Ghana - Oil Palm Project

Document of The World Bank FOR OFFICIAL USE ONLY Repot No. 5105 PROJECT PERFORMANCE AUDIT REPORT GHANA: OIL PALM DEVELOPMENT PROJECT (CREDIT 531-GH) May 31, 1994 Operations Evalulation Departmrent This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents bmy wot otherwise be disclosed without World Bank authoriaztion.  Weights and Measures 1 hectare (ha) = 2.47 acres 1 kilogram (kg) = 2.2 lbs. 1 metric ton (t) = 2,204.6 lbs. Abbreviations ADB Agricultural Development Bank EDF European Development Fund FAO Food and Agricultural Organization of the United Nations FFB Fresh Fruit Bunches GOPD Ghana Oil Palm Development Corporation IDA International Development Association ICB International Competitive Bidding IRHO Institut de Recherches pour les Huiles et Oleagineux LCB Local Competitive Bidding NIFOR Nigerian Institute for Oil Palm Research OPRC Oil Palm Research Center RMWA Regional Mission in West Africa SFC State Farms Corporation  PROJECT PERFORMANCE AUDIT REPORT GHANA: OIL PALM DEVELOPMENT PROJECT (CREDIT 531-GH) TABLE OF CONTENTS Page No. Preface ...................................... Basic Data Sheet ........................................... i Highlights .iii PROJECT PERFORMANCE AUDIT MEMORANDUM I. SUMMARY .1 II. MAIN FINDINGS. 4 A. Project Success and Durability . B. The Land Compensation Issue. 6 C. Labor ...................................... 7 D. The Outgrower Approach. 8 PROJECT COMPLETION REPORT I. Background ........................ 11 II. Project Formulation ...................................... 12 III. Implementation . ..................................... 15 IV. Evaluation of Project Impact ............ ................. 28 V. Benefits and Economic and Financial Evaluation .... ....... 29 VI. Institutional Performance ................................ 30 VII. Conclusions . ............................................. 33 Annexe Tables Planting Program . .............................................. 37 Estimated and Actual Costs and Financing ...................... 38 Status of Funds as at February 1983 ........................... 39 Schedule of Disbursements ........ ............................. 40 Project Costs . 41 Imports of Oils and Fats for Industrial Use ..... .............. 42 Inflation Rates and Factors ........... ........................ 43 Breakdown of Field Establishment Costs ...... .................. 44 Farm Budget ............................................... 45 Economic Millgate Price of Palm Oil .-------------------------- 46 Economic Millgate Price of Palm Kernel ...... .................. 47 This document has a restricted distribution and may be used by recipients only in the performance of their official dutics. Its contents may not otherwise bc disclosed without World Bank authorization.  Table of Contents (coat.) Page No. Annexe Tables Yield Projections for Nucleus Plantation and Outgrowers ....... 48 Details of Economic Operating Costs ........................... 49 Details of Project Costs ...................................... 50 Economic Rate of Return ....................................... 51 Assumptions Underlying Economic Rate of Return Calculations ... 52 Map - 10271 R (PPA)  PROJECT PERFORMANCE AUDIT REPORT GHANA OIL PALM DEVELOPMENT PROJECT (CREDIT 531-GE) PREFACE This is a performance audit of the Oil Palm Development Project in Ghana, for which Credit 531-GH was approved on February 18, 1975, in the amount of US$13.6 million. The Credit is expected to be closed fully dis- bursed on June 30, 1984. The audit report consists of an audit memorandum (PPAM) prepared by the Operations Evaluation Department (OED) and the Project Completion Report (PCR) dated Jume 1983, prepared by the Bank Resident Mission in West Africa following a country visit in September 1982. The audit memorandum is based on a review of the Appraisal Report (No. 173a-GH) dated February 3, 1975, the President's Report (No. P-1431-GH) of February 6, 1975, the Credit Agreement dated March 5, 1975 and the PCR. Internal Bank memoranda on project issues as contained in relevant Bank files have also been consulted and Bank staff associated with the project have been interviewed. A copy of the draft PPAR was sent to the Borrower for comments on January 27, 1984. However, none were received. The audit finds that the PCR covers adequately the project's sali- ent features and the PPAM agrees with its conclusions. In addition to summarizing the objectives and results of the project, the PPA! expands upon certain aspects of land acquisition and land compensation, the availability of labor and the outgrower development approach because of their importance to this and other similar projects undertaken in the same region. The valuable assistance provided by the Government of Ghana, the project staff and the officials involved in the project during the prepara- tion of the PCR is gratefully acknowledged.  - ii - PMUnF.CT FF:MIIt"ANFC. AUtIT Rpnlutl GUANA: OIL rALM DEVR.IJvPEmr ISaAC? _ _..J! I 533___) ______ BASIC DATA SIUFXT KET ruOJEC? DATA Appraisal Actual or Actual as Z of rEtimate Estimat_d Actual Apprasal Lalate Total Project Cost (US$ million) 22.5 47.8 214 Credit Amount CUSS million) 13.6 13.6 100 Date Board Approval - 0211/75 Date Effectivene 06J13/75 06124/76 40 Date Physical Components Completed 12/31/82 03131/84 132 Proportlon then completed (2) 100 94-112-951a Closins Date 12/31/83 06/30/84 124 Economic Rate of Return (X) 16.6 12.0 Financial Rate of Return (2) Inmtitutional Performace 80 Agrunomic Performance - - 90 Number of Direct Ueneficiarls (1984) Smallholders/outgrouera 250 318 12? Staff and Laborers 850 900 106 CUHUATIVE DISSURSEUEUOTS 7177 m78 Fn79 rTS0 "el 782 7T83 7164 Appraiaal Estimate CUSS milllon) 1.3 2.3 3.6 5.9 9.3 11.3 13.6 13.6 Actual (US5 mlllon) 0.97 1.77 2.77 4.97 8.47 11.77 13.1 13.6 Actual aa * of eutisate 74 77 77 84 91 104 96 100 Date of final dlsbursement: *me 1984 MISSI0N DATA Date No. of Specializations Performance Types of Mission (No.lYr.) Persons Mandays Represented k Ratinx /d Trend le Problem /f Identification 03171 NIl N/A Preparation 06/72 MIA N/A Appraisal 10/72 5 75 a.bc - - Subtotal 75 SupervisIon 1 12/75 1 2 c not effective N/A N/A Supervision 2 04176 1 5 d oot effective N//A R/A Supervislon 3 09/76 2 12 b.c 2 1 n Supervision 4 03/77/b 2 2 a*d - - Supervision 5 06/777ii 1 6 b Supervlison 6 10/77 1 5 a I 1 Supervision 7 05/78 1 3 a 1 2 11 Supervision 1 11/?8 2 8 a,d 1 2 T Supervision 9 11/79 1 6 c 2 3 T.r Supervimion 10 05/80 2 17 ce 2 3 r,0 Supervision 11 10180 3 15 a,;,f 2 2 n 0 Supervision 12 05/81 4 18 a*.c.dd 2 2 O.P Supervision 13 12/81 1 7 a 2 2 0,Y Supervision 14 09/82 1 9 a 2 3 r.n.r Subtotal 115 Total 190 OTRZ PROJECr DATA Borrover RepublIc of Chana Executing Agency Chana 011 Paln Development Corporatlon Fiscal Year July I - June 30 Name of Currency (abbretlatlon) Cedi (9) Currency Excharge Rate: Apprai-al Year Average ISS1.00 - 1.15 lntervening Tears Averape 5DI_00 - 9 2.21 Completion Year Average US$1.00 - 1i 2.75 Y11ow-on ProjeCL:/1 Name: (Pioponed) Second 011 Palm Development Project Credit Number: NMA Credit Amount: US$15.0 million Date Board Approval: Planned June 1984 /a 942 of the nucleus estate. 117Z nf amallholders;autgrowere anod 952 of the nil m111 comronent. 7b A Reck-to-Office Report wan prepared Instead of a Full Supervieinn Rrpnrt. 7r a- agriculturalist: b - agricultural economlit; c - financial anmlyat; d - ecnnostmt; C - con.lts.n; f - vp. Id I - problem-free or minor prol.lemn; 2 m derate pruhlea-; and 1 - mjor problems. 7e 1 - I Ipr-.ving; 2 - slam ine-.-r;; aed 3 - detrinratting. If F- finmmrll: N - mnnnimpl.]; T * trChmmlf'l P - polItical; and 0 - otier. 7k ImA rrntedr Cmv-r.msrnt n 11. udvancem of S';857.00t) n May 9813.  - ili - PROJECT PERFORMANCE AUDIT REPORT GHANA OIL PALM DEVELOPMENT PROJECT (CREDIT 531-GH) HIGHLIGHTS The project was the first phase of an oil palm development program and consisted mostly of establishing and operating a 4,000 ha nucleus planta- tion, assisting outgrowers for planting 1,200 ha, constructing a palm oil mill and training personnel. The project objectives were to generate an annual production of 14,000 tons of palm oil for internal consumption and 3,000 tons of palm kernels for export, and to provide employment to about 250 farming families and 600 staff and laborers on the oil palm nucleus planta- tion and in the oil mill. The project, to be implemented over an eight-year period, was to be executed by the Ghana Oil Palm Development Corporation (GOPD) an autonomous agency which was to operate as a commercial entity with the assistance of a firm of consultants experienced in oil palm development. Project start-up and implementation experienced delay due to the political and economic problems of the country, difficulties in acquiring land for oil palm planting in the nucleus area, the lack of labor and conr tinuous shortages of building materials. As a result of difficulties encountered during project implementation, the planting program was delayed by two years but attained 5,145 ha, or 99X of the appraisal target. Construction of the oil mill was successfully completed on schedule. Actual project costs totalled US$47.8 million, instead of US$22.5 million estimated at appraisal, an increase of 114%, due mostly to high inflation, the high cost of land compensation and an increase in oil mill size and capacity. The project ERR is now estimated at 12%, compared with 16.6% esti- mated at appraisal. The project cost increase has been partly compensated by increases of world market price projections for oil palm products. The proj- ect benefited about 600 permanent staff and laborers of GOPD and about 318 farming families, 27% more than anticipated at appraisal. Considering that the project was implemented in a difficult working environment and at a time when Ghana faced considerable political and economic problems, the project was a success despite its cost and time over- runs. The autonomy of the execuring agency, the efficiency of consultants and Government's support were the main reasons for the project success. However, the project's durability remains uncertain because the training objectives, aimed at having qualified Ghanaians replace expatriate staff, have not been met. The following points may be of particular interest: - training efforts are frequently lost to projects because of high staff turnover, poor design of the training component, and the  - Iv - difficulty for expatriate-) to Implement project vorks wlthin a tight schedule vhile at the same time having to traln counterpart (PPAM, paras. 18-20; and PCR, pars. 7.06c); when land expropriatlon is a prerequisite for successful project execution, a cadastral survey and a land compensation plan should be completed before appraisal, and Government Rxpropriation pro- cedures should be reviewed by the Bank (PPAM, paras. 21-23; and PCR, para. 7.07); in little-populated regions, the only way to attract and retain good laborers is to provide them with housing and to support ford crop production (PPAM, paras. 24-25; and PCR, paras. 3.35 and 7.04); and as in some other Bank-assisted oil palm projects, there is a risk that recovers of credit might prove difficult as farmers may evade repayments by selling part of their crop on the parallel market instead of through the official mill channel (PPAM, paras. 26-27).  PROJECT PERFORMANCE AUDIT MEMORANDUM GHANA OIL PALM DEVELOPMENT PROJECT (CREDIT 531-GH) I. SUMMARY Background 1. In the early 70s, agriculture, forestry and fisheries employed about 60Z of the Ghanaian population and contributed to about 40% of its GDP. Cocoa was the main export commodity and generated about 622 of the country's export earnings Ghana's main economic problems were dominated by the need to reduce the bAlance of payments deficit, achieve self-sufficiency in food and diversify its agricultural production, emphasizing import substi- tution projects. To achieve its objectives the Government was considering, inter alia, the development of oil palm, because of the suitable climatic conditions of the southern part of the country and the increasing internal demand for palm oil. 2. Palm oil production in 1970 was estimated at about 20,300 tons, of which 19,000 tons was used for human consumption. Internal demaLnd for palm oil was projected to increase to about 69,000 tons in 1985 and to 80,000 tons in 1990, well in excess of the estimated production of 41,000 tons and 45,000 tons respectively. The country's main sources of palm oil were the natural wild palm groves, small private plantings and some large-scale plantations. Project Design 3. The project was the first phase of an oil palm development program and consisted of: Ca) establishing and operating a 4,000-ha nucleus plantation; (b) providing technical services and credit facilities to develop 1,200 ha of outgrower plantings; (c) constructing and operating a 10-ton/hour pain oil mill; (d) financing infrastructure development and operating costs; (e) training personnel; and (f) carrying out field trials. The project was to be implemented over an eight-year period, and was to include expenditures up to the year when plantings were brought into har- vest. Planting would be completed during the first five years. Total project cost was estimated at US$22-5 million of whic l IDA was to provide  -2- US$13.6 million (60%). The project was to be executed by the Ghana Oil Palm Development Corporation (GOPD), which was to operate as a commercial entity, and had been established under the existing Statutory Corporation Act. To ensure efficient management, GOPD was to recruit key staff by entering into a contract with an internationally recognized firm experienced in oil palm estate management. 4. The nuclems plantation with outgrowers approach was expected to offer advantages in the spheres of employment, extension and large-scale central processing. The employment opportunities on the nucleus plantation would provide a regular source of income for members of the outgrower commur- nity and would serve as a valuable extension factor in the outgrower planting sector of the project. Participating outgrowers would be limited to those with land suitable for oil palms within a 15-mile radius of the nucleus estate. Credit would be supplied to successful applicants in the form of planting material, fertilizers and other elements supplied in kind, and cash for up to 80% of estimated labor costs. Loans would be recovered from the proceeds of Fresh Fruit Bunchl/ (ffb) sales. Project Implementation 5. Although the Credit Agreemeir: was signed on March 5, 1975, the project did not become effective until Jui.e 30, 1976, because a number of conditions of effectiveness remained unfulfilled, particularly the appoint- ment of a Managing Director, the contract with consultants and the acquisition of 1,200 ha of land for oil palm planting in the nucleus area. Project implementation also experienced some delay, mostly due to continuous difficulties in acquiring and transferring land to GOPD as a result of slow compensation of dispossessed farmers and dissatisfaction of the latter with the compensation rate. In addition, the Government's policy and funding were not adequate for a satisfactory solution of the land compensation issue. 6. Manual land clearing also proved difficult and lengthy. In Septem- ber 1976 it was decided that clearing would be done mechanically. Three bulldozers were ordered, and arrived in April 1977, but their performance was below standard because of poor maintenance and lack of spare parts. Clearing operations improved markedly in 1979/80 when two bulldozers of another type were purchased and an expatriate workshop manager was recruited. 7. The availability of sufficient labor remained a constant problem. The labor policy, set up at appraisal, to employ only non-resident laborers proved unrealistic as (i) most workers living in the project area were mainly seasonal and made up of smallholders with a low output; (ii) the dispossessed farmers, dissatisfied with the rate and the pace of land compensation, generally refused to work on the project; (iii) the pro_ect lacked the means to safely transport large numbers of workers from distant villages; and (iv) salaries paid by GOPD were lower than in other oil palm estates. This unfavorable situation was reversed in 1981 when the plantation 1/ Fresh Fruit Bunches (ffb) are the produce of the oil palm.  -3- workers benefitted from World Food Program rations and when a housing program for workers was finally adopted. 8. The initial outgrower program changed at the end of 1977 when it became apparent that it was seriously delayed by the lack of access roeds and the lengthy delays in obtaining the legal documents for outgrowers. At the suggestion of the Managing Director, a block of smallholders was created conr- tiguous to the nucleus estate. Each smallholder was to receive on lease from GOPD 8 ha of land to be planted under oil palm at an annual rate of 4 ha, and 1 or 2 ha for growing food crops. The introduction of the "smallholder- policy (farmers' taking up land for cultivation of oil palms), as dis- tinguished from 'outgrowers- (farmers receiving assistance to plant their own land) was reflected in an amendment of the Credit Agreement in 1980. The smallholder program made a good start, as land was allocated to nearby villagers, but because of difficulties in obtaining additional land, the out- grower concept was reintroduced later. At project completion 318 farmers (202 smallholders and 116 outgrowers) or 40Z more than expected at appraisal have benefitted from the program. 9. As a result of delays and difficulties encountered during project implementation, the planting program, for both the nucleus estate and the smallholders/outgrowers was delayed by two years. However at project comple- tion in September 1982, the overall planting program had attained 5,145 ha or 99X of the appraisal target (94% of the target for nucleus estate, and 112% of the target for the smallholder/outgrower component). 10. The construction program, consisting of offices, houses for senior staff, a training and a health center, workshop, storage facilities and a small dam, was delayed because of continuous shortages of building materials, and is expected to be completed in 1984. Construction of the oil mill was started in June 1980 and successfully completed in November 1982, one month ahead of schedule.2/ I1. Actual project costs totalled US$47.8 million, instead of US$22.5 million estimated at appraisal, an increase of 114%. The main reasons for this cost escalation were an overvalued exchange rate, high inflation, addi- tional buildings and heavy equipment for land clearing, the high cost of land compensation, much higher local wages than anticipated and a larger oil mill. Project Impact 12. Yields at full maturity are expected to be lower than anticipated (about 10% for nucleus estate and 20% for smallholders/outgrowers) because of the likelihood of droughts, the shortage of inputs in Ghana and possible problems which could result from the departure of expatriates before a replacement team of Ghanaians is ready to take over. The project ERR is now estimated at 12%, compared with 16.6% estima-ed at appraisal. The project 2/ In 1977 the decision had been made to increase the oil mill capacity from 10 tons to 15 tons per hour.  cost increase has been partly compensated by increases of world market price projections of the palm produce as compared with appraisal estimates. 13. The social impact of the project is not negligible since about 600 permanent staff and laborers and as many casual workers are employed in proj- ect works. In addition the project contributed to lncreaslng the incomes of 318 smallholder/outgrower families representing 1,500 to 2,000 people. About 900 additional farmers have applied to become smallholders or out- growers, should a follow-up project materialize. 14. The project was less successful in improving the capability of Ghanaian nationals to operate and manage an oil palm development project. Training of national staff proved insufficient, and there is no evidence that the training program met the project objective to have qualified Ghanaians replacing expatriates at the time of project completion. In addition, a num- ber of trained national staff have been suspended or have resigned following civil disturbances and political problems during project implementation. 15. A second-phase project aiming at consolidating the Phase I project and expanding both outgrower plantings and the capacity of the oil mill was appraised by the Bank in April 1981. However, submission of documents to the Board was deferred after the change of Government In December 1981 and pending agreement of the Government with the IMF and the Bank on macro- economic policy issues. In October 1983, the project was re-appralsed fol- lowing the Government's announcement of major policy reforms. Board presen- tation of the second project is now scheduled by June 1984. II. MAIN FINDINGS A. Project Success and Durability 16. Considering that the project was implemented at a time when Ghana faced enormous political and economic problems, resulting In the failure of a number of Bank-assisted projects, this oil palm development project can be termed a success despite its large cost overrun. In the audit's view, and as correctly pointed out by the PCR, the project success was mostly due to (i) the relative autonomy of the project, which was under the responsibility of a statutory corporation, not part of the government bureaucracy; (ii) the efficiency and wide experience of the consultants; and (iii) the government support despite its financial difficulties. 17. As already mentioned in a number of OED rep-rts4/ an autonomous agency, able to integrate production, transportation, processing and marketing, is generally more efficient than a ministerial dep:irtment for pro- moting crops which have to be processed. In this respect, :ne project was no 3/ See particularly PPAR of Ivory Coast: Cotton Areas Rural Development Project (OED Report No. 4568).  - 5- exception; it undoubtedly benefitted from the statutory autonomy and the undeniable flexibility of GOPD, particularly during the difficult and troubled period of the country- The need for technical assistance from a firm of consultants experienced in oil palm planting and management was recognized at appraisal and proved to be essential for the project success since little expertise in oil palm existed in Ghana before the project. The team of expatriates was efficient and dedicated amid difficult working con- ditions- The consulting firm fully supported the project and agreed to leave its management team in Ghana awaiting a follov-up project. Finally, the Government was cooperative and supported the project even in the 1981/82 period when financial problems became particularly severe and management more difficult because councils of workers were given the right to intervene in day-to-day management. 18. The question of the project's durability, however, remains un- answered. It is clear that the training objectives, aimed at having quali- fied Ghanaians replace expatriate staff, have not yet been met. One of the reasons was that some top Gnanaian staff have been suspended or resigned following the revolution. In addition, it is not evident that the training program4/ was adequately designed and the personnel who attended the courses made the best use of their training for the benefit of the project. Should a Bank-assisted second-phase project not materialize, and the technical assistance stop within a few months, the economic viability of this first- phase project could be seriously jeopardized. 19. The project illustrated the shortcomings of the training component in a number of Bank projects. The PCR correctly mentions that it is diffi- cult for expatriates to train counterparts during a tightly scheduled plantation program; besides, expatriate technical and managerial specialists do not always have the ability and skills to train their counterparts. In addition, as noted in other Bank-financed projects,5/ the benefits derived from training efforts are frequently lost to the project because staff turnover is high in many countries and trainees are often transferred elsewhere soon after their training. Another problem, which has affected many plantation projects under parastatal management, is that conditions of employment often compare unfavorably with those in the private sector and prevent attracting and retaining staff of the calibre required. 20. The experience of this, as well as other projects of the same nature proved that (i) the type and timing of training should be clearly defined at appraisal and, if necessary, designed with the assistance of a training specialist; (ii) expatriate technicians and managers should not be entirely relied upon for training purposes and a training specialist should 4/ A two-week course in Swaziland for assistani plantation managers, a one-month course on harvesting in Ivory Coast for the senior assistant plantation manager, ten-month courses in the Netherlands and Malaysia for mill engineers and a three-year course in agricultural economics in Australia for an assistant plantation manager. 5/ See particularly OED Report No. 4749 (Dominican Republic: Yaque del Norte Irrigation Project); Yemen Arab Republic: Southern Uplands Regional Development and Grain Storage Projects (under preparation).  - 6 - be included in the expatriate team when training is considered essential for the durability of the project and (iii) when costly training abroad is included in the training program, beforehand assurance would be _btained from the trainees that they would serve the project for a number of years before being transferred elsewhere.6/ B. The Land Compensation Issue 21. The project faced a continuous problem of land acquisition because the Government proved unable to quickly and adequately compensate the dispossessed farmers. This led to serious delays in the land clearing and planting programs, to a marked hostility of the population against the project and to severe labor problems. Although the cost of land compensation turned out to be 550% higher than anticipated at appraisal, there is no evidence that dispossessed farmers have been adequately compensated and/or relocated. 22. The land availability and land acquisition and compensation issues have been noted in a number of projects of the same nature (as well as in livestock and irrigation projects) and have been raised in some PPARs.7/ With the benefit of hindsight, it is clear that the land tenure and land com- pensation aspects of the project were not sufficiently scrutinized at appraisal. The Staff Appraisal Report only stated (para. 3.06) that 'there does not seem to be any problem in acquiring the land needed for the nucleus plantation and recommended to obtain assurances that -the Government take all requisite steps for the acquisition and retention by GOPD of all land required to establish a 10,000 acres plantation prior to June 30, 1976- (para. 4.06). In sum, it was thought at appraisal that since the Government was entitled to acquire land, there would be no land acquisition problem. In addition, land compensation was essentially a local cost and therefore, according to Bank rules, no Bank funding was earmarked for this purpose. In fact, it became apparent during project implementation that the procedure and methods of compensation were slow and inadequate (PCR, para. 3.05); the comr pensation rate was somewhat arbitrary and continuously questioned by dispossessed farmers; and the Government lacked the necessary funds. 23. The experience of this project confirms OED findings that when land expropriation and land compensation are a prerequisite for project execution, a cadastral survey and a land compensation plan should be completed before project appraisal. The Government expropriation procedures should be reviewed by the Bank and the compensation rate and timing should be 6/ Some Bank staff argue that it is often better to let trainees go than to attempt to hold them against their will. On the other hand, other staff note that a high turnover of competent national staff may result in an increasing dependency on expatriates. 7/ OED Reports lo. 3972 (Ghana: Livestock Development Project); No. 4979 (Dominican Republic: Yaque del Norte Irrigation Project); and No. 3971 (Ethiopia: Lower Adiabo Development Project).  -7- determined during project preparation and agreed upon at negotiations. Finally, it should be a condition of effectiveness of the Loan/Credit that the Government has acquired all land or most of it as needed for project implementation. C. Labor 24. The availability of sufficient labor was another important problem of the project during implementation. Unlike most of the cree crops projects in Africa and elsewhere, the project policy was to employ only non-resident laborers. Therefore, no villages were built in the estate and the project had to rely on costly and unsafe transportation of laborers from surrounding villages. The situation was complicated by the shortage of food on the proj- ect site, the low salaries, and the dissatisfaction of the dispossessed farmers who were the main source of labor and often refused to work on the estate. As a result the rate of absenteeism was abnormally high. A further risk was a rapid turnover of the labor force and the difficulty of training people in skilled operations like harvesting. In 1976, a reversal took place and the decision was made to establish villages at the project site. But the construction cost proved to be much higher than at appraisal time and progress in construction was slow because of shortages of building materials, mostly cement. 25. Again, the experience of this project confirmed those of other similar projects and showed that in the little populated regions, where tree crop development usually takes place, the only way to attract and retain good laborers is to provide them with housing and to support food crop production.8/ Like in other tree crop projects, the type of housing has also been a matter of controversy in this project- While a Bank consultant and the Project Manager recommended temporary wooden houses, the Government gave preference to houses made of concrete. In the Bank-assisted rubber projects in Cameroon,9/ and the oil palm and coconut projects in Ivory Coast,10/ villages were also built in concrete but the design of the houses and, more importantly, the siting and layout of the villages, proved tc be poorly adapted to the socio-economic context of the region and the population. The umodern blocks construcred under these projects soon turned to urban slums rather than rural 'Villages and raised discontent among dwellers. In view of these past experiences one may wonder whether a mare traditional type of village and traditional construction material (like that prevailing in land-settlement projects in Malaysia) would not have been better accepted by laborers and proved cheaper than the villages built under these projects. 8/ On this subject see also Cameroon: Niete Rubber Estate Project (OED Report No. 3507). 9/ OED Report No. 3507. 10/ See Ivory Coast: Impact Evaluation of Oil Palm and Coconut Projects (OED Report under preparation).  -8- D. The Outgrower Approach 26. The PCR states that the nucleus plantation with outgrowers approach proved a feasible excercise. This approach was also generally successful in other Bank-assisted tree crop projects in Southeast Asia and Africa. In Ghana, like in other countries, the Bank development strategy was based on credit in kind and cash to outgrowers. With the benefit of experience gained under other oil palm projects the credit system proved a dangerous financing mechanism for farmers who are not familiar with its requirements and evade repayment by selling part of their crop on the parallel market instead of through the official oil mill channel. In addition to a low recovery of loans from outgrowers the adverse effect of this practice is that processing plants do not operate at full capacity, which keeps processing costs high and jeopardizes the economic viability of the project. 27. Since a free market of oilpalm products probably exists in Ghana like in neighboring countries, GOPD will have to face the same issue as in other projects when repayments by outgrowers are being started, i.e., when the oil palms are nine years old. The best way to solve this problem will '. to ensure that (i) the ffb price paid to outgrowers remains competitive with the one of the free market, (ii) collection of ffbs is timely and properly organized and (iii) payments to farmers are made on a regular basis. In this respect the follow-up project appears essential to protect and bring to full development the investments made under the Phase I project and more particularly to organize an adequate harvesting, collection and payment system for outgrowers/smallholders.  GEEAN OTM PALH DEVELOPMENT PROJECT (CREDIT 531-Ga) PROJECT COMPLETION REPORT JXLu.E 19B3 The World Bank. Resident Mission in Western Africa  t I r~~~~~~~~~  - 11 - I. BACKGROUND 1.01 Ghana, vith an area of about 239,000 kmd, has a population of 11.7 million, growing at over 3 percent a year. With an average population density of 48 per sq. km there is no overall scarcity of land, although in certain regions like the Central and the Eastern, population density can exceed 100 per km2. A significant shift of rural population 1/ to the urban areas has taken place since the early 1970's when some 70 Z of the total population was deriving a livelihood from agriculture. With a declining GDP per capita of approximately USS 420, Ghana was, in 1980, among the 34 least developed countries in the world. 1.02 Some 12 Z of the country's 22.7 million ha of land is under cultivation; nearly 55 Z under bush-fallow and the remaining 33 Z under forest. Shifting cultivation with long periods of fallow is the predominant farming system. Important food crops include maize, which constitutes slightly over 50 Z of the overall cereals production of 700,000-800,000 tons, sorghum and millet, groundnuts and rice. The starchy staples are estimated at 5 million tons, 50 Z of which is cassava, 30 Z plantain and 15 x yams. Cocoa covers some 50 Z of all cultivated land; oil palm is grown on about 111,000 ha whereas cotton, coffee and sugarcane remain minor crops. 1.03 The Bank's main objectives in Ghana is agricultural sector have been to assist the country in achieving self-sufficiency in food and raw materials for agro-industries. Total sector lending over 1969/80 amounted to USS 105.5 million and comprised 8 projects. These included one fisheries, two cocoa, a sugar, a livestock, an oil palm and two agricultural development projects. 1.04 Oil palm cultivation has been important in Ghana for several decades. However, it was onlv late in the sixties that more attention was devoted to the crop. The country's main sources of palm oil were the natural wild palm groves, small private plantings and large scale plantations. Departing from previous interventions in production activities through heavy investments in large capital intensive State Farms, the Government decided in the earlv seventies to give more consideration to smallholders and large scale nucleus estates. The country's overall production reached 20,000 t of palm oil in 1970, while recent forecasts put the internal demand for the commoditv at 70.000 t in 1985 and 90.000 t in 1990. This is well in excess of the pfoiected national output of 75,000 t of Dalm oil in 1990. 1.05 The nast development of oil nalm in Ghana has mainlv been through the establishment of three large-scale Dlantations e State Farms Corporation (SFC) which has about 14,200 ha planted including 400 ha under outgrowers; Benso - Unilever - with about 4,000 ha and Twifo - CDC - which out of a targeted 5,300 ha has achieved only 1,000 ha so far due to land acquisition problems. The Oil Palm Development project for which an IDA credit of USS 13.6 million was approved in February 1975, helped the Government expand the area under oil palm through the implementation of both a 4,000 ha nucleus estate and 1,200 ha of smallholder/outgrower plots. The project, located in the southern part of Ghana, - some 160 km North-East of Accra -, also provided for the construction and operation of a 10 t per hour ffb oil mill, the building of access roads in the vicinity of the nucleus estate, the training of personnel, 1/ World Development Report 1982. Table 12. It is estimated that from 1960 to 1980, the rural labor force declined from 64 x to 53 % of the country's total vorking population.  - 1Z- the carrying out of field trials and the allocation of credit to smallholders/ outgrowers. The project was to be implemented over an eight year period at a total cost of US$ 22.5 million of which the IDA credit represented 60 Z. Outgrowers were to contribute 20 Z of farm labor requirements during establishment, equivalent to some US$ 120,000 or about I Z of total project costs, whereas the Government was to provide the remaining US$ 8.8 million equivalent or 39 Z of total project costs. The credit agreement between the Government of Ghana and IDA was signed June 24, 1976 and the credit was expected to be fully disbursed by the new closing date of June 30, 1983 - six months ahead of the original closing date. 1.06 This report is based on a review of IDA records, the findings of a supervision mission which visited Ghana in August/September 1982 and views expressed by Government officials during the above mission. Because of the heavy workload prior to handing over the project to the Government in September 30, 1982, the IRHO management team submitted only a brief and incomplete analysis of the project's past problems and achievements. At the time of the mission, about 98 % of the IDA credit had been committed, the appraisal planting targets were 94 Z complete for the nucleus estate and 112 % for the smallholders/outgrowers and the construction program for workers' housing, which was much larger than anticipated at appraisal, was in full swing. The oil mill was commissioned in October 1982, one month ahead of schedule. II. PROJECT FORMULATION A. Identification, Preparation, Appraisal and Board Recommendations. 2.01 An identification report for an oil palm development project was submitted to IDA by the Ministry of Agriculture, in November 1970. The proposed project consisted of the implementation of some 4,050 ha each in the Kade and Twifo areas of the Eastern Region. It was later agreed during discussions held between the Ministry of Agriculture and an RMWA mission in May 1971, that the project should instead (i) be limited to a single area; and, (ii) provide for both a smallholder and a nucleus estate component. At the specific request of the Ministry of Agriculture, preparation of the project was left with their planning unit. However, as a result of insufficient know-how within the Ministry and a lack of coordination between the soil research institute and the planning unit, progress was slow and in March 1972 the Government asked the Bank to take over responsibility for final project preparation. The Bank retained the services of the "Institut de Recherche pour les Huiles et Oleagineux" (IRHO), a French institute specialised in oil producing crops, which initiated the study immediately. 2.02 The IRHO preparation report was received by RMIWA on August 8, 1972 and reviewed.thereafter. Despite the short notice and scarcity of available data, it was felt that IRHO had presented a well balanced report. REWA concluded that the project was sound and ready for appraisal. Over an eight year period subdivided into a two phase oil palm development operation the main components were to include : a) planting of a 5,200 ha nucleus estate and 4,800 ha for smallholders; b) building of roads; c) construction of a 20 t ffb/hour oil mill to be implemented in 10 t stages; d) providing mediumrterm credit to smallholders e) backing the local oil palm Research; and f) recruiting expatriate technical assistance. Depending on the disbursement period and the phasing of the second project, costs for the first project were estimated at US$ 7.5 to US$ 8.5 million.  - 13 - 2.03 The appraisal mission composed of HQ staff visited Ghana in October 1972. Issues raised by Government during appraisal included : a) the need f or the internationally recognized management firm retained, to take a minority share-holding position in the Corporation, in order to secure the financial interest of a private partner - the idea was subsequently dropped as it appeared unlikely that an international firm would wish to become involved in the equity of GOPD during the limited time span of the managing contract; b) the key positions to have more than one Ghanaian deputy appointed so as to provide flexibility in relation to the ultimate replace- ment of expatriate staff - instead, trainLng provisions were included in the contract and no more than one deputy was appointed per expatriate. (Para. 3.28); c) emplovment of a Plant Breeding Consultant - Government initially thought that this was unjustified whereas IDA was convinced that the services of a consultant would have a positive impact on the project -- the above employment was later approved by Government; d) whilst recognizing the need for formulating a pricing policy for oil palm, Government was reluctant to enter into an agreement which involved consulting IDA when fixing prices - although Government later reversed this initial view, the related covenant -section 3.04 - was barely adhered to (para. 3.50). 2.04 In the issues paper dated November 24, 1972, the appraisal mission stated that it had no major differences in the general proposals put forward in the feasibilitv study although it lengthened the period of development; the mission proposed that the project would comprise the development of 5,200 ha of oil palm as Phase I of a program ultimately scheduled to expand to 10,100 ha. Essentially, Phase I would involve i) establishing and operating 4,000 ha of nucleus estate; ii) developing 1,200 ha of outgrowers plantings; iii) constructing and operating a 20 ton ffb/hour oil mill; iv) establishing a fruit collection system for outgrowers; v) training personnel; and, vi) carrying out applied research in selected spheres of study. Planting would be completed during the first 5 years of the 8 year project. In fact the feasibility study 11 had planned an 8-year planting program subdivided into two 4 year phases. Some 5,200 ha would have been established from year 2 to 5 and 4,600 ha from year 6 to 9. The implementation history of the project shows that the program proposed by the appraisal mission was more realistic. 2.05 In November 1972 project costs were estimated by the appraisal mission to be USS 8.6 million and included expenditures up to the year when plantings were brought into harvest; they were therefore extended from year 1 through year 9. It was the appraisal mission's intention to recommend that disbursement of IDA funds be limited to 7 years with 8 and 9 financed entirely from local funds. The foreign exchange component was estimated at 47 Z of total project costs. Applying the world market prices forecast for 1980 by the Bank Economic Analysis and Projections Department and shadow pricing the cedi at 1.70 to the US$ 1.0 the economic rate of return was 9 Z. 1/ Feasibility study p. 48 para. 4.2.  - 14 - 2.06 A number of issues and policy matters on which agreement needed to be reached at negotiations were reviewed during the decision meeting 1/. These included : a) interest rates for lending to farmers and credit channels, b) employment of expatriate staff, c) feeder road construction, d) commodity pricing policy, and e) exchange rate. 2.07 Project processing was delayed for various reasons, including the prevailing economic situation and the need to look for a confinancer (the Kuwait Fund was a possibility). Soils surveys took longer than expected as well as the establishment of the project entity, due to Government's indecision on whether to establish COPD as a limited liability company or as a statutorv cornoration. As time gassed. costs had to be recalculated and rose to USS 13.5 million in Anril 1973: USS 18.9 million at the time of negotiations in May 1974: and due to increases in price contingencies. USS 22.6 million in February 1975 at Board approval. IDA's contribution was increased from US$ 8.8 million to US$13.6 million representing 60 % of costs. No cofinancer was found. 2.08 Points raised by Executive Directors, prior to Board presentation included queries on : i) the reason why the estate/plantation approach had been adopted, which seemed to have failed in the past and also might be in conflict with the Bank Group's commitment to rural development (para. 2.10); ii) what use would be made of the wood from trees to be felled during land clearing (para. 3.13); iii) whether there was an alternative oil generating crop which could be produced more competitively since there was a high tariff on imported palm oil (para. 4.01) and iv) whether consideration had been given to possible soil erosion due to land clearing (para. 4.04). 2.09 The project was presented to and approved by the Board on Februa-rv 18, 1975. The Credit Agrement was signed Marched 3, 1975 and the Credit became effective June 30, 1976. B. Proiect Description 2.10 The project comprised the first phase of an oil palm development program which was to utilize the nucleus plantation with outgrovers approach. It was thought that such an approach would offer advantages in the spheres of employment, extensior. and large scale central processing. The employment opportunities on the nucleus plantatiop were to provide a regular source of income for members of the outgrower community. Finally, the technical experience gained by the outgrowers working in the nucleus plantation under well supervised conditions was to serve a valuable extension factor in the outgrower planting sector of the program. 2.11 The project was to be carried out over eight years from 1975 to 1982 and involve the planting of 5,200 ha of oil palm as a Phase I of a total 10,100 ha planting program. Amendments to the project credit agreement were proposed to Government on November 6, 1974 with specific mention of the construc- tion of a 10mile access road and the provision of technical assistance to the Agricultural Development Bank (ADB). Government approved these amendments at the end of January, 1975. The project was thus to include : (a) establishing and operating a 4,000 ha nucleus plantation, (b) providing technical services and credit facilities to develop 1200 ha of outgrower plantings adjacent to motorable roads leading to the plantation, 1/ Decision Memorandum of December 20, 1972 and Addendum dated February 28, 1973.  - 15 - (c) constructing and operating a 10 ton/hour palm oil mill (designed for possible extension to 20 tons) and establishing a fruit collection system for outgrowers, (d) building ten miles of road in the vicinity of the nucleus pi itation, (e) training personnel, (f) carrying out field trials with oil palms, and, (g) providing technical assistance to the Agricultural Development Bank (ADB). 2.12 Plantings, which were to be completed during the first five years, would all be in bearing within 9 years. The project was to be implemented by the Ghana Oil Palm Development Corporation (GOPD), a statutory corporation established by the Government and to operate as a commercial entity with responsibility for developing and operating the nucleus plantation, constructing and operating the palm oil mill, organizing and managing credit and extension services, and fruit collection for outgrowers. The board of GOPD was to formulate policy and approve operational programs, delegating responsibility for their execution to the Managing Director. Finally, GOPD's senior staff positions were to be provided under contract by an internationally recognized firm experienced in oil palm planting and management. 2.13 The prcject's main impact at full development would be to generate an annual production of 14,000 tons of palm oil and 3,000 tons of palm kernels. The project would sustain and develop an industry generating a gross income of US$ 17 million annually in a poor and relatively less developed region of Ghana. Project induced oil production would substitute for imports that otherwise would be required and kernels would be exported. Finally, there would be several unquantifiable benefits including : i)providing employment for about 250 farmers and 600 staff and laborers at GOPD headquarters and on the plantation; ii) employing and training skilled labor in construction, roads, water and electrical supply; and, iii) establishing an efficient marketing system for smallholders 'palm produce. III. IMPLEMENTATION A. Effectiveness and Start-up 3.01 Six conditions of effectiveness were included in the Development Credit Agreement for the project, namely that : a) Government had acquired and transferred to GOPD not less than 1200 ha of land suitable for oil palm planting in the nucleus area; b).GOPD had entered into a contract with a plant-breeding consultant on terms and conditions acceptable to IDA; c) financial arrangements, acceptable to IDA, between Government and GOPD for providing Government and IDA funds to GOPD had been concluded; d) a person acceptable to IDA had assumed full time duties as Managing Director of GOPD;  - 16 - e) GOPD had entered into a contract with consultants experienced in oil palm planting and management acceptable to IDA; and, f) the Form of the loan agreement for outgrowers had been approved by IDA. 3.02 Although the Credit Agreement between IDA and the Government was signed on March 5, 1975, the IDA Credit did not become effective till June 30, 1976, after the date of effectiveness had been repeatedly postponed from June 13, 1975 to , successively, September 30, 1975, December 31, 1975, March 31, 1976, April 30, 1976, and May 31, 1976. At first the delays were essentially due to the principal requirements relating to the appointment of a Managing Director and the contract with consultants remaining unfulfilled. Also the Government needed additional time for the acquisition and transfer of land to GOPD. By May 1976, three conditions still had not been met, namely, the management contract, the contract with the plant-breeder and the legal opinion. The project was finally declared effective on June 30, 1976. B. Physical Implementation 3.03 Project physical implementation experienced a one vear delav from the date of effectiveness due to : (i) remaining difficulties in acquiring and transferring the necessary land to GOPD as a result of slow compensation of dispossessed farmers by Government; (ii) the late switch from a manual to a mechanized land clearing policy; and. (iii) late arrival (April 1977) of the heavy eauipment. As a result, in 1978 the appraisal planting targets had to be revised. The availabliliry of sufficient and suitable land remained a major preoccupation throughout the project. A land survey promised by Government for Septeuoer 1971 r-,as only completed in March 1974. It was therefore difficult for the preparation and appraisal missions to. challenge the Government's assessments that there would be no problem in getting the amount of land necessary for project implementation. Also difficult to assess was the likely participation of farmers in an oil palm project, in a predominantly cocoa growing area. Concern about Government's ability to acquire the land and the cooperation of the affected farmers was expressed within the Bank as early as June 1972 and well into 1974. According to comments on the Bank's proposals by German Aid officials in February 1974, the Chief of the Kwae village knew hardly anything about the project and was opposed to replacing cocoa trees by oil palm. In February.1975, the Bank was still enquiring about the steps Government was taking to acquire land for the nucleus plantation. Land for the nucleus plantation was finally made over to Government by Decree in September 1975, under the Administration of Lands Act. Bovever, for technical reasons related to the transfer of this land from Government to GOPD, this vas revoked on March 12, 1976. Subsequently the transfer was made under the 1962 State Lands Act by Executive Instrument 30. A 50-year lease for the land was signed by both the Chairman of GOPD and a representative of the Commissioner of Lands on April 3, 1976. This step theoretically fulfilled one of the remaining conditions of effectiveness which was that Government had acquired and transfer- red to GOPD, not less than 1200 ha. Compensation 3.04 Land compensation remained a bone of contention throughout the project mainly for three reasons. Firstly, the Government intended to resettle the Kvae  - 17 - villagers outside the plantation and at the same time allocate them some 1000 ha of land in the adjacent Aiyaole forest reserve for village food crops in place of the conceded land. The Ministry of Forestry never complied with this latter commitment. As a result, the Kwae villagers hampered project implementation, especially the final clearing and planting phase. It therefore took six years to plant 5,145 ha when appraisal estimates were for 5,200 ha to be completed within four years. 3.05 Secondly, too many Ministries were involved with the project. While direct responsibility for the project was with the Ministry of Agriculture, the Bank's contact was the Ministry for Finance. The allocation of land from the forest reserve had to be approved by the Ministry of Forestry, while the rate of compensation was decided by the Ministry of Finance and distributed through the Ministry of Lands, Natural Resources, Fuel and Power. Despite several requests, Project management never obtained more than global figures on the amounts paid to farmers. 3.06 Thirdly, farmers were dissatisfied with the compensation rate of 4 0.50 per cocoa tree paid to them till mid-1980. This was especially the case in the concession for smallholders at Okaikrom where the villagers refused to move from approx-imately 385 ha of land needed to meet 1981 and 1982 planting targets. Some 3,000 ha of cocj:: had been compensated at that low rate, totalling about 4 14 million. By end-1960., farmers were asking for e 5 to d 8 per tree, depending on the age. Claims for compensation for other c-rops were made as well. Faced with a 0 60 million payment adjustment for compensation of the initial 3,000 ha, in addition tJ compensation for a remaining 1,000 ha, the Government in 1982 ordered an in-depth investigation of past and present claims. Project management which had always deplored the lack of cooperation with officials on the compensation issue, highlighted that most farmers'claims were over-inflated while seldom challenged in the field. Real progress only started by end-1982 when payments of up to d 2000 were made to individual farmers. This represented part-payment for some growers and complete payment for others. 3.07 Oherall, Government had no clear cut policy to zackle the issue. It had little means to cross-check farmers claims and, at times, completely lacked the necessary funds. This situation resulted in the planting targets not being met as farmers on eight occasions, between February 1980 and September 1982, obstructed deforestation, demarcation and windrowing operations, each disruption lasting for several weeks, even months. Land Preparation 3.08 It was decided during project preparatiorr and appraisal, that land would be cleared manually by local labor and the use of chainsaws. Import substitution and the creation of local employment were the project philosophy. The manual option was clearly retained during an April 1972 pre-appraisal mission which stated that mechanical land clearing and windrowing prior to planting were not acceptable. This approach was completely reversed in 1975 by the IRHO management team winich supported mechanical clearing instead, going so far as threatening to withdraw from the project should their view not be adopted. The reversal was based on the prevailing Ghanaian labor situation characterized by labor shortages and possible difficulties with the labor unions. As a matter of fact, only 20 ha had been cleared manually over a three month period when the project was initiated. IDA indicated, in April 1976, that it had not objection to mechanical clearing provided that unemployed labor in the project area be given job opportunities. Finally, the GOPD Board, On the advice of IRRO and with previous agreement from the Bank, decided on September 7, 1976, that land clearing wDuld be done mechanically.  - 18 - 3.09 Because of the two year delay in starting project implementation, the project was allowed to forego ICB in purchasing the land clearing machinery. Three bulldozers were ordered in October 1976 and arrived in April 1977. The equipment chosen proved to be inadequate for land clearing. There were no canopies to protect the drivers, gearing was too high and the pushers too weak. Furthermore the overall performance of the machines was belov standard because of poor maintenance, lack of spare parts and insufficient training of the operators. Some 100 ha vere cleared per month when the normal average should have been 200 ha. In July 1980, an expatriate workshop manager was recruited, although not provided for at appraisal. Maintenance improved markedly and the land clearing program made up for some of the delay. Two more bulldozers of another type were purchased and became operational in March 1979. This had a positive impact on project implementation. Overall, the land clearing program was as follows : 1977 1978 1979 1980 1981 1982 Total Deforestation (ha) Manual 8 42 162 128 214 84 638 Mechanical 190 734 780 1 399 354 808 4 265 Windrowing (ha) 1/ 190 574 734 936 487 647 3 568 Total hours 2/ 1 766 2 052 1 B42 2 777 262 1 942 10 641 Deforesting (hours/ha) 3/ n/a 2.8 2.4 2 * 2.4 Windrowing (hours/ha) 3/ n/a 4.5 4.4 3.7 3.5 * 1/ As at September 30, 1982. 2/ of bulldozer for both deforesting and windrowing. 3/ Figures unreliable. * Work was often interrupted by dissatisfied farmers. 3.10 Once the surveying and demarcation had been completed, bulldozers equipped with tree-pushing and dozer blades carried out land clearing. Trees too large for the tractors to push over were felled by chainsaw. All vegetation was then windrowed in alternate rows oriented North-South, using stacking rakes. The alternate windrowing practice, one row in every two, was aimed at minimizing soil erosion and loss of fertility. Strict monitoring of land clearing operations prevented the top soil being unduly disturbed. Finally the legume Pueraria phaseolides was sown shortly before the rainy season. Because local labor refused to harvest puerarie seed, for fear of snakes, the project had to import about 30 t of seed from the Ivory Coast, at a cost of about CFAF 16 million, or US$ 64,000 equivalent, in foreign currency, excluding transport. 3.11 There is no mention in the appraisal report of how the vast quantity of wood and timber were to be disposed. No proper survey quantifying the merchantable timber on the area cleared was carried out. As a result, the project relied on various contractors taking away whatever wood they wanted at bargain prices. A better use of the timber could have been t'-rough the project's large construction program and/or by making charcoal 'ihi--h fetched high prices in urban centers. Roads 3.12 It had been agreed during negotiations that Goverrent would build a 10 mile (16 km) feeder road in the vicinity of the nucleus plantation and would maintain all roads within a 15 mile radius of Kwae but not on the nucleus  - 19 - plantaticn. In fa.:t, due to the lack of equipment, Government did not keep to its commitment. Road construction was therefore mainly entrusted to a contractor, supervised by GOPD's building consultant. Three types of roads were built; namely. i) 4 m wide East-West roads, ii) 6 to 8 m Nortb-South roads, and iii) 6 to 8 m boundary roads. Some 92 Im of well graded and graveled roads were thus constructed. Maintenance included the newly opened roads as well as the 30 km portion between Kade and Kwae. 3.13 The pace of road construction work dropped sharply in 1979-80 when the contractor started facing financial problems. As a result, the road construction program was delayed, which in turn had a negative impact in 1982, when 1978-plots were ready for harvesting. By mid-1982 GOPD had no other alternative than to actively participate in the program, when both a new bulldozer and a loader became operational. No roads were built for outgrowers. The project had, by the end of 1982, constructed an additional 24 km of roads of which 5 km was for smallholders. Field Establishment 3.14 Oil palm plantations for both tne nucleus estate and the smallholders/ outgrovers component were established from 1977 to 1982, as shown in Annex table 1. It was anticipated at appraisal that the bulk of the project seed requirements would be supplied by the Oil Palm Research Center (OPRC) at Kusi, by IRMO and by the Nigerian Institute for Oil Palm Research, NIFOR. In fact for quality reasons, most of the seed came from the IRHO Research Station at La Me, Ivory Coast, at d 4 per seed compared to 4 2 for local seed. Also, supplies from other sources like Benin were limited. Since tnis situation was expected, funds had been included in the project for a plant-breeding consultant to visit O?RC tvice yearly. The consultant was to assess the suitability of plant material produced by OPRC and monitor the seed breeding and propagation- programs. To that effect Government had agreed to provide OPRC with foreign currency for buildings, laboratory equipment, vehicles, irrigation equipment and seed, for a total amount of about USS 230,000. Engaged in early 1978, the consultant made two short visits to Ghana. He then refused to return when it became obvious that no foreign exchange could be expected from Government and that consequently none of his recomuendations would be taken into consideration. 3.15 The plant-breeder's analysis-revealed that seed production techniques at the Kade station were less than adequate though some of the existing genetic material could have allowed the production of excellent variety combinations ie. 5.368 Deli x 32.3005 NIFOR. With an improvement of production techniques Kade would thus have been able to supply about half of the country's required planting materials. Wnile the European Development Fund, EDF, was willing to consider provision of tne necessary foreign exchange, Government preferred to wait for the 1980-81 Lome Convention before asking for funds from EDF. However since no action was-taken, Ghana became almost entirely dependent on oil palm seed from abroad. 3.16 The planting material received by GOPD was in both the form of germinated and heat treated seed, at the rate of 280 seeds per hectare to be planted. The former needed to be dispatched within three days while the latter allowed for a more flexible dispatching schedule. While germinated seed was planted upon arrival, heat-treated seed required 15 days to germinate. The seed was placed in a pre-nursery for 3-4 months for easier control and better homogeneity of the plants, as well as because of continuous delays in the delivery of large polybags. After proper selection, some 220 plants-were left out of each batch of 280 seeds introduced for transplanting in the nurseries  - 20 - where seedlings remained for between 8 and 12 months. Urea and the compound fertilizer 15-15-6-4 were applied and pest control was taken care of by using various pesticides. Plants were grown under shade until 1981. However Temik had been used on a trial basis since 1978179 against blast disease on unshaded plants. In 1982, the Plantation Manager decided not to shade the nurseries and obtained healthy seedlings. 3.17 Density of planting was 143 trees per hectare, planted in triangles at 9 meter intervals. Transplanted seedlings were protected in the field by wire netting sleeves to prevent damage from rodents among which the grass-cutter or agoutis was the most common. Maintenance 3.18 Regular maintenance work was carried out on the plantations as they were established. This consisted in maintaining clean circles around each young plant while controlling the growth of the cover crop Pueraria. Circle weeding w-as done either manually or using chemical herbicides. Also the regrowth of vegetation in the wuindrows had to be periodically cut back, using machettes. Additional maintenance work included the yearly application of chemical fertilizers and the replacement of trees up to year 3 - less than 5 Z. Finally, castration palm oil trees required special attention, trained staff and a firm stand from project management as the practice was unpopular in Ghana. The castration, which is common on IRHO plantations throughout the world, and which is still being advocated, was necessary even the moreso as the oil mill was not ready to process the harvest. Castration of oil palm trees consists in the removal of all young male and female inflorescences and bunches. This practice which is meant to favour the growth of the young palms until they are three to four years old, is carried out every month co=mencing in the sixteenth month. At that tine some 50 per cent of the trees produce their first inflorescence. In fact, experiments have shown that where plantations are castrated at an early stage, production rises markedly in the following years and usually makes up the difference. At the time of the mission, most palm trees looked strong and healthy. Based on leaf analysis and girth measurements and taking into account the possible recurrence of severe droughts like the one that occured from December to April 1983, IRRO estimates the average yields at full maturity at about 12 t/ha ffb on the nucleus estate and 7 to 8 t/ha ffb on the outgrowers/smallholders plots. The oil extraction rates should be 21 and 20 percent respectively. These figures compare with the slightly higher productivity forecasts made at appraisal of 13 t/ha ffb on the nucleus estate and 10 t/ha ffb for the outgrowers with an overall oil extraction rate of 21.5 per cent. Finally, of concern was the fact that no application of fertilisers was planned for the first half of 1983, due to the lack of foreign exchange. Smallholders/Outgrowers 3.19 The IRHO feasibility study in 1972, defined smiallholders as farmers receiving project assistance to plant oil palm on their own land while having access to credit facilities. A rather similar definition was given to outgrowers in the appraisal report. However, due to the lengthy delays in obtaining the necessary legal documents for the outgrowers and the lack of access roads, the Managing Director suggested, in late 1977, that a block of smallhol- dings be created in the area located north-west of and contiguous to the nucleus estate. Smallholders were therefore farmers taking up project land leased from GOPD for cultivation of oil palm. Their selection by a loan committee including  - 21 - Project staff, representatives of the Ministry of Agriculture and of ADB was based on the candidates having a minimum land, labor and funds. GOPD provided chainsaw teams for deforestation, as well as personnel for surveying, lining, and pegging. Each smallbolder received 8 ha of land to be planted under oil palm at an annual rate of 4 ha and about 1 to 2 ha for growing foodcrops. 3.20 The introduction of the term smallholders, as distinguished from outgrowers defined in the Development Credit Agreement, required the amendment of the Development Credit Agreement, the Project Agreement, the Financing Agreement and the Agreement between GOPD and Smallholders. This was officially agreed to by the Bank, in June 1980. The initial smallholder program did not experience many difficulties as land was allocated to nearby villagers. Credit was also supplied by GOPD to participants in the form of planting material, fertilizers and other non-labor elements supplied in kind, as well as up to 80 Z of estimated labor cash requirements. The project provided technical assistance throughout the development phase and vas responsible for fruit collection. In return, the smallholders were committed to sell their harvest to the project for processing at the central mill. The production of fresh fruit bunches (ffb) by smallholders was still too small by the end of the on-going pro-ect to assess whether smallholders were actually complying with the above comnitment. 3.21 Most smallholders felt that 8 ha of oil palm would require too much work as their families decreased in size and they got oldey. Because of increasing difficulties in obtaining additional land for the smallholder program, the project reintroduced the outgrower concept by which farmers, within a 15 mile radius of the nucleus estate and near a motorable road, could grow oil palm and obtain both advice and credit facilities. The outgrowers planted an average 4 ha of oil palm each, which tends to substantiate the smallholders earlier statements that 8 ha were too nuch. By Septenber 1982, the project was looking after 202 smallholders and 116 outgrowers, as against 152 outgrowers planned at appraisal. Labor 3.22 The availability of sufficient labor remained throughout one of the Bank's main concerns. During preparation in 19i2, indications were that up to 1000 laborers per day were available, based on the assumption that the manual land clearing policy would be adopted. In March 1974, queries were again put to Government, as to both the exact labor situation and the likely cooperation of dispossessed farmers. A complete reversal took place in September 1976 for reasons mentioned in para. 3.10; other reasons included (i) the labor policv at the time of appraisal was to employ only non resident labore_s - tnis proved unirealistic as most workers living outside the estate were mainly seasonal, made up of smallholders with a low output and who therefore were very costly; (i;) the dispossessed farmers, dissatisfied with both the rate and the pace of compensation by Government, refused in many cases to work on the project; (iii) the project had limited facilities to accommodate a sizeable labor force and lacked the means to safely transport large numbers of workers from distant villages; up to 30 % of working hours were wasted on transportation - a housing program for the workers was subsequently adopted on which progress has been very slow due to acute shortage of building materials;  - 22 - (iv) the salaries paid by GOPD vere always lower than on the other oil palm estates - this unfavorable situation was gradually reversed starting in May 1981 when the project benefitted from the World Food Program following arrangements with the Government. Rations of rice, fish and edible oil were distributed on a monthly basis to both the project staff and to the plantation workers. The distribution was further extended, a few months later, to the smallholders/outgrowers as well to the labor they employed. Given the at times, scarcity of basic food stuffs, the World Food Program had indeed a positive impact. Labor absenteism dropped sharply from 30 % to about 18 2 during the peak season whereas farmers showed keener interest in the project. Construction Program 3.23 The project construction program as described in the SAR mainly included offices, houses for senior staff, a training and a health center, a workshop, storage facilities and a small dam. IDA was to finance USS 1.79 million, representing 58 Z of total expenditures. Realizing the high cost versus the low output of non-resident labor, project management in September 1976 appealed to the GOPD Board to consider establishing modern villages at the project site to house a corps of f-eld and factory workers. Meanwhile the cost of housing had already increased by 2-3 times since appraisal. This led a Bank consultant to recommend temporary instead of permanent housing in order to reduce expenses. The Managing Director presented in October 1976, a i 5.3 million (US$ 2.37 million) building program to be spread over five financial years, aimed at housing at least half the labor force required to run the nucleus estate. 3.24 In addition to the contractor which was responsible for the building program, the project hired the services of Consultant Engineers to oversee the construction work. Progress remained slow due to continuous shortages of building materials and cost overruns were substantial. Again, in 1978 suggestions were made, this time'by the Plantation Manager, for the project to provide temporary wooden houses for workers, The Government consistently reiected this possibility, preferring houses made of concrete and the Consultant Engineers showed little enthusiasm for investigating what could have turned out to be a cheaper solution. In March 1980, the Ministry of Finance expressed considerable concern at the high unit costs of construction. In order to speed up the construction program while lowering the costs, a Bank mission suggested in September 1980, that a Brick plant be imported. Following positive results from soil surveys which located high quality clay at the project site, a USS 240,000 brick and tile plant was ordered from Brazil. It arrived in February 1982 but was still in crates by June 1983. The equipment arrived in February 1982 followed a few weeks later by the arrival of the Brazilian technicians. The latter were to put up the small brick plant and co=nission the equipment. Once at the project site the technicians finding the living conditions unbearable left without notice, after rwo days. Although several complaints were lodged with the Braziiiar. Embassy in Accra, no replacement team had been sent by the end of June 1983. So far, houses have been built for the senior and most of the middle level staff. However, none is as yet available for laborers. Oil Mill 3.25 The contract to build the palm oil mill at Kwae was awarded by the Board of Directors of GOPD, on January 21, 1980. The award of contract caused complaints from bidders (see para. 3.45). Started in June 1980, the mill was coamissioned on November 8, 1982, one month ahead of schedule. This was a major achievement given the country's poor economic situation.  - 23 - Training 3.26 Both Government and the Bank rightly considered the training of the local staff as a major goal. As such, training provisions vere specifically included in both the contract vith the managing firm, IRBO and in the project agreement (Section 2.03) with the aim of having qualified Ghanaians replace expatriate staff when time came. The project thus financed : i) a two-week course at the Mananga Agricultural Management Center, Swaziland for four out six Assistant-Plantation Managers. A ban on overseas training prevented the sending in 1982, of the two remaining candidates; ii) a one montb training session on harvesting and collection systems, in the Ivory Coast, for the Senior-Assistant-Plantation Manager; iii) a ten month course on operating an Oil Hill, in both the Netherlands and Malaysia, for two mill engineers; and iv) a three year course in Australia, for an Assistant-Plantatioa Manager to obtain a M.A. in Agricultural Economics. 3.27 By contrast, the middle level staff received little training. In fact, the training needs had not been well identified at project inception. Furthermore, responsibility for training had been left with expatriates whose ability to train had not been assessed and whose primary duties were of a technical and executive nature. Also the difficult circumstances in Ghana at the-period did not leave much time for training. Field Trials 3.28 Although the project was to carry out field trials with oil palms, no specific funds were earmarked for that purpose 1/. Tne applied research program was thus to be limited. Emphasis was to be given to nursery and field establish- ment practices together with harvesting and post maturity maintenance techniques. Cooperation between GOPD and OPRC was initially very good. However, due to lack of funds it soon became difficult for the Research center to participate as anticipated. Because GOPD's financial contribution was small - US$ 30,000 and limited to the plant-breeder consultant's visits which were discontinued at an early stage (para. 3.16 and 3.17) little progress in seed production was achieved. GOPD's contribution to research as part of its field operations included information on : i) deforestation - windrowing every four instead of two rows, to improve the output of the heavy equipment; ii) nurseries -- suppressing shade after using Teik, a standard practice elsewhere; iii) plantation - basing fertilizer applications of nitrogen and phosphorus on leaf analysis; iv) anti-pest measures -- applying free acid tar on young trees to prevent destruction by grass-cutters which are common rodents; v) harvesting - introducing transportation methods of ffb by workers, similar to the methods prevailing in Asia and in other parts of ;'est Africa using two baskets and a bamboo pool and, vi) landscape - construction bunds according to contour lines, on plots showing marked water deficit. None of the above was original work but merely an adoption of methods evolved in other oil palm growing countries. 1/ SAR : See project description p. 7 and Project Cost Amnex 5 table 1.  - 24 - Technical Assistance to ADB 3.29 Technical assistance to improve ADB's lagging accounting system and controls, credit policy and procedures was originally included in the Sugar Rehabilitation project - Credit 354-GE. Two consultants Btarted their assignment in early 1973 and published their findings in 1973 and 1974. It was however felt that in order to enable the consultants to implement their recommendations a 27-month extension of their assignment was necessary. In view of difficulties in reallocating further project funds under che Sugar Rehabilitation Credit, IDA agreed to include the financing of the extension (about US$ 200,000) under the oil palm project Which was then being processed. A contract for a one-year extension of the contract was signed in November 1974. 3.30 Most of the consultants' recommendations have been implemented however, many of the latter were found to be of a standard difficult to attend locally and had therefore to be modified. One component of the consultants' assignment which seems to have succeeded has to do with training. It is nevertheless unfortunate that inspite of the apparent improvements in ADB's internal operations under the Sugar Rehabilitation project, the results would indicate that the underlying objective of the exercise was not achieved. C. Project Costs and Financing 3.31 Actual project costs including duties and taxes, for the seven years during which 99 % of the plantings were completed totalled US$ 47.8 million instead of US$ 22.5 million (SAR). This is 114 Z above the appraisal estimates. With the exception of Technical Assistance costs to ADB (USS 179.000) the actual cost of every component greatly exceeded appraisal estimates. In US Dollar terms the actual total project cost was more than double the appraisai estimare. Tne increase was even higher in cedi terms, by almost five times. .;he following main factors were responsible for the wide, unfavorable differences between actual and appraisal costs (See Annex-tables 2-3-5). 3.32 Exchange Rate and Inflation (i) The variation in the exchange rate between the cedi and the US dollar from US$ 1 - i 1.15 in 1975 to US$ I - 4 2.75 in 1980 onwards; and (ii) very high annual inflation rates in Ghana, during project field establishment. Inflation rates reached 116 Z in both 1977/78 and 1980/81. The lowest inflation rate during the period under review was 44.3 Z in 1979/80. Although the exchange rate varied it did not reflect the true value in dollar terms 1/. As a result, it should be noted that the dollar equivalent of local costs is extremely high because of conversion at the offical rate of exchange which was significantly overvalued during the last few years. Most of the cost overrun reflects this since differences between appraisal and actual costs are mainly in local terms. 3.33 Land Compensation. Initially estimated at $ 534,000 at appraisal, the total payment for land compensation of dispossessed farmers reached 4 14 million (US$ 2.9 million equivalent) by September 1982 with still more to be claimed. As the price paid per cocoa tree until 1979 was estimated to be insufficient, 1/ At the end of 1982 the black market exchange rate for the cedi was US$ 1 - 60 to 70 i in contrast to the official rate of US$ 1 - d 2.75.  - 25 - Government raised compensation from 4 0.5 to between