1978 under the State Agricultural Corporation Act of 1972 to implement this and any such future projects. It became an agency under the then Ministry of Agriculture and Lands (MAL) and later, after some changes in national ministerial organization, under the Ministry of Agricultural Development and Research (MODR). During project implementation, the Secretary, MAL or MADR, liaised from time to time with the NADSA Board of Directors, which was a seven-member board appointed by the Minister. Through its control of the Board, the Ministry was, in essence, in control of NADSA. In the first year and a half of the project, NADSA and its operations were heavily publicized with political overtones, and the Ministry took a strong direct interest in NADSA. It was not uncommon to use NADSA in local political rallies as an example of what could be done and of what the government had actually achieved. 5.02 However, once the first phase of physical implementation was complete and the settler selection process about to begin, socio-political problems, common in land reform activities, began to surface. The problems of fairness - 42 - in settler selection and displacement of traditional estate workers became much larger than ever anticipated. It became a significant political liability, with much pressure being put by local members of parliament on NADSA to favor their constituencies. In trying to accommodate these pres- sures, the settler selection process began to violate the stipulations set out in the DCA. With this growing local political involvement and the clear potential for political embarassment, the Ministry began to take less of a direct interest in NADSA activities. Monitoring of the project at ministerial level faded, and NADSA was left to continue to run its own business with a much reduced profile. 5.03 Political pressure eventually led to a ministerial enquiry into NADSA's activities and resulted ultimately in transfer of project lands to the LRC (3.13). In June 1982, GOSL, following recommendations from IDA, made a request for ending IDA's association with the project and for accelerated repayment of the credit. 5.04 GOSL performance in terms of consistency in support and compliance with DCA covenants was weak. Although GOSL gave many indications and assurances of support for the project, there are several aspects which sug- gest that this support was not unconditional nor based on a full awareness of the local political sentiments. The shortcomings of GOSL in this respect are summarized below: (a) GOSL failed to carry out its commitments in several DCA covenants which were crucial to project success (para 3.26); (b) No sincere cooperation was obtained from JEDB and SPC in terms of absorbing excess labor and properly transferring land to NADSA while continuing to run tea factories as agreed (para 3.07); (c) GOSL did not carry out the settler selection process as agreed (para 3.10); (d) By insisting that NADSA obtain IDA reimbursements directly and by allocating the GOSL cost component only as a budgetary allocation, NADSA experienced increased shortage of funds (para 3.25). 5.05 Generally, GOSL did not put adequate pressure on its agencies (JEDB and SPC) to get certain DCA covenants complied with. It is likely that, had the transfer of labor problem and management of transferred estate factories, staff, and labor been dealt with properly and as prescribed in the DCA and minutes of negotiations, the project could have succeeded. - 43 - B. NADSA Management 5.06 The impact of the different Chairmen and General Managers (GM) who were responsible for NADSA's activities cannot be readily determined. However, there is general agreement among IDA project staff that the depar- ture of the first GM was a great loss to the project. He had worked on the preparation of the project and was most sensitive to the sociological aspects of the project. It may even have been the case that with his continued presence, the social issues would have been recognized much earlier and dealt with with greater sensitivity. However, conflicts between him and the Minis- ter led to his departure from NADSA within the first year of project implementation. 5.07 The NADSA Board of Directors, responsible for making NADSA policy decisions, was poorly composed. It did not include representatives of all the agencies which were either directly or indirectly involved in the project. This resulted in poor coordination of project related activities and no forum in which Lo exchange ideas with representatives of all agencies and factions affected by the project. 5.08 Following IDA's separation from the project, 0OSL reorganized and strengthened the NADSA Board. It did so by bringing the Board directly under the control of the Ministry by making the Secretary, MADR, Chairman of the Board and including representatives from SPC, JEDB, the Treasury, Department of Agriculture, Minor Export Crops Department, and the two labor unions. By including representatives from departments and organizations central to many of the problems encountered in the past, these problems have become easier to resolve. For example, the Treasury has agreed to provide funds to NADSA to repay the SLRS 10.2 million EPF funds used (para 5.13). Also, a specific program has been worked out to transfer excess labor to SPC and JEDB estates, and incorporating labor unions into the process has appeased the confronta- tions of the past and has made it easier to resolve conflicts of interest. 5.09 The above experience suggests strongly that projects in which dif- ferent social, economic, and institutional groups are affected or are directly or indirectly involved need a coordinating body which offers a forum for continuous liaison between these groups. It is very likely that had the Board originally been composed of these members, many of the serious problems encountered could have been avoided. Also, it is apparent that a project in which important social issues are involved requires managerial support that goes beyond the conventional orientation of management to include sociologi- cal awareness and sensitivity. 5.10 Despite the already-mentioned unanticipated responsibilities imposed on it, NADSA did perform well. It was swift in getting the project underway in respect of fieldwork and procurement, in the face of the many constraints. It is indeed difficult to appreciate NADSA's overall performance in the shadow of the overwhelming forces working against it. These include the forced take over of estate factory operations as well as estate staff and labor, and also the local political interferences into what was supposed to be straight forward NADSA operations, such as the settler selection and land - 44 - allocation activities. In view of these constraints, it must be concluded that NADSA was definitely an effective machinery in performing its respon- sibilities. That it did not succeed is more a reflection of the lack of GOSL support and commitment, than of an inadequacy in NADSA operations. C. Labor Issues 5.11 The problem of excess labor resulting from implementation of the settlement program was expected to be dealt with by repatriation and transfer of estate labor to other estates. The issues paper, dated November 30, 1977, following the appraisal of the project raised the issue whether "... IDA should support a project which would cause hardship to a segment of the project area population, albeit temporarily". The paper recommended "that the project be supported despite this social cost. Government is committed to the Tamil repatriation program and the mission was assured that everything would be done to alleviate hardship. Some job opportunities will be avail- able in the tea rehabilitation project area and priority for them would be given to those displaced by the diversification project." 5.12 The decision memorandum subsequently stated that "it was decided that, in supporting this project, IDA should seek Government assurances that all measures would be taken to provide, either through relocation to other estates, or through accelerated repatriation, for those workers not eligible for settlement under the project." And the DCA did include an assurance to this effect. However, no detailed program was worked out to phase in this transfer and the timetable agreed upon with GOSL for SPC and JEDB to absorb the excess labor had not been explicitly endorsed by the labor unions, and was not implemented by SPC and JEDB. The degree to which the laborers would resist the transfer was also not fully anticipated. The shortage of living accommodation in the up-country estates, where some estate workers were initially being transferred to, became an obstacle to further transfers. Although employment opportunities existed, the lack of accommodation made such transfers unpopular among estate workers. To make matters worse, the project had made no specific references to the need for labor leaving incen- tive payments in cost estimates, and so no such payments were budgeted for NADSA eventhough they are now recognized to be essential to obtaining any movement of excess labor. 5.13 When the excess labor was first approached with the proposition for transfer to the SPC and JEDB estates which had shortage of labor, several problems began to surface: (a) There was general resistance to move, a sociologically natural and expected reaction; (b) Poor housing conditions on many of the estates proposed for labor transfer, further reduced their willingness to move; - 45 - (c) NADSA, facing a lack of funds for its own operations, had resorted to using up the Estate Workers Provident Fund (EPF) and, consequently, did not have the funds to pay the workers what was correctly due to them. Consequently, they did not leave; (d) Even if NADSA did have the funds, there was uncertainty as to how much was due each worker and as to which agency (previous private owners, LRC, SPC, or JEDB, or NADSA) would be responsible for these payments; and (e) The unions and NADSA could not reach agreement on the proper level of labor leaving incentive payment. 5.14 The individual sense of resistance to move was consolidated at the labor unions which pronounced a policy of resisting the proposed labor trans- fer. Indeed, the union adopted and lobbied for the position that the project should not be implemented and that the area should be returned to estate run tea activities in the areas which had been targeted for dversification and settlement. A situation of confrontation emerged to the extent that sabotage, in the form of plundering of some farmsteads established by NADSA, could be observed. But NADSA, with backing from its Ministry, continued its work and slowly the reality of the transformation which was taking place penetrated the union ranks, and the union began arguing for a financial remuneration for its members to move to other estates. Although this was the pattern of events in most areas, in some areas the political influence of the union action was powerful enough to block the diversification and settlement program. The transfer of project lands in the Hewaheta electorate, already diversified by NADSA, to the LRC for return to estate tea operations, repre- sented the culmination of this power struggle. This transfer of project lands, on which considerable project expenditures had been made, was the issue which finally brought matters to a head and eventually led to IDA's decision to end its association with the project. 5.15 Thus, it can be concluded that the problems related to excess labor resulting from diversification and settlement were at the core of the project failure. Besides these problems, other social and political issues inter- fered with implementation of the project, and these are discussed below. D. Socio-Political Issues 5.16 At appraisal, it was anticipated that 4,500 families would be selected as settlers and would be employed by NADSA to implement the field workers. According to the selection procedures agreed between GOSL and IDA, displaced families of citizens or those that were to be granted citizenship before June 30, 1981 from the 61 project estates would have first priority. Once there were no more eligible families within this group, selection would continue among landless families from villages in the project catchments, and, finally, among unemployed youth from the project catchnents. Both Sinhalese and Tamils would be eligible as long as they satisfied the criteria. However, several aspects of the selection process brought about a - 46 - low participation response from Tamils and involved strong local political influence. 5.17 Sociologically, Tamils are a minority in Sri Lanka and have tradi- tionally stayed close together in cluster strongholds. Since the number of Tamils which would be eligible for settlement was relatively low, the likelihood of Tamils remaining in the traditional clusters through random land allocation would be small. This appears to have contributed sig- nificantly to the apparent lack of interest of Tamils in the settlement program. So did the fact that application notices had been prepared only in the Sinhalese language. In addition, the unions were encouraging the Tamil estate laborers to resist participating in the hope of reversing the project. 5.18 Because of this low response from Tamils, the first category for eligibility was quickly exhausted and selection then moved to the second category which involved landless families from villages in the project catch- ments. At this point, several members of Parliament (MP) began blocking the settlement program because of their concern that the displacement of families and random land allocation was affecting the membership of their con- stituency. In addition, in many instances, their constituency was putting pressure to ensure a return of the land to its previous owner before nationalization or to oppose the settlement program in the hope of reversing the project. In other instances, MPs would insist on having a say in the settler selection and land allocation process to ensure allocation of settle- ments to members of their constituency. This introduced a politically sensi- tive dimension into the settler selection process. In certain areas, some settlers selected by NADSA following the correct procedures, but not endorsed by local MPs, were ordered by the MPs to be evicted two days after having been handed over their homesteads. Political realism finally convinced NADSA management to take into account the MPs wishes. 5.19 Whether the project failed primarily because of improper management of labor transfer, settler selection and land allocation, or due to incon- gruencies of the project area population will remain uncertain. However, what the experience here undeniably confirms is that a project as this one, which cuts deeply into the long standing traditional social fabric of land ownership and employment, cannot succeed without having a strong grass-root political support. 5.20 GOSL has now attempted to appease local political pressures by reduc- ing the farmstead size from 2.2 ac to 1 ac, thereby allowing a greater number of families to be accommodated within the project. The unions have also now agreed to the transfer of the excess labor to the Rewaheta area under LRC management and to other SPC and JEDB estates following agreement on severence payments. What remains uncertain is the viability of such small farmsteads as a family farm unit. Most likely, it may not generate sufficient income for a family, forcing off-farm employment which may be to the detriment of the quality of the farmstead and to the economic success of the project. - 47 - VI. IDA PERFORMANCE A. Introduction 6.01 In this chapter IDA's performance is reviewed in terms of the extent to which it correctly anticipated problem aspects of the project and the extent to which the project design included safeguards to deal with these aspects. Furthermore, the impact of supervision missions in terms of iden- tifying problems, proposing remedies, and obtaining IDA action to deal with these problems is also evaluated. The main problems responsible for inhibit- ing a successful completion of the project are summarized below: (a) Socio-political problems surrounding settler selection and land allocation; (b) Problems related to the transfer of excess labor out of the project area; (c) Inadequate project funding; (d) Unanticipated project expenditures; (e) Reduced intensity of IDA supervision as a result of excellent progress in physical implementation. These problems are discussed in greater detail below. B. Socio-Political Problems 6.02 A two staged process for selecting settlers from among the project area residents had been agreed upon between GOSL and IDA. Sri Lankan citizens, whether Sinhalese or Tamil, would be given equal consideration. Highest priority would be given to estate workers displaced by the project, followed by landless families from adjacent villages, and finally unemployed youth from surrounding catchment areas within the project area. Although assurances were sought and obtained to follow this process, the project did not set out an institutional framework to deal at grass root level with the process of settler selection. The SAR made only casual reference to the need to inform the labor force on the project area estates of the potential effects of the project on their future, but no institutional mechanism was developed or proposed which would take care of specifically this aspect. As a consequence, there was no channel for feedback from project area residents and no mechanism through which to address these concerns. 6.03 Clearly, ex-ante, there was little experience with these aspects and it was hoped that the repatriation program and labor transfer plans together with the settler selection criteria proposed would succeed, as assured by GOSL throughout. But exactly because of this lack of experience in this area, the need was even greater to have within the institutional framework a - 48- sociological support unit to provide feedback of grass root sentiments. It is in this respect that project appraisal would likely have benefitted from more technical support in sociology. 6.04 Early in project implementation a supervision mission (Septem- ber/October 1978) identified a tendency for local political interference in the settler selection process and sought ministerial assurances of protection to NADSA from these tendencies. No further reference of political influence into the selection process and land allocation were referred to in supervi- sion reports until much later. 6.05 The original intent was to first finalize settler selection and then have the selected settlers form the labor force to implement the physical works. In this way, the labor force would have a sense of affinity with its work. But because the selection process was heavily delayed, this did not happen. As the physical presence of the project became more and more visible, vocal opposition to it increased. 6.06 In retrospect, it seems that IDA should have suspended disbursements until the conditions outlined in Schedule 3 of the DCA, requiring settler selection to be completed by the end of the first year (i.e., July 1979), had been satisfied. IDA had the remedy, in this Schedule, to deal with this problem early on. At that point in time, however, it was thought that GOSL would indeed deal with this issue as it repeatedly promised to do. It was left to GOSL to resolve this an an internal matter; it was in fact never resolved. 6.07 No specific scheme had been worked out at project appraisal to allo- cate land among settlers. Thus, when the project began, NADSA proposed allocating land on a random basis to the selected settlers which appeared equitable to IDA and this method was subsequently adopted. It is only after some time that the socio-political implications of this scheme became evi- dent. The implied scattering of the minority Tamils away from their tradi- tional cluster settings and among the Sinhalese was threatening and unaccep- table to most Tamils. This traditional preference was further entrenched by the then recent outbreaks of violence between the two groups. Because of that, the Tamil response to participate in the program was low and practi- cally no Tamils applied for consideration under the program. What seemed like a measure to ensure equity among elligibles had resulted, whether inten- tional or not, in a clear bias against Tamils. To further worsen the matter, announcements and application instructions were only published in Sinhalese. 5.08 These matters were not forseen at appraisal and no reference is made to a proposed framework or mechanism for allocating the land among selected settlers. The discussion on land allocation was limited to arrangements for ownership title, temporary leasing, and cost recovery. It did not discuss or suggest the process by which the actual allocation would take place. Again, it seems that a grass-roots involvement of participants or their legitimate representatives to air their reactions to proposed methods for allocating the land could have high-lighted these issues early on in project implementation, and offered a forum in which to discuss them. This would have provided a - 49 - mechanism by which to obtain project participant sentiments and allow pos- sibly for incorporating then into a more appropriate scheme. Furthermore, greater support at appraisal from a sociology expert would probably have helped to identify some of the above mentioned social impediments to project success. C. Labor Transfer 6.09 The SAR correctly estimated that a substantial number of laborers would either not be eligible or prefer not to participate in the program. It suggested that these surplus labor families would be relocated to other estates belonging to JEDB or SPC, or repatriated to India. An agreement was sought and obtained that GOSL would take all measures possible to achieve this. The agreement spelled out a time-phased program to carry this out, but it did not spell out or refer to a specific program with details about vacan- cies in different estates to which the labor would be transferred. Neither did it insist for credit effectiveness or simply through a dated covenant that specific arrangements and a plan be worked out with JEDB and SPC, as well as with the workers, for implementing such a transfer. In restrospect, it is also apparent that the degree to which laborers would be willing to move to other estates offering employment opportunities was overestimated and that the extent and impact of inadequate housing facilities on those estates in diminishing the willingness of laborers to move was not anticipated. D. Project Expenditures 6.10 The project suffered heavily from inadequate funding due to unan- ticipated expenditures. The major contributing factors were: (a) GOSL introduced changes in housing design, from a temporary structure to a semi-permanent one; (b) Unanticipated heavy build-up of XADSA staff and forced take over of estate labor payroll; (c) Unanticipated responsibility imposed on NADSA for running of tea factory operations in project area. (d) Unconventional GOSL funding arrangements for NADSA resulting in high overdraft interest charges; (e) Higher than anticipated costs for planting material; (f) Larger than anticipated increase in general price level, particularly for civil works; and 6.11 Already in August of 1979 a supervision mission projected cost over- runs over SAR estimates of US$2.6 million. Of this increase, it attributed about US$1.1 million to housing, US$1.1 million to salaries, and US$0.4 - 50 - million to planting material. As Table 3.2 indicates, even with the trun- cated project as proposed by GOSL, cost overruns are projected to be about 72% of original US dollar cost estimates. Although GOSL agreed to and did provide additional funding to cover the increased cost of housing resulting from changes in its proposed design, no such additional funding was made available to cover the much larger staff costs, forced responsibility for tea factory operations and estate labor payroll, and other expenditures not part of the project but forced upon NADSA. 6.12 The increase in estate labor and staff costs resulted primarily from NADSA being forced to run the tea factory operations in the project area combined with delays in transferring excess estate labor out of the project area to other SPC and JEDB estates, as agreed. NADSA, forced to take over these operations if it wanted to receive the agreed upon project lands, did. so in the hope of pushing ahead with the project. At appraisal, it was agreed between GOSL and IDA that NADSA should not be burdened with management of the tea factories on the estates taken over for diversification. SPC and JEDB were to continue to run these factories with their own staff and in fact, agreement had been reached between IDA and GOSL that proper transport and processing of green leaf from lands transferred to NADSA would be arranged between NADSA and GOSL. Thus, responsibility for the tea operation activities, having been pushed onto NADSA, was counter to understandings reached with GOSL and became a serious burden to NADSA, both financially and administratively. Although IDA supervision missions immediately identified the problems brought about by this imposed situation, no firm IDA action was taken to reverse it. Consequently, it continued to be a serious financial and managerial drain on the project. 6.13 To further exacerbate the funding problem, actual domestic inflation was considerably higher than appraisal estimates, so that allocations made for the project within the five year national plan budget, in real terms, were below the needs of the project. Also, GOSL had adopted a financing arrangement with NADSA whereby budgetary appropriations for the project would be made only in the amount of the GOSL contributions to the project. For the IDA reimbursable cost component, NADSA would have to rely directly on withdrawals from credit funds. This arrangement was contrary to that generally applied to IDA projects in Sri Lanka, where IDA reimbursements go directly to the Treasury and budgetary allocations contain grants to the implementing agency, covering the full costs. The system imposed on NADSA had two immediately identifiable impacts. Because NADSA funding became directly dependent on IDA reimbursements, the incentive to timely submit withdrawal applications on the part of the implementing agency was great. This had a positive impact on disbursement performance. However, it also implied a serious financial burden on NADSA. It meant that NADSA had to obtain commercial bank loans to cover these expenditures until they were reimbursed without obtaining GOSL funding for the resulting interest charges. These interest charges were substantial (approximately 20% a.p.r.). 6.14 This financial pressure had a secondary detrimental impact on the project. Finding itself with insufficient funds, NADSA resorted to using the Estate Workers Provident Fund (EPF) to finance its activities. This was not - 51 - only illegal, but it also meant that when proposing to estate workers, not selected as settlers, to move to other estates where employment opportunities existed, NADSA was not able to pay them the EPF which was appropriately due to them. And so, even those who were willing to move in principle would not leave. The financial constraint had been transformed into a further impedi- ment to solving the social aspects of the project. 6.15 Supervision missions, early on, repeatedly stressed the dangers of inadequate project funding and funding arrangements on the project. But again, in the hope that GOSL would deal with this issue, based on verbal ministerial assurances, no specific dates and conditions were set by IDA to resolve this issue. And GOSL never provided any additional funding to cover these unwarranted financial burdens imposed on NADSA. E. Coordination 6.16 The project envisaged that RADSA would obtain land covering 61 estates from JEDB and SPC and that JEDB and SPC would continue to run the tea factory operations on those estate lands. The project also relied on JEDB and SPC to absorb the excess labor resulting from the diversification and settlement program into other of their estates where a shortage of labor existed. 6.17 However, no coordination mechanism to liaise with representatives from JEDB and SPC was included in the appraised organizational framework. Such a coordination unit could have improved liaison and coordination between these agencies and should have also included reprasentatives from the relevant labor unions representing estate workers, as well as representatives from the Treasury. This in fact is what GOSL has done subsequent to IDA's disassociation with the project (para 5.03). It has set up a new NADSA Board composed of these members, including representatives from the Department of Agriculture, and the Minor Export Crops Department, and chaired by the Secretary, MADR, himself. This is in contrast to the original set-..p in which the Chairman, even though appointed by, was not directly linked to the MADR, and in which the composition did not include representatives from all these agencies which were either directly or indirectly involved in the project. Because of that, NADSA tended to function isolated from the Minis- try, little supervised by it, and poorly coordinated with other agencies involved in the project. With the new Board now in place, this has changed considerably and it appears that most of the original inter-agency problems are being resolved with much greater ease. F. IDA Supervision Performance 6.18 In the first year following Board presentation and approval of the project, three supervision missions visited the project area. Another three missions took place during the following year. Most of the problems dis- cussed in this completion report were identified and repeatedly focussed on by these missions. However, as also noted in this report, at no time did IDA - 52 - threaten or even suggest suspending disbursements unless those issues were resolved. Part of the problem was that the highly successful progress with the physical implementation of the project in terms of land demarcation, diversification activities, farmstead and homestead development, civil works, procurement, and disbursements diverted attention away from the social and financial problems which were beginning to threaten the project. In fact, the project was repeatedly praised locally as well as within IDA for its success, with the physical performance serving as evidence of its achieve- ment. 6.19 Each supervision report stressed the importance of completing the settlement phase of the project which continued to be seriously delayed. However, the excellent physical and disbursements progress together with government assurances that the settler selection process would be undertaken and completed shortly, had created a false sense of success. Because of that, the frequency of supervision missions was reduced. In fact, following the July 1980 mission, it was not until October of 1981 that the next mission visited the project. 6.20 That mission found that disbursements had gone up from US$1.4 million in July of 1980 to US$3.9 million, and it also identified severe problems facing NADSA on all fronts. The settler selection process had not progressed and cost overruns were beginning to cause major problems. Although it is conceivable that more frequent IDA supervisions over this period might have kept the focus on these issues, the lack of action in the past suggest that no greater commitment would have resulted on the part of GOSL, unless, of course, pressure had been introduced in the form of suspension of disburse- ments. That would have been effective only at an earlier stage when disbur- sements were still low. With almost 90% of the credit disbursed by October 1981, suspension had little significance then. 6.21 The above suggests that from early on in project implementation, supervision missions did correctly identify what were to become the major causes of project failure. What in retrospect is clear is that IDA should have taken strong action to seek a resolution of these problems. VII. CONCLUSIONS 7.01 Although the project did not achieve its original targets and cost overruns had brought the economic rate of return down to an unviable level, the concept of the project may still be sound. IDA's disassociation from the project stemmed primarily from delays in finalizing settler selection, which, in turn, were causing unacceptable cost overruns and from the transfer of parts of the project area on which investments had taken place to an agency outside the project. However, subsequent to IDA's withdrawal, GOSL made an in depth reevaluation of the project and is now continuing the project on a revised basis. Even with the project area reduced by the transfer to LRC of the Hewaheta area, the revised GOSL program now calls for a total of 5,430 families to be settled, compared with the 4,500 the project originally aimed - 53 - to achieve. By reducing the average size of farmsteads, GOSL is attempting to accommodate more settlers thereby reducing the social problems. It is now attempting to complete the settlement program by the end of 1983. 7.02 The fact that GOSL continued the project subsequent to IDA's withdrawal, is not in and of itself an endorsement of the project concept. It should be emphasized that the project was identified and appraised while Sri Lanka was governed by the Sri Lanka Freedom Party (SLFP), which, at least at that time, was sympathetic towards redistribution of public estate lands. The present United National Party (UNP) came to power in 1977 and has been more oriented towards strengthening of the public estates corporations. Although the present government has ongoing settlement as well as tea diver- sification programs, its tendency to diversify public estate lands and activate settlement on these lands is much below that of the previous govern- ment. Nevertheless, it is attemptig to bring the ongoing project to a politically acceptable conclusion. To this end, it is reducing the farmstead size to accommodate a larger number of settlers. Although GOSL may recognize that the reduced farmstead size may not be an economically viable unit for a family, it is a practical compromise in the face of the confronting social issues. 7.03 It is likely that were it possible to have forseen some of the problems that came to confront the project, and had the project had the design characteristics and institutional support systems to absorb and deal with these problems, the project may have succeeded. Thus, it is felt that in no way should IDA's withdrawal from participation in this project repre- sent a rejection of the concept of diversification and settlement in the Sri Lankan mid-country areas. However, this project has provided useful experience for any future such projects. This experience would be valuable for similar projects both in Sri Lanka and elsewhere. Some of the main conclusions from this experience are summarized below: (a) Whenever land allocation (land reform) is involved, the project must have strong political support at government level, but most impor- tantly at grass root level; (b) In order for the above to be possible, the local project area popula- tion must have a good understanding of the project and the way in which it is likely to affect them; (c) To this end, the project must include a sociological support unit, similar to an extension unit, whose role is to communicate with and inform the project area population. There should also be a forum for representatives of the people to communicate and coordinate with project management; (d) Had the appraisal mission included a sociologist, it is likely that the socio-political issues would have been more fully identified and formal institutional mechanisms included in the project design to address them; - 54 - (e) If the project relies on other agencies for its implementation (e.g., JEDB and SPC to absorb excess labor) these agencies should be part of a coordination unit; (f) If this project were repeated, it would be crucial to ensure that the settler selection process had been finalized before beginning Yith the physical project works. By having the selected settlers comprise the labor force, a strong sense of affinity with the work would exist. Also, settlement could immediately follow completion of field works; (g) IDA must maintain a constant frequency of supervision, regardless of the changing appearance of success from time to time. Often, the awareness of an imminent supervision mission serves as an inducement to getting certain things done; (h) IDA should include in the DCA only covenants which are important to the project's success and then enforce them immediately when they are violated. In retrospect, this particular project may have succeeded had IDA acted strongly early on with respect to the settler selection issue; and (i) When much uncertainty surrounds the project, its design should be more flexible to allow for changes in view of changing and unforseen circumstances. 7.04 In retrospect, it seems that IDA's withdrawal from participation in the project may in fact have added awareness and vigor to the project. For a long time, the project was lying in a stalemate state with no decisions being taken and little progress on both the physical and social fronts. Immediately following IDA's withdrawal, a ministerial review of the status of works was ordered, and proposals emerged for completing the program in a revised fashion as discussed earlier. Thus, although IDA's separation from the project and GOSL's accelerated repayment was a costly proposition to the Sri La-nkan Treasury, it may at the same time still prove to have served as the stimulus needed to bring the project to a second-best completion. Va&mmtua8T 25' 80ý30 SRI LANKA TREE CROP DIVERSIFICATION PROJECT (TEA)' Trin-, s SRI - Project Boundaries Settlement Areas LANKA Fores[ Areas ndi --- Roads Ocean lE] N --+-- Roilways AREA OF MAP e COLOMBO I ~Rivers 7°' ---- Provincial Boundary 0 1 3 4 5 in n -Oc n r------L-------- MILES 80. 20- Oorakanda- Oya/ Alapolo -710 Aruru MAHA OYA D.b. KUD YA - Var eri o I GURU GODA ,,OYOr 2, und ° CATCHMENT • GAHA A CATCHMENT AMMeMeodDa d, M'Irkviandond nnkande Bulathkca25tiv -lann 8 ,20os l2!BO 0 IBRD 13383 Bo-'a' 8o-4ta. FEBR UARY 1978 (TEA) I KANDY Poradeniya Gaalh "li ReGampola °0 Erin ntern Hill \/onør•H[ Dartry mount Galpaya iold Havane obage Oy0 5 Nu LAMB TBA t THME T \Rajcela Tomen gla Kaoiwafta 35' 0*0