Uganda - Smallholder Tobacco Project

Document of FILE COPY The World Bank FOR OFFICIAL USE ONLY Report No. 4235 PROJECT PERFORMANCE AUDIT REPORT UGANDA SMALLHOLDER TOBACCO PROJECT (CREDIT 212-UG) December 23, 1982 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. WEIGHTS AND MEASURES 1 acre .405 hectare 1 hectare = 2.47 acres 1 kilogram = 2.2 pounds ABBREVIATIONS BAT - British American Tobacco Company IDA - International Development Association MMC - Ministry of Marketing and Cooperatives NTC - National Tobacco Company OED - Operations Evaluation Department PCR - Project Completion Report PMB - Produce Marketing Board PU - Project Unit (in the Min. of Marketing & Cooperatives) SLA - Subsidiary Loan Agreement UCB - Uganda Commercial Bank EXCHANGE RATES Currency Uganda Shilling (USh) Year: Appraisal Year Average US$1 = USh 7.00 Intervening Years Average (1971-74) US$1 = USh 7.14 Intervening Year (1975) US$1 = USh 8.25 Completion Year (1976) US$1 = USh 8.31 FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT UGANDA SMALLHOLDER TOBACCO PROJECT (CREDIT 212-UG) TABLE OF CONTENTS Page No. Preface ................................................................. i Basic Data Sheet ...................................................... ii Highlights ......................................................... PROJECT PERFORMANCE AUDIT REPORT I. SUMMARY ..................................... ....1....... II. ISSUES ....................................... .......... 4 A. Project Management and Organization .................. 4 B. Technical Feasibility ................................ 6 C. Tobacco Prices ...............................7......7 PROJECT COMPLETION REPORT I. Background .................................. .......... 11 II. Identification, Preparation and Appraisal ................ 13 III. Project Implementation ............................ ...... 17 IV. Agricultural Impact ..................................... 33 V. Economic Returns ......................................... 37 VI. Institutional Performance ............................. 38 A. Institutional Design ........................ ...... 38 B. The Project Unit .................................... 39 C. Reporting ............................................ 41 D. Accounts and Audits .................... ............ 41 VII. Bank Performance ......................................... 42 VIII. Conclusions .................................. ......... 45 Annexes I - IX Map - IBRD 2707 Smallholder Tobacco Project 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 UGANDA SMALLHOLDER TOBACCO PROJECT (CREDIT 212-UG) PREFACE This is a performance audit of the Smallholder Tobacco Project in Uganda, for which Credit 212-UG in the amount of US$4.0 million was approved in July 1970. The Credit was closed on March 31, 1976, with a delay of three months. The Credit was fully disbursed with the final payment made on April 8, 1976. The audit report consists of a memorandum prepared by the Operations Evaluation Department (OED) and a Project Completion Report (PCR). Because of the political turmoil in the country it was not possible to visit Uganda to obtain information to prepare a PCR at the time of completion of the project. The delayed PCR was prepared by the Resident Mission in Eastern Africa and is based on a review of Bank/IDA files and recent reports of the Uganda National Tobacco Company, which has taken over the Smallholder Tobacco Project, and on the report of a recent sector mission to Uganda. The audit is based on a review of the Appraisal Report (No. PA-33a) dated April 13, 1970, the President's Report (No. P-850) of July 8, 1970, thA Development Credit Agreement (No. 2124-G) dated July 24, 1970, the PCR, relevant Bank/IDA files, the minutes of Board discussion, and interviews with Bank staff who have been associated with the project. A copy of the draft report was sent to the Borrower on October 7. 1982 for comment. However, none have been received. On the basis of this abbreviated procedure the audit agrees with the general findings of the PCR. The audit memorandum elaborates on the issues of design of management, technical feasibility, and product pricing which are important for future projects to be supported by Bank/IDA financing in Uganda and other countries.  - ii - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET UGANDA SMALLHOLDER TOBACCO PROJECT (CREDIT 212-UG) KEY PROJECT DATA Actual or Actual as % Appraisal Estimated of Appraisal Total Project Cost (US$ million) 7.26 10.95 152 Credit Amount (US$ million) 4.00 4.00 /a 100 Date Physical Components Completed 06/75 n.d./b - Proportion Completed on Above Date (%) 100 50 T 50 Proportion of Time Overrun - n.d. - Economic Rate of Return (%) 27 Negative - Financial Performance - Very poor Institutional Performance - Very poor CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS US$'000 FY71 FY72 FY73 FY74 FY75 FY76 Estimated 1,250 2,337 3,086 4,000 - - Actual - 397 1,762 2,171 2,341 4,000 Actual as % of Estimated 0 17 57 54 59 100 Principal Repaid (09/30/82): US$0.10 million MISSION DATA Sent Month/ No. of No. of Man- Date of By Year Weeks Persons weeks Report Identification n.d. n.d. n.d. n.d. n.d. Preparation RMEA n.d. n.d. n.d. n.d. n.d. Appraisal HQ 08/69 n.d. 4 n.d. 04/13/70 Supervision I HQ 07/70 .4 1 .4 08/06/70 Supervision II RMEA 05/71 2 3 6 06/08/71 Supervision III RMEA 12/71 1.6 3 4.8 02/01/72 Supervision IV RMEA 04/72 1 2 2 n.a. l Supervision V HQ 06/72 1.6 1 1.6 07/24/72 Supervision VI RMEA 05/73 .4 2 .8 05/08/73 Supervision VII RMEA 03/74 1.5 1 2 03/22/74 Supervision VIII RMEA 10/74 .5 2/e 1 11/15/74 Disbursements RMEA 03/75 .4 1 .4 07/09/75 9.4 9.0 OTHER PROJECT DATA Original Actual or Plan Revisions Estimated First Mention in Files or Timetable n.d. - 03/68 Government's Application - - 03/68 Negotiations n.d. - 03/70 Board Approval n.d. - 07/21/70 Credit Agreement n.d. - 07/24/70 Effectiveness Date 10/31/70 12/31/70, 03/01/71, 07/28/71 05/01/71, 08/02/71, Closing Date 12/31/75 03/31/76 /f 04/76 Borrower Republic of Uganda Executing Agency Ministry of Marketing and Cooperatives Follow-on Project None /a Taking exchange adjustments into account, disbursements totalled US$4.16 million. /b Not determined. 7 Since reports on physical achievements ceased after March 1975, the estimate is based on key indicators as of late 1974. /d Report not on file. 7e Mission included disbursement officer. If See PCR, para. 3.45; but according to Statement of Development Credits (September 30, 1982) Closing Date not officially extended.  - iii - PROJECT PERFORMANCE AUDIT REPORT UGANDA SMALLHOLDER TOBACCO PROJECT (CREDIT 212-UG) HIGHLIGHTS The project, approved in July 1970, aimed to increase tobacco pro- duction through the provision of credit for the construction and improvement of individual tobacco curing barns and cooperative bailing centers; fuel wood plantations, a seasonal input revolving fund, management unit facilities, and project preparation were also included. Total project costs were estimated at US$7.3 million. The project got off to a slow start and effectiveness was delayed by seven months. Security and political uncertainty continually plagued the project following the January 1971 coup d'etat and although supervision con- tinued, IDA staff was unable to visit project sites. From what limited information that is available, it is clear that the project was not success- fully implemented: yields were lower, planted area was much less and costs were much higher than estimated at appraisal. Overall, Uganda exports fell from a record 1,000 tons of flue-cured tobacco to an average 200 tons for 1978-80. It can be reasonably inferred that the rate of return is negative. In sum, the project failed to achieve any of its objectives (PCR, paras. 8.01-8.03). Despite the circumstances prevailing during implementation, im- portant lessons can be drawn for further IDA/Bank lending from this experi- ence: the state of preparation before appraisal (PPAM, paras. 14, 20-21 and PCR, paras. 2.01-2.04), thoroughness during negotiations (PCR, paras. .3.05-3.07), and adherence to covenants (PPAM, para. 11). Other points of interest are: - The Bank decided not to suspend disbursements (PPAM, para. 13 and PCR, paras. 7.08-7.10); - Institutional design was unrealistically complex (PPAM, paras. 16-21 and PCR, paras. 6.01-6.07); - Tobacco production is not a simple technique or production process and necessary conditions for its success were not forthcoming (PPAM, paras. 22-26); and - Lack of adequate price incentives (PPAM, paras. 27-28 and PCR, paras. 4.04-4.11).  - 1 - PROJECT PERFORMANCE AUDIT MEMORANDUM UGANDA SMALLHOLDER TOBACCO PROJECT CREDIT 212-UG I. SUMMARY./ 1. The main objective of the project was to increase flue-cured tobacco production from about 2,700 tons to 5,300 tons at expected full development in 1977-78. Most of this increase in production was to be exported, raising Uganda's flue-cured tobacco exports from about 700 tons to 3,100 tons. About 6,000 small farmers were expected to benefit from the project. 2. Increased production was to be achieved through provision of credit to individual growers to construct and improve tobacco curing barns, credit to unions of cooperative tobacco societies and their constituent tobacco growers cooperatives for construction of baling centers, tobacco stores and office facilities, and for establishment of fuelwood plantations, investments by the Produce Marketing Board (PMB) in additional storage facilities in Kampala, provision of a revolving fund to finance -- through the Cooperative Unions -- seasonal production inputs, and establishment of headquarters and field centers for project administration in the Ministry of Marketing and Coopera- tives, and the Cooperative Unions. The project also provided funds for preparation of further agricultural projects in Uganda. 3. Total project cost was estimated at US$7.3 million of which US$4.0 million would be financed by Credit 212-UG, US$1.2 million by the Government, and US$2.1 million by the cooperatives and farmer beneficiaries. 4. The management structure of the project as envisaged at appraisal was complex, and a large part of the project's success depended on the effi- cient cooperation and interaction of the several agencies in an untried complex structure. 5. The project got off to a slow start. Effectiveness was delayed mainly by failure on the part of the Borrower to execute subsidiary loan agreements with the three cooperative unions and PMB. The credit became effective in July 1971 after a delay of seven months and four extensions of the effectiveness date. 6. Security and political uncertainty continually plagued the project following the military coup d'etat in January 1971. Little, if any, of the expected benefits have been achieved by the project. Because of the inability of IDA staff to visit project sites, it has not been possible to obtain 1/ Adapted from PCR. - 2 - accurate information on project accomplishments. According to available information, total flue-cured tobacco production approached the amount fore- cast only in 1972 when 3,200 tons were reportedly produced compared with 4,000 tons estimated. Total flue-cured tobacco production only averaged about 920 tons in 1978-79. In the West Nile district, where yields were expected to average 450 kg/ac, they fell to 400 kg/ac and planted acreage fell 1,000 ac short of the revised target of 5,000 ac. 7. In the Middle North, plantings were at little more than a third of the 6,530 ac planned and, in Kigezi, two-thirds of the modest target of 800 ac were planted. Similarly, yields in the Middle North and Kigezi remained at around 200 kg/ac compared with 450 kg/ac envisaged at appraisal. Exports, likewise, were disappointing. After a record 1,000 tons of flue-cured tobacco were exported in 1972 compared with 970 tons in 1969, overseas sales declined every year since--except for a recovery to 995 tons in 1976--to an average of 200 tons for 1978-80. The development of fuelwood plantation fell way short of the plan of operations that was developed after the project was approved, i.e., 7,736 acres vis-a-vis 15,000 planned. Eight thousand acres were esti- mated in the appraisal report, but this turned out to be unrealistic after further analysis. 8. Precise figures on total cost of the project at completion are not available. From information that is available, it is roughly estimated that total project cost exceeded US$11.0 million, or 52% above the US$7.3 million estimated at appraisal. If the project had been implemented as envisaged at appraisal, total cost likely would have amounted to US$11.9 million. On the physical side, only 1,802 of the planned 4,711 tobacco curing barns had been constructed or renovated by mid-1974. Only around 40% of the expected baling centers, housing, offices, and supply stores for cooperative unions and local cooperative societies were completed by the latter part of 1972-early 1973. At that time Asian contractors were expelled and cost for all project build- ings had escalated to about US$10 million compared with US$7.7 million bud- geted in the appraisal report. Efforts to resume construction with local contractors in late 1974 met with limited success due to shortages of vehi- cles, spare parts, and building materials. For these reasons and the lack of local funds, it appears that the building program was not completed. 9. A supervision mission in April 1972 reported that the outlook of the project was bleak and that unless changes were made the net benefits of the project (economic rate of return) would be negative. The mission recommended that disbursements be suspended until a new plan could be drawn up which would show a satisfactory rate of return to the project. As a result of this recommendation, a special review mission was sent to Uganda in July 1972. The mission concluded that the project could be made viable by expanding the tobacco acreage in both the West Nile and Middle North areas. In order to complete such a plan, it would be necessary to change the completion date of the project from 1974 to 1976. -3- 10. IDA accepted these recommendations and requested the Government, if it agreed, to submit a new plan of operation that would incorporate such proposals. A proposal to implement part of the plan was sent by the Project Director to the Ministry of Finance in February 1973. Apparently the proposal was overcome by events, the security problem became serious, especially in the Middle North in the fall of 1972. There is no record of any further correspondence on the proposed expansion of the project. 11. Most of the covenants were partially or completely disregarded by the borrower after 1971, in spite of the insistence of IDA and the best efforts of project management to honor them. Accounts were not adequately maintained for lack of financially competent staff; audit reports are vir- tually non-existent or highly qualified and were out of date when they were submitted to IDA. 12. Progress reports were rare, late, censored, and insufficiently comprehensive to supplement the findings of IDA supervision missions. Hence, a clear view of progress of the project as a whole and of individual com- ponents was lacking, especially after 1972 when statistics became unreliable and supervision was made difficult by the political authorities and security risks in the country. Equipment required under the project was either destroyed or diverted to other uses during years of political upheaval after 1972. 13. In spite of most adverse project performance and deteriorating economic conditions in the country, IDA agreed to continue disbursement for verifiable and approved expenditures, after repeated progress reviews in 1972, 1973 and 1974 and reassessment of project risks by higher Bank management. Supervision of the project became difficult from 1973 onward due to restric- tions imposed by the Government and security problems in the field. IDA considered suspending disbursements in 1972 and again in 1974, but lacking incontestable evidence obtainable only through supervision missions--and because of censorship imposed on reports to IDA by project management--it was concluded by higher management that suspension of disbursements would have undesirable side effects.1! 14. The project began with the handicap of inadequate preparation and being implemented by an untested organizational structure marked by in- experienced, weak and fragmented management. Successful implementation would have depended on the consensus of three Government organizations working 1/ On this subject, see also PPAM, Ethiopia - Lower Adiabo Development Project (Credit 516-ET), OED Report No. 3971 dated June 17, 1982 and PPAM, Zambia - Integrated Family Farming Project (Loan 882-ZA), OED Report No. 3746 dated December 30, 1981. harmoniously with each other and with one cooperative union in each district and their constituent tobacco growers, cooperative societies; all seeking to attain similar objectives. The establishment of a project unit (in the Ministry of Cooperatives) dependent upon staff seconded from the Ministry of Agriculture, led to early problems that became progressively more acute. Prices paid to farmers for flue-cured tobacco were not increased to stimulate production and recruit new growers. Thus, while the PMB export marketing margins increased from USh6/-/kg in 1969 to over USh9/-/kg in 1974, failure to raise prices to farmers made it impossible for all but the most efficient and perseverant producers to continue in flue-cured tobacco production. Since the margins needed by PMB to cover costs of minimal marketing services for export never exceeded USh5/-/kg, farmers were, in fact, consistently "taxed" in the order of USh2/- to USh3/-/kg while their share of final tobacco prices shrank from 50% in 1968-69 to 40% in 1972-73 and 37% in 1973-74 on the average. 15. It has not been possible to calculate an economic return to the project due to the lack of data as cited above. However, because no tangible economic benefits were attained and cost exceeded appraisal estimates, it can be inferred that the economic rate of return to the project is negative. II. ISSUES A. Project Management and Organizationl/ 16. The PCR notes that the institutional design [management orga- nization] was unrealistic-ally complex for a fairly simple and technically straight-forward project. It suggests that project management should have been structured like the relatively successful Smallholder Tea Project (Cr. 109-UG) that was approved in 1967 and completed in 1972 (PCR, para. 7.04). Prior to the project, tobacco production had been primarily under the aegis of a private tobacco company, BAT.2! Beginning in the latter 1940s and early 1950s, this company had experimented with several types of production and processing organizations for flue-cured tobacco in the West Nile and Middle North. First, the company provided farmers with seedlings and bought the uncured (green) tobacco from them and cured it in its own barns. Later, BAT changed to a system in which farmers had their own curing barns and sold cured tobacco to the company. The change to curing by farmers came when tests showed that curing in small barns was feasible and that a higher quality of tobacco was produced. The new organization was designated the Master Farmer Scheme. The success of this scheme was achieved through i) company financed or guaranteed loans for barn construction and tobacco production and ii) company provision of strict supervision and extension of adequate technical 1/ This section, somewhat in the vein of the PCR, elaborates on the possi- bility that project preparation was inadequate and, if management design had been different, the project's impact would have been greater even under the trying circumstances during which it was implemented. 2/ British American Tobacco Company. - 5 - advice to farmers. Production under this scheme progressed more satisfac- torily in the West Nile than Middle North. BAT then decided to transfer its main activities to the West Nile, but agreed to continue processing cured leaf in the Middle North area. It also encouraged farmers in the area to form cooperatives to further tobacco production, which they did. In 1966, the Government initiated an expansion program covering existing and additional areas in the Middle North. At the same time, it encouraged BAT to get out of tobacco production and pushed for a transfer of this production to the cooperative program. 17. At the time of project preparation, it was the government policy not to establish additional parastatal institutions for the development of individual crops when existing institutions, in particular the cooperative movement, could be employed if suitably strengthened. The project organiza- tion as appraised, comprised five elementsl/: (a) a project unit within the Ministry of Marketing and Cooperatives headed by a project director, and a project committee responsible for determining project policies and for coordinating project activities; (b) three cooperative unions, one in each project area (serving 53 primary cooperative societies), responsible for carrying out the project in their respective areas; (c) the Uganda Commercial Bank (UCB), acting as a Government disburse- ment agent for loans to the three unions; (d) PMB, responsible for establishing additional tobacco storage and for tobacco marketing; and (e) extension service personnel of the Departments of Agriculture, Forestry, and Cooperative Development, responsible for providing technical assistance and advice to tobacco growers who would be members of the project unit and who would work under the direction of the project director and union managers. 18. A project committee, established by ministerial executive order, was responsible for overall direction of the project and coordination of project entities. It was made up of senior representatives of the Departments of Agriculture, Forestry and Cooperative Development, the Produce Marketing Board, and the Uganda Commercial Bank. The project director, attached to the Ministry of Marketing and Cooperatives, was responsible for the overall execu- tion of the project. The day-to-day execution of the project rested with the three cooperative tobacco unions and their associated cooperative societies. 1/ Staff Appraisal Report, para. 6.02. - 6 - 19. The leadership of the project committee passed from the Ministry of Marketing and Cooperatives to the Ministry of Agriculture, Forestry and Cooperatives in 1971 following the military coup d'etat. There is no evidence that the project committee had any significant influence on the project at any time during implementation. The audit agrees with the PCR's assessment that the institutional design was unrealistically complex (PCR, aras. 6.01-6.07) for a fairly simple and technically straight-forward project1U. As designed, success of the project depended on too many agencies, most of which did not have a permanent vested interest in the project or, as for PMB, had other means of pursuing its interests (high margins and marketing of other crops). 20. An unanswered question is what was the origin of this, especially in retrospect, inoperative management structure? Unfortunately, because the project was prepared 13 years ago, Bank staff have little recollection of preparation and appraisal of the project. They do, however, remember that the project was hurriedly prepared to please the Government after IDA had rejected a broader credit project. Some IDA staff feel that the project also was inadequately appraised. The preparation mission had expected that BAT would continue to be involved in the project for some time to come after appraisal. It primarily had the experience and expertise to supervise the development and expansion of tobacco production in Uganda. This continued involvement did not take place. If this were to happen, it would have had to have been on an informal basis. BAT involvement is not mentioned in either the SAR or the Credit Agreement. As noted above, the Departments of Agriculture, Forestry and Cooperative Development were responsible for providing technical assis- tance and advice to tobacco growers under the appraisal design, a function that had been mainly carried out by BAT during development of the tobacco industry in Uganda prior to the project. 21. The origin of the project management organization concept remains a mystery. The audit can only conclude that staff involved in preparation and appraisal should have been wary about the likelihood that an untried, loose- knit, organizational structure could manage such a widely dispersed project. More time should have been spent at preparation and appraisal in thinking about and designing an effective management organization. Under the circum- stances, only a much smaller project would have been viable. Although such a project may not have been fully successful due to the ensuing events that tend to obscure its viability, it likely would have made more progress in the first two years of the project implementation period. B. Technical Feasibility 22. As noted above, the audit takes issue with the PCR statement that the project was "fairly simple and technically straightforward". It may be a matter of interpretation, but tobacco production is not a simple technique or production process. A large amount of technical knowledge and experience is 1/ The audit agrees with this statement except for the underlined portion, see following section. - 7 - required for growing tobacco. Tobacco seedlings are very susceptible to disease and must be carefully nurtured. The growing of quality tobacco requires good fertilization and the prevention of attacks by diseases and pests. Finally, the flue-curing of tobacco requires constant, expert atten- tion.1/ 23. As already noted above, BAT contributed to and supervised many of the production activities in the early development of tobacco in Uganda. During the 1950s and 1960s BAT provided the seedlings to farmers, supervised the field production process, and bought "green" tobacco from farmers. The company cured the tobacco, graded it and processed it. 24. Under the IDA Credit, 40% of the increase in tobacco plantings was expected to occur in the Middle North where BAT had been unsuccessful. While some progress was made early in the project in the West Nile, none was made in the Middle North, primarily due to ineffective cooperative union managers, and unexperienced and unmotivated extension staff. Conversely, the Middle North benefitted from having the Deputy Project Director, a tobacco production expert, located there. 25. In retrospect, the expansion program proposed for the Middle North was unrealistic given the past history of developments in the area and the knowledge and experience required by farmers for tobacco production. 26. Further, a recent Bank reportZ2 questions whether Uganda has sufficient natural advantages to re-enter the world tobacco market on a large scale. Bank staff also feel that a thorough study is needed to determine Uganda's competitive position in tobacco production, both in foreign markets and locally. C. Tobacco Prices 27. Another factor contributing to the lack of progress in the Middle North as well as in the West Nile and Kigezi areas was the lack of adequate price incentives. At the time of preparation, it was realized that the prices to be paid to farmers under the project would be lower than they had been under the BAT regime which sold its tobacco on the domestic market, where prices were fixed above world market levels. At preparation, it was antici- pated that for the output of the project, most of which would be exported, farmers would receive a price 25% below the price they had been receiving for tobacco they had been supplying to BAT. This lower price incentive was 1/ See for example, T. B. Hutcheson, T. K. Wolfe and M. S. Kipps, The Production of Field Crops. McGraw-Hill, New York, 1948. 2/ Uganda - Country Economic Memorandum, Report No. 3773-UG, dated March 31, 1982. - 8 - compounded during project implementation by the fact that the PMB, in the face of rising export prices, extracted higher and unjustifiable operating expenses and gross margins from its tobacco sales. IDA supervision missions repeatedly discussed this anomalous pricing policy with the Government without effect. 28. From the beginning, it would have been desirable for IDA to have had some agreement with the Government on prices paid to farmers for tobacco that was to be followed. However, when the project was prepared and appraised, little attention was being given to prices received by farmers by Bank/IDA and instead of providing price incentives, disincentives were readily accepted by the Bank. -9- UGANDA SMALLHOLDER TOBACCO PROJECT Credit 212-UG Project Completion Report Eastern Africa Projects Department June 18, 1982 Northern Agriculture Division  - 11 - UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report I. Background 1.01 After the mid-1960's, the agriculture of Uganda entered a period of slow growth, with agricultural GDP increasing at a slower rate than the 3% population increase, and sluggish agricultural export earnings. This prompted the Government to assign priority to investments in commercial agriculture - e.g. tea, beef and tobacco - which seemed amenable to quick modernization and capable of diversifying and increasing agricultural exports. 1.02 Uganda's major agricultural exports, coffee and cotton which jointly accounted for 88% of total exports in 1968, had suffered setbacks since 1964 due to low and unstable prices. Tea, the third major agricultural export was slowly growing in importance and benefitted from a 1967 IDA credit for expansion through a smallholder project. Since tobacco production had become well established in Uganda, but provided only 2% of total export value, it appeared logical to aim at expanding its output for export and thereby diversifying the country's agricultural exports. The international market for flue-cured tobacco appeared favorable in the late 1960's and Ugandan growers had shown responsiveness to technical innovation and market demand. Moreover, since flue-cured tobacco is a high value, low risk, and labor intensive crop, requiring little capital, it suited the conditions of Uganda's small scale farmers much more than either a project covering several crops requiring multiple technical "packages," or one with a crop such as sugarcane which would expose smallholders to high risks, entail a high degree of integration with processing and require relatively large scale and tightly-managed operations. 1.03 Tobacco has been and remains a relatively minor crop in Uganda, although commercial production has a history dating back to 1928 and the 1940's when flue-cured tobacco production became well established. After coffee, cotton and tea, tobacco was alleged to be Uganda's most important cash crop for some 15-20,000 smallholders, although in 1969 it probably accounted for less than 1% of the total value of crops produced in the country. Among the main cash crops in 1969, seed cotton represented about 20% of total value crop production, coffee 13%, tea 8%, sugar 7% and sesame seed 2%. Nevertheless, commercial production of flue-cured tobacco, stimulated by the British American Tobacco Co. (BAT) had made progress in the 1950's among smallholders in West Nile and Middle North districts, and after 1960 in Kigazi district. 1.04 Progress in tobacco production during the 1940's and 1950's, however, was not free of the growing pains involved in introducing a new crop. Much trial and error was required by the BAT and the farmers themselves in order to organize production, curing and marketing in financially attractive ways for both BAT and the farmers, and to make the - 12 - final product competitive with flue-cured tobaccos from Rhodesia and the USA. Although tobacco production in each of the three districts evolved differently, i.e. according to the responsiveness of farmers and the needs of BAT, it began initially with farmers producing green leaf from plots averaging one-half hectare, and selling it to BAT for curing in large central barns. Since leaf grown over large areas varied widely in quality, curing large quantities in central barns became uneconomical and was subsequently shifted to small individually or communally owned barns. Under this system after 1960, BAT financed directly or guaranteed loans for barn construction according to its specifications, provided close supervision and technical assistance and purchased the cured leaf. The emergence of primary cooperative societies and cooperative unions in the three districts in the 1960's facilitated the provision of inputs, including fuelwood for curing, and marketing of cured leaf, thereby achieving a degree of market integration that BAT had not previously successfully achieved. 1.05 Total tobacco production in Uganda amounted to almost 2,900 tons in 1959, increased to 3,500 tons in 1969, reached a peak of 5,000 tons in 1972, and declined precipitously thereafter to 1,392 tons in 1979. Flue-cured tobacco, however, began to assume importance in the late 1950's, as its production rose from an annual average of 600,000 kg in 1959-60 to over 2,330 tons during 1967-68, and eventually attained an average of 3,000 tons during 1971-72. Thus, the proportion of flue-cured tobacco rose from about 22 to 54% of total tobacco production over an eight year period, and eventually almost completely displaced fire-cured tobacco by 1979. 1.06 The increase in flue-cured tobacco production and the multiplication of growers' cooperatives were closely related events with the Government's interest and involvement in tobacco production and marketing in 1966. Tobacco growers' cooperatives and cooperative unions spread rapidly after 1966 apparently with prompting of Government, which vigorously encouraged cooperative activity in agriculture. The Government's involvement apparently began in 1966 when an international tobacco manufacturer - Rothmans of Pall Mall Ltd. - offered (or agreed) to purchase high quality Ugandan flue-cured tobacco from middle North District. These would be the first significant Ugandan exports of tobacco, except for small quantities previously marketed in Kenya. In response to the opportunity, the Government encouraged expansion of tobacco acreage and launched a drive to increase the number of cooperative societies first in middle-North, and by 1968, to other districts. 1.07 In 1968-69, the Government, through the Department of Agriculture, also gradually began to assume supervision and agricultural extension responsibilities for tobacco, which had previously been borne by BAT. BAT (for reasons which are more open to conjecture than supported by facts) was curtailing its tobacco field operations and, with its cured leaf supply assured, was concentrating on producing cigarettes only for the domestic market. It can also be inferred from the Staff Appraisal Report (SAR) that, having established a sure supply of tobacco for its own manufacturing for the domestic market, BAT's lack of interest in tobacco exports was inconsistent, or even clashed with inter alia, the export promotion policy of Government. It is a fact, moreover, that an arrangement (formal or tacit) existed in 1969 whereby, according to the - 13 - SAR, BAT "would continue to own and operate its own leaf handling plants and cigarette factories, and the Government would be responsible for production and export marketing." The acquisition of BAT facilities in Middle North and West Nile districts by the respective Cooperative Unions in 1970, which was a condition of effectiveness of IDA, was among the last steps toward the eventual withdrawal of BAT from Uganda. 1.08 When IDA agreed to consider financing a smallholder tobacco improvement project in 1969, it had approved five credits to Uganda, among which two had been for agriculture, i.e. a smallholder tea project (Credit 109-UG) in 1967 and a beef ranching project (Credit 130-UG) in 1968, one for education and two for roads. In addition, the Bank had made our loan to Uganda for power development, which was guaranteed by the United Kingdom and four loans for common services in the East African Community. The project has since been followed by only one credit, for education in 1971, after which no further credits were approved until 1980. 1.09 The IDA credit of US$4 million, approved in July 1970, was intended to finance 55% of total project costs (US$7.3 million) , to cover US$2.6 million foreign exchange and US$1.4 million of local currency costs. Government was expected to provide 16% of project costs (US$1.17 million equivalent) and farmers and cooperatives would finance 29% (US$2.17 million) from their own funds. 1.10 This PCR was prepared exclusively by Bank staff stationed in Nairobi. It is based on a detailed review of project files and recent reports of the UNTC. No field visit to Uganda was deemed necessary in view of the long period since project completion in 1976 and the economic dislocations that have affected the country, including tobacco producing areas, during the last few years. Bank staff deems it sufficient to assess the status of tobacco production and marketing from NTC reports and participation in a recent sector mission by agricultural staff stationed in Nairobi. II. Identification, Preparation and Appraisal Identification and Preparation 2.01 Before IDA was asked by the Government of Uganda to finance the Smallholder Tobacco Project, the Government had requested IDA assistance in financing a general agricultural credit project. Such a project was appraised by IDA in 1968, but was found insufficiently prepared. IDA was skeptical that such a project could be prepared because it would have entailed a long period of preparation during which a careful assessment of technical feasibility of a diverse range of crops and of the agricultural credit systems would have to be made. Consequently, the Government proposed and IDA agreed to concentrate on a promising single crop such as flue-cured tobacco, since considerable progress had been made in producing it, at least in West Nile district. This was fully in accordance with Government policies and priorities, as indicated in the previous section. Moreover, sufficient progress was being made under the 1967 Smallholder Tea Project (Credit 109-UG) , under the guidance of a parastatal organization, to suggest that a somewhat similar project could be formulated for flue-cured tobacco. - 14 - 2.02 The project was prepared by the Government with assistance from the Bank's Permanent Mission in Eastern Africa. No documentation exists in RMEA to determine the respective roles and views of the Bank and the Government of Uganda in project design and preparation. It appears from the SAR, however, that there was a consensus regarding project design since the project coincided with the withdrawal of BAT from flue-cured tobacco production and the Government's growing involvement in all aspects of production, marketing and exports. The Bank, during either preparation or appraisal, seems to have considerably influenced project design by insisting on: (a) carefully circumscribing the amount and terms of agricultural credit to tobacco growers and cooperative unions; (b) establishing a Tobacco Capital Account; (c) reserving for IDA the unusual prerogative of final approval over changes in operating charges by cooperatives, export marketing arrangements and requiring payment of uniform tobacco prices to farmers in the three districts; (d) consolidation of the finances of cooperative unions and local growers' cooperative societies; and (e) purchasing certain BAT assets needed for the project. 2.03 Since the project contained no complex engineering for civil works or equipment, and was merely an expansion of a promising activity conducted over twenty years by individual farmers and BAT, its preparation was a simple and straightforward task that could be done with the assistance of Bank staff. Nevertheless, it is significant to the state of project preparation in 1969 and the conjectural - if not speculative - nature of appraisal at that time that (a) the formulation of a detailed Plan of Operation for project implementation was a condition of credit effectiveness; and (b) the Plan of Operation, when issued, was subsequently amended on two occasions. The 1971 Plan of Operations stated in considerable detail the quantitative objectives and features of the project, thereby amending similar data that had been the basis of the SAR. For example, the tobacco acreage target during project implementation was reduced from 11,800 acres in the SAR to 10,654 acres in the 1971 Plan of Operations; the figures for wood-fuel plantations of the SAR and the 1971 Plan of Operations were not compatible; the tobacco-curing barn building program likewise was significantly different from that of the SAR. - 15 - Difference between SAR and 1971 Plan of Operation SAR P Oper Tobacco Planting (acres) Total 11,800 10,654 New 4,708 3,302 Tobacco Barns Proposed (No) Existing 6,199 5,729 Conversion 1,600 831 New 4,711 2,922 Eucalyptus Woodlots (acres) Existing 10,366 11,030 New 7,968 15,146 2.04 The deviation of the Plan of Operations from the SAR, scarcely more than one year after appraisal of an ongoing crop production and marketing program, leads to the inference that a well prepared assessment of investment needs and thorough feasibility study were not available to the appraisal mission. Had the project been well prepared, the Plan of Operation formulated as a condition of effectiveness would have been unnecessary. A work program, however, would have been required in its place. Project Objectives and Description 2.05 The project's main objectives were to raise annual flue-cured tobacco production from about 2.7 million kg to 5.3 million kg. Flue-cured tobacco exports were projected to increase from 700-800,000 kg annually to 3.1 million kg. These objectives were to be reached by increasing the number of flue-cured tobacco producers by some 6,000 farmers in the three districts, expanding the area planted and simultaneously raising yields, as summarized below. Before Project With Project 1/ District Area Yields Production Area 2/ Yields Production (acres) (kg/acre) ('000 kg) (acres) (kg/acre) ('000 kg) West Nile 3,630 590 2,238 5,000 450 2,250 Middle N 2,978 191 589 6,000 450 2,700 Kigezi 484 199 128 800 450 360 7,092 416 3/ 2,955 11,800 450 5,310 1/ SAR. 2/ At full development in 1977-78. 3/ Weighted average. - 16 - 2.06 The project objectives appeared simple and straightforward and their attainment essentially entailed adding physical infrastructure, inputs and credit, technical assistance and improving organizational efficiency, to a process that had functioned with some success by trial and error under the guidance of BAT and the producers' organizations. No drastic increases in output or yields were forecast in West Nile, which already was the major tobacco producer, with 50% of Uganda's area and about 75% of the flue-cured tobacco output. In contrast, Middle North and Kigezi were expected to nearly double their combined tobacco area from 3,500 to 6,800 acres, and their joint output was expected to increase fourfold from 700,000 kg to 3 million kg. For these objectives, the project included the following components and activities: (a) Credit to tobacco growers for (i) construction of 4,711 new tobacco curing barns and, in West Nile, conversion of 1,600 barns from four to six-tier; (ii) financing the first year costs of production inputs of new tobacco producers, up to a total of about U Shs. 2.8 million. Existing producers would continue to receive credit, through their cooperatives, from commercial banks and the Uganda Commercial Bank (UCB); (b) Credit to the three Cooperative Unions for (i) construction of baling centers needed by 38 local cooperative societies (5 in West Nile, 33 in Middle North and Kigezi) ; (ii) building and equipping supply stores and offices for the Middle North and Kigezi unions; and (iii) planting about 8,000 acres of eucalyptus to satisfy part of the fuelwood needs for tobacco curing; (c) Building and equipping two tobacco stores in Kampala for the Produce Marketing Board (PMB) , thus enabling it to handle marketing of 5.5 million kg of leaf annually; (d) Financing the establishment (offices, equipment, housing, vehicles etc.) and operating costs of central and district project administration under the Ministry of Marketing and Cooperatives; (e) Technical assistance for studies and preparation of new agricultural development projects. 2.07 Total project costs were estimated at U Shs. 50.9 milion, equivalent to US$7.27 million, costs were allocated as follows: - 17 - Project Costs US$ Equivalent (million U Sh.) ('000) On-farm investment 21.4 3,063.0 Cooperatives (Unions) 13.6 1,939.3 Investment (8.3) Operations (5.3) Tobacco Storage (PMB) 4.4 628.6 Project Management 11.0 1,571.9 Investment (2.6) Operations (8.4) Technical Assistance 0.5 71.4 Total 50.9 7,274.2 2.08 An IDA credit of US$4 million was approved by the Board on July 21, 1970 to finance 55% of project costs, including 100% of foreign exchange costs (US$2.6 million) and 30% of local expenditures (US$1.4 million equivalent). III. Project Implementation 3.01 If project implementation is assessed according to the comprehensive and strict criteria of project performance presently employed by IDA (and the Bank) , the project was handicapped from the outset by delayed effectiveness, and weak and fragmented management at its apex and at the level of cooperative unions and societies. Subsequently, after the 1971 military overthrow of the Government, political and economic conditions deteriorated, initially with disruptive effects on staff morale, security and farmers' incentives in Middle North, which subsequently spread to the other two districts. Fear of reprisals, e.g. arrest, imprisonment and even threats of death, intimidated both project staff and farmers. Hence what began as "moderate" start-up difficulties in 1970-71 became major problems throughout the project's implementation and the political upheaval from 1972 to project completion in 1976. 3.02 Since tobacco production was well established in two of the three producing districts and no unusual conditions of credit effectiveness were imposed, it was expected that the project would get a quick start. This would have enabled project management to focus its efforts during the early period on strengthening the cooperative societies and unions and working out administrative and financial procedures with Government entities that shared responsibility for project implementation. - 182 Credit Effectiveness 3.03 Eight conditions of credit effectivness were set in the Credit Agreement. October 31, 1970 was set as the date for fulfilling these conditions, or slightly more than three months after signing the Credit Agreement and seven months after negotiations. The conditions were: (a) Establishment of the Project Committee, consisting of senior representatives of the Ministries of Agriculture and Forestry, Marketing and Cooperatives, the PMB, UCB and each Cooperative Union; (b) Appointment of a Project Director; (c) Preparation and agreement with IDA of a Plan of Operations; (d) Statutory measures under Uganda's External Loans Act of 1962 to allow relending of IDA funds to the cooperative unions; (e) Execution of subsidiary loan agreements with cooperative unions and the PMB; (f) Agreement between the Government and commercial banks that the latter would continue lending for seasonal production to tobacco producers; (g) Agreement with UCB that it would act as the Government's disbursing agent for loans to cooperative unions and the revolving fund for new seasonal credit; (h) Purchase of BAT facilities by the West Nile and Middle North Cooperative Unions in their respective regions 3.04 The above conditions of effectiveness were finally satisfied only after four extensions, or 270 days after the date originally set. As shown in Annex I, after the first extension in December 1970, all but two of the conditions were fulfilled. 3.05 What was conceived as a simple matter of executing a series of ten standard subsidiary loan agreements (three with each cooperative union and one with the PMB) under Section 4.06 of the Credit Agreement, and the usual legal opinions and ratifications that these entailed, became the reason for three of the four extensions of the terminal date of credit effectiveness. It appears that during negotiations, in March 1970, no issues arose from the text of the draft Credit Agreement concerning relending the proceeds of the credit to individual tobacco producers. Although it was clearly the intent of IDA that part of the credit (under Section 4.06) would be relent directly to growers through the union, the statutory constraints in Uganda and the ensuing legal ramifications seem to have been overlooked during negotiations, since there is no mention of SLAs in the record of negotiations. - 19 - 3.06 Since the issue first arose before expiration of the original terminal date for effectiveness, it appears, in retrospect, that there was sufficient ground for recovening the negotiating parties in order to find a way out, in the context of the already signed Credit Agreement. Instead, much time was spent on correspondence between IDA and the Borrower about whether the proceeds of the credit could be relent by the unions directly to tobacco producers, or whether the credits ought to be made to the 53 cooperative societies to which the producers belonged. In the opinion of the Solicitor General of Uganda, sent to the Bank one month before expiration of the first terminal date, the latter should be the case. But since IDA considered the individual cooperative societies too weak organizationally and financially, its requirement that the unions relend directly to farmers for tobacco barns and seasonal credit was deliberate. Nevertheless, IDA (reluctantly) suggested, (after the second extension) and the Government eventually agreed, that the unions could relend first to their member societies which in turn would relend to their individual members. Although this was in compliance with the Credit Agreement, the introduction of an additional party in the lending process entailed revision of the SLAs. Upon expiration of the third extension for effectiveness, the SLAs had not yet been revised to the satisfaction of IDA, therefore, a fourth extension was granted, to August 2, 1971. Meanwhile, the Government of Uganda was overthrown in Feb. 1971 and power was assumed by the military forces. The new Government merged the Ministry of Marketing and Cooperatives with the Ministry of Agriculture, abrogated existing cooperative legislation and regulations and enacted new ones, which voided SLAs already drafted. These new ones were, once again, found legally unacceptable by IDA. 3.07 Finally in early June 1971, the Government sent a mission to Washington consisting of the Project Director and a legal advisor to discuss these matters with IDA. During the meetings, agreement was reached on the SLAs, the legal opinion and the text of a letter to IDA from the Minister of Agriculture and the Registrar of Cooperatives. Upon receipt of the agreed documents and other evidence, on July 28, 1971, the Credit was declared effective. Project Start-up 3.08 Starting the project depended on concerted action to activate the five principal entities which comprised the project organization: (i) the project unit, headed by a Project Director, in the Ministry of Marketing and Cooperatives; (ii) the district Cooperative Unions and their constituent local societies; (iii) the Produce Marketing Board (PMB) responsible for tobacco storage and marketing; (iv) the Uganda Commercial Bank, as financial agent of the Government and source of credit for cooperatives along with commercial banks; and (v) the Ministry of Agriculture, Forestry and Cooperative Development, which was to assign the required technical field staff and equipment, i.e., the vital limbs of project field operations to serve the cooperative unions. 3.09 The Project Manager was appointed well in advance of the original date of credit effectiveness and the Project Committee was established shortly thereafter. The August 1970 supervision mission reported "no problems" foreseen in fulfilling "particular covenants" of the Credit - 20 - Agreement, and expressed confidence that qualified Deputy Project Director (expatriate) had been selected, while little action had been taken to recruit Cooperative Union Managers, also expatriates. The mission also outlined the required details of the Plan of Operation, which was a condition of effectiveness. It stressed that since the Plan of Operation was a blueprint for project implementation, it should have the formal endorsement of "other Ministries whose Departments will be committed to provide -- staff and/or services to the project." 3.10 The Plan of Operation was submitted to IDA in January 1971, and was approved by IDA, without concern over the differences between it and appraisal objectives reported in paras. 2.03-2.04. However, IDA reiterated to the Government that, although the Plan of Operation had been approved by the Project Committee, it lacked the endorsement of the three key Ministries (Agriculture, Finance and Marketing and Cooperatives). 3.11 In 1971, the first full year of project implementation, the first signs of breakdown of public administration in Uganda began to show. Difficulties began to set in regarding the authority of project management over matters of staffing, salaries, procurement and the work program. Insufficient control of Project management over technical staff seconded from various Departments, mainly the Ministry of Agriculture, particularly affected performance of the following technical services: (a) Tractor hire services for land preparation and transport; (b) Transport of fuelwood to assure a timely and sufficient quantity of wood for leaf curing; (c) Employment of a Chief Technical Officer to direct (i) the soil survey needed before expanding tobacco acreage in Kigezi and (ii) designing practical ways to solve the fuel shortage and improve furnace design for tobacco curing barns; (d) Posting of Forest Department technician to accelerate fuelwood plantations; (e) Use of experienced BAT field staff who, although they lacked the formal education qualifications of the civil service, they knew tobacco production, leaf curing and marketing; (f) Employment of qualified accountants for the Cooperative Unions, especially those in West Nile and Middle North, where accounting deficiencies and arrears were carried over from previous years. This would have required at least doubling the salary level of accountants to make them competitive with the private sector. 3.12 It was also becoming evident that the role and performance of the PMB, UCB and Central Tenders Board, acting independently, was not conducive to a coherent project organization. The PMB services were deemed unsatisfactory. It extracted high margins from its purchases relative to its poor fuelwood and tobacco transportation services, and was late in paying farmers for their tobacco. Already it appeared that the tobacco price structure unduly favored PMB, and should be reviewed in light of - 21 - incentives needed by farmers. The Central Tenders Board was not functioning and was a serious source of delays in the award of contracts for procuring goods and services. It particularly affected the purchase of vehicles and building materials at the outset. The financial management of becoming unwieldy with responsibility split between UCB and Project management. The UCB, which received a 2% charge against all project transactions, in addition to interest on loans to cooperatives, was not equipped to provide the full range of financial services needed by the Project, including supervision of all accounts and preparing requests for disbursements to IDA. Financial management responsibility was transferred to Project management in July 1971, but the difficulty in hiring qualified accounting staff remained. 3.13 The preceding problems were symptomatic of political interference with project implementation and of the fragmentation of responsibilities for which the existence of a Project Committee could not compensate. It affected the 1971 total tobacco production which was about the same as 1969, in spite of an increase of about 1,000 acres. The immediate causes were enumerated in para. 3.11, plus the demoralization of field staff due to lack of facilities and transport. A first symptom of the latter was the resignation in August 1971, of the (expatriate) West Nile Union Manager over his allegations that the Government withdrew the originally agreed terms and conditions of employment. Irrespective of the claim made, and the explicit IDA approval of Government action, the lack of competent managers secure in their posts and with sufficient autonomy to administer the project, remained one of the project's main liabilities from 1971 onwards. 3.14 One event of passing significance is that leadership of the Project Committee had passed from the Ministry of Marketing and Cooperatives to the Ministry of Agriculture, Forestry and Cooperative Development, by virtue of the merger of the two Ministries shortly after military coup d'etat. Although there is no evidence that the Project Committee was a useful means for decision-making before the merger of the two Ministries, de facto it ceased to exist after the merger. Therefore, there never really was an opportunity to test the proposition that the Project Committee could be useful. 3.15 By the end of 1971, project start appeared to have been all but aborted, in spite of uncompromising forebearance by IDA and its unobtrusive efforts to support and strengthen Project Management. Procurement continued to be held up by the Central Tenders Board, no disburement applications had been submitted to IDA, and accounts and audits were so unsatisfactory that IDA requested that the auditors be replaced. Because of import restrictions imposed by the military Government to ease foreign exchange scarcity, shortage of "strategic" materials such as cement was affecting the building program in Middle North and Kigezi, where such facilities were most needed. 3.16 The findings of a December 1971 supervision mission clearly reflected the start-up difficulties as they affected physical implementation of the project as follows: (a) Progress was deemed satisfactory in West Nile with respect to both tobacco acreage and production, fuelwood plantations, and construction of barns and baling centers. The farmers' resourcefulness and cooperatives' long experience with production of flue-cured tobacco and strong cooperative union management during 1971 enabled the District to overcome adverse circumstances i.e. unreliable AMD tractor services, shortage of fuelwood due to poor transport services of PMB and lack of vehicles for extension work. (b) In Middle North, where most of the expansion in tobacco acreage, fuelwood planting and barn construction was expected, little had been achieved. Area planted in 1971 was 800 acres less than in 1970, due largely to inefficient tractor services, ineffective cooperative Union Managers, and ill-equipped and unmotivated extension staff. More ominous, however, were the political disturbances in the area, with reports of farmers leaving their land. (c) In Kigezi, where tobacco acreage and production were not expected to increase substantially until after soil studies were done, the 1971 results were on par with modest expectations. Prospects for achieving project objectives were dim, however, due to delays in the construction of barns and baling centers and shortfalls in fuelwood plantings. 3.17 In response to the mildly critical letter containing the highlights and recommendations of the December 1971 supervision mission, the Ministry of Finance acknowledged the disappointing performance in Middle North and the delays in the other districts. In its reply to IDA, the Government cited extenuating circumstances rather than project management and the high degree of Central Goverment control, as major causes. IDA was assured that these extenuating circumstances, i.e. drought, terms of appointment of Union Manager, lack of vehicles, .unreliable tractor services - no longer prevailed or had been overcome by better planning and coordination. More funds were promised to step up fuelwood plantings in West Nile. IDA was assured that the Central Tenders Board was fully constituted and ready to consider the award of bids, and that (30) vehicles had been ordered. On the question of the audit of project accounts and those of the Cooperative Unions, IDA was reassured that the Auditor General was expected to audit Project accounts; the 1970 accounts of West Nile's Cooperative Unions had been audited, and those of Kigezi and Middle North would be done by new private auditors. The Ministry of Finance confirmed that import restrictions had been imposed with a 100% advance deposit requirement (to enforce quotas) , but did not respond to IDA's suggestion that for Project goods, pre-payment requirements should not exceed 10% of the import value. 3.18 Assurances of the Ministry of Finance in February 1972 did not result in improved project implementation throughout 1972. Two IDA supervision missions, conducted at close intervals in April and June 1972, reported further setbacks, leading to the conclusion that the project was no longer economically viable. The May mission judged the situation to be sufficiently serious to recommend suspension of IDA disbursements. - 23 - 3.19 Development in Middle North was virtually stalled for the reasons mentioned in para. 3.16, in addition to insolvency of the Cooperative Union due to cumulative arrears of U.Shs. 4 million in loan repayments by the local societies for debts incurred since 1968; in West Nile arrears amounted to U.Shs. 1.7 million. Moreover, total project costs were estimated to have increased 23-25% above approval estimates. High overhead costs (because of over-staffing) of the PMB made it difficult to raise tobacco prices paid to farmers. In current terms, farmers received 85% of the 1968 price for their tobacco, even though wholesale and export prices had increased 13% and 15% respectively since 1968 (Annex 8). 3.20 There was also sufficient evidence that the Plan of Operation was no longer a realistic guide for physical implementation, since it had become obvious that the Middle North targets could not be achieved, and the overall investment program and schedule had been seriously disrupted. Consequently, during the June 1972 supervision mission, the project management and IDA agreed that a new Plan of Operation should be prepared, envisaging the following: (a) expansion of West Nile tobacco area by 1,000 acres, to compensate for the shortfall in Middle North; (b) extending the expansion of tobacco production contemplated in the implementation schedule beyond 1974, in order to enable Middle North to plant 2,368 acres. The foregoing was proposed and eventually agreed with a view to restoring economic viability to the Project from a projected -4.8% IRR with the prevailing state of affairs, to +14% if corrective measures were taken. Under the proposed revision, the total additional tobacco area under the project would have been 6,177 acres vs. 3,302 in the original Plan of Operation (assuming, of course, no reduction in the original area) . It is incomprehensible how, in light of previous and existing difficulties in 1972, the obsession to achieve a desired rate of return could lead Project management and IDA to agree rationally on such unrealistically high objectives. 3.21 A final setback in August 1972 was the expulsion of virtually all Asians from Uganda, which included many suppliers, builders, architects etc. under contract, and some staff employed by the Project (Chief Accountant and Cooperative Union staff) . This meant that inexperienced staff had to be hired; nearly all construction came to a halt, and alternative sources of input and material supplies had to be found. In Middle North and Kigezi, Project and Cooperative Union staff and farmers were reported to "have either fled or been killed" while fear for their personal safety intimidated the remaining staff as well as farmers. Revised Plan of Operation 3.22 Because of a lack of staff, more time was needed to prepare a new Plan of Operation than was anticipated in June-July, 1972. A new draft was personally prepared by the Project Manager and his Deputy and informally submitted to IDA in April 1973. It included an increase of 1,000 acres of tobacco in West Nile and of 2,369 acres in Middle North. In West Nile the - 24 - Plan also proposed (i) the provision of 100% of West Nile fuelwood requirements (vs 60%) by increasing the planned eucalyptus woodlots from about 6,500 to 7,800 acres; (ii) building an additional 326 tobacco curing barns; and (iii) forming two new cooperative societies. In Middle North more drastic alterations to the original Plan of Operation were proposed to justify and support a 2,369 acre expansion in tobacco acreage, as follows: (a) 4,760 increase in eucalyptus woodlots; (b) about 2,200 additional curing barns; (c) construction of 3 additional baling centers in expansion areas; (d) increasing field staff (tobacco supervisors) by 22; (e) purchase of additional vehicles, housing etc. 3.23 In view of the U.Shs. 15.3 million increase in project costs that the new plan entailed, in addition to the deterioration of Uganda's economy and poor fulfillment of financial obligations under the first Plan of Operation, IDA requested from the Government (i.e. Ministry of Finance, Planning and Development) a commitment to provide the additional funds needed. A hiatus of several months ensued during which no action was taken by the Government until the revised Plan of Operation was formally sent to IDA by the Ministry of Finance in January 1974. IDA deemed it unsatisfactory however, since it excluded the Kigezi area, lacked corresponding financial projections and a commitment by the Ministry of Finance to finance incremental costs. Consequently, on the basis of careful reviews of country and project conditions, IDA felt that the 1972 agreement on the need for a revised Plan of Operation had been "overtaken by events" and that "little purpose would be served by attempting to draw up another plan." Since a revised Plan of Operation was accompanied by a request to extend the closing date, IDA also informed the Ministry of Finance in March 1974 that it would not be necessary to extend it beyond December 1975, as originally scheduled. Physical Implementation 3.24 Start-up difficulties, inflation, shortages of materials and inputs, and the quickening breakdown of public safety were not conducive to the execution of key components and activities contemplated in the original implementation schedule. The absence of monitoring and evaluation in project operations and the lack of any comprehensive progress report until late 1974 hinders a detailed assessment of achievements except for certain key components, shown in Annex II. Moreover, Government censorship of Project management's reports from 1972 until project completion in 1976, further limited and, it appears, distorted the extent to which physical implementation was disclosed to IDA. On-farm Investment 3.25 This consisted mainly of the construction and renovation of tobacco curing barns. Only 1,802, or 38%, of the total number of new barns to be built (4,711) were reported to have been built by mid-1974. Viewed - 25 - against the 1971 Plan of Operation, however, the 1,802 new barns represent 57-62% of the target. Even though the largest number of barns were built in Middle North, the achievement was less than 50% of the target (27% of SAR) largely because of the withdrawal of farmers from tobacco production in that District. In West Nile, the achievement by 1974 can be considered satisfactory under the circumstances. In Kigezi this and other components of the project never really got under way. New Tobacco Curing Barns Built SAR Plan of Operation Built District Target Target to 1974 West Nile 233 871 587 Middle North 4,329 2,202 1,182 Kigezi 149 79 34 4,711 3,152 1,802 Seasonal Credit for New Producers 3.26 Since total tobacco acreage decreased, there is little evidence to lend credence to reports that new tobacco producers were recruited. Nevertheless, reports available for only the project's first two years indicate that U.Shs. 657,000 was provided to new producers for first year inputs, compared with U.Shs. 1.2 million anticipated in the SAR for the first two years. As indicated in para. 4.02, withdrawal of funds from a revolving fund established for this purpose became an issue between IDA and the Government. Fuelwood Plantations 3.27 In addition to increasing the number and capacity of curing barns, successful increase of flue-cured tobacco production required an expansion of eucalyptus woodlots by the Cooperative Unions to supply the fuel for curing the tobacco leaf. It was originally thought that about 8,000 acres would suffice for the three Districts, since the existing woodlots and new plantations would provide about 60% of fuelwood needs of West Nile and Kigezi, and about 40% in Middle North. The remaining fuelwood, according to the SAR, would be cut from natural timber. This seemed unrealistic to Project management, since the accessible wood supply from natural timber was dwindling and difficult to transport. Consequently, the 1971 Plan of Operation proposed expanding new eucalyptus woodlots to about 15,000 acres, vs. 8,000 acres estimated in the SAR. After three years of project implementation, 7,736 acres were reported to have been planted, which is 97% of appraisal requirements but 62% of the more realistic target set in the Plan of Operation for the same period. The fuelwood acreage reported in 1974 is, however, of dubious validity since new plantings were calculated by the Cooperative Unions on the basis of seedlings obtained from nurseries, rather than actual planting, and no account was taken of several hundred acres lost to fire and poor survival after transplanting. The plantation inventory recommended by RMEA in 1974 was never conducted. - 26 - Fuelwood Plantations SAR Plan of Operation Planted District Target Target to 1974 ---------------------Acres------------------- West Nile 1,056 6,498 3,772 Middle North 5,760 7,135 3,330 Kigezi 1,152 1,514 635 7,968 15,147 1/ 7,737 1/ Four-year program. A target of 12,512 acres was set for the comparable (1970-73) period. Cooperative Investment and Construction 3.28 Construction of 38 baling centers, housing and offices, and supply stores for the Cooperative Unions and local societies was initially delayed by the slow process of the Central Tenders Board. When work finally started in 1971, shortages of building materials, foreign exchange restrictions on imports slowed down progress. The expulsion of Asian contractors and technicians brought construction to a standstill after mid-1972, with buildings about 40% completed. Meanwhile, due to cost increases, the funds originally allocated for all construction, including Project administration buildings, had been exceeded. About U.Shs. 10 million was spent for all project buildings by early 1973, compared with U.Shs. 7.7 million budgeted in the SAR (U. Shs. 5 million for cooperatives and about U.Shs. 2.7 million for Project administration facilities. 3.29 Efforts were made to resume construction work in late 1974, with limited success due to shortages of vehicles, spare parts and difficulties in obtaining materials. Only renovation and repairs on 14 existing baling centers were done in 1974. In mid-1975, it was reported that contracts had been awarded to local builders to complete 25 baling centers, at a cost of almost U.Shs. 24 million and two zonal headquarters for U.Shs. 3 million. Since, as of July 1975, there were insufficient funds in the IDA credit to finance the reimbursable expenditures that the revised building program entailed (about U.Shs. 10.8 million) and Government lacked the necessary local funds, it appears that the original building program was not completed. Central Project Administration 3.30 In addition to the poorly documented and incomplete construction program, the lack of spare parts for project vehicles acquired under the project from 1974 immobilized most of the project staff. Tractor services from the AMD completely collapsed in 1973, with the result that farmers could not rely on AMD tractors for the timely cultivation of their fields and transporting firewood. Twenty tractors were acquired by the Project in 1974 (not financed from the IDA credit) , but these were sufficient for less than half of tobacco acreage requirements. - 27 - 3.31 While overall implementation of project components and activities was falling short of agreed objectives, two components were in fact eliminated. One, technical assistance for studies and preparation of new agricultural projects was eliminated by default on the part of Government. The other, building two new tobacco stores for PMB in Kampala was deleted when, in connection with the Plan of Operation, IDA declined to finance a more costly and politically motivated proposal to locate the store in Arua, in West Nile District, rather than in Kampala. 3.32 Concerning the -preparation of new agricultural projects, for which U.Shs. 500,000 was allocated in the project, the first steps were taken in December 1970, when the Government sent IDA a report of consultants identifying the following possible projects for further study: (a) mixed farming (beef, milk and food crops) in South Central Uganda; (b) production of flowers and ornamentals for exports; (c) integrated agricultural development in cotton producing areas; (d) improved cotton ginning, transport and marketing. No records exist to indicate that the consultants conducted a thorough identification of the possible prospects, either of their intrinsic merits or sectoral priorities. IDA reviewed the proposals and in March 1971, recommended to Government that it should focus on a single project, preferably cotton ginning, which could produce tangible results. The Government agreed but since one cotton production and marketing season was necessary to conduct cotton ginning tests and collect data on cotton transport and marketing before a feasibility study could be prepared, it decided to postpone project preparation by one year. IDA followed up on the matter in January 1972, to find out that no action had been taken. Presumably, in view of the turbulent course of events in the process of implementing more vital components than this one, the preparation of a new project was deemed inconceivable and impractical by IDA. 3.33 Central tobacco classification and storage facilities for PMB operations in Kampala, for which U.Shs. 4.4 million had been allocated in total project costs, were never constructed or equipped with project funds. The SAR envisaged two stores administered by PMB, each of 40,000 sq. ft.; these stores would have supplemented BAT facilities in Kampala which were inadequate for the volume of flue-cured tobacco forecast for 1974-75 and thereafter. Although tobacco production did not increase as projected in the SAR, PMB nevertheless proposed in 1973 the construction of a classification and storage warehouse in Arua, West Nile District, rather than in Kampala. The first indication that PMB was autonomously proceeding with this component was a letter of July 9, 1973 from the Ministry of Works and Housing to IDA seeking the Bank's concurrence with the appointment of consulting engineers. No supporting documentation was enclosed. Aware of the shortfalls in tobacco production and other difficulties facing the Project, IDA did not immediately concur and requested (i) an assessment of the need for additional tobacco storage facilities; and (ii) consultations with PMB and Project management. Following an assessment by PMB of its - 28 - needs and its proposal to erect a tobacco collection and classification store in Arua, IDA (i.e. RMEA) , in September 1973, withdrew its objection to the appointment of the consulting engineers, and requested review of final design before invitations to tender were issued. 3.34 Preliminary designs were prepared by the consultants by early November 1973, and informally submitted to and discussed with RMEA. Moreover, the consultants and the Project Manager conveyed to RMEA the Government's wish to waive ICB for the purchase of construction materials and equipment. IDA dismissed the possibility of waiving ICB and, verbally communicated this to the Project Manager and PMB. Moreover, IDA became aware that the relocation of the store from Kampala to Arua was in response to political pressure by the President of Uganda, rather than for economic reasons, and that the Project Manager had been threatened with death if the Arua store was not built. Consequently, in view of the preceding and the high costs (U.Shs. 6 million vs 4.4 million in SAR) IDA deferred approval of construction plans until (i) final design and specifications were submitted and examined by independent consultants; (ii) the revised Plan of Operation (para. 3.23) was submitted; and (iii) total project costs and commitments under the credit could be reassessed. The final designs and specifications were eventually prepared but never submitted to IDA. Moreover, in the review of project costs, it was evident to IDA that even without financing the store, remaining funds from the credit were sufficient to meet only the project's operating expenditures during the remaining two years (i.e. 1974 and 1975) . Finally, after being held in suspense for several months, the Government abandoned its effort to finance construction of the Arua store from the IDA credit, upon being informed that the Credit was overcommitted at the same time as IDA dismissed the revised Plan of Operation and the possibility of extending the closing date of the Project (para. 3.23). 3.35 The matter of the Arua store did not end, however, with the formal notification of IDA to Government in March 1974. Twenty months later, in November 1975, the consulting engineers reported to Nairobi Bank staff that they had not been paid their US$78,000 fee by the Government (i.e. PMB) for final designs, and sought payment from the IDA credit. The cost of the Arua store had escalated to U.Shs. 7 million and construction was to be financed by the Uganda Development Bank. By closing date, however, construction had not started. There is no record to indicate that the consultants were paid from IDA funds, or if they were ever paid. 3.36 Research on (a) tobacco-growing techniques and (b) finding the most economical heating system for flue curing barns, was never conducted by the Ministry of Agriculture; contrary to Section 4.08 of the Credit Agreement. Similarly, under Section 4.12 of the Credit Agreement, the suitability of soils for producing tobacco by new growers in the Kigezi area was supposed to be determined by appropriate soil tests. This was not done. Project Costs 3.37 Project costs were originally estimated at U.Sh. 50.9 million, including U.Sh. 10.5 million for farmers' contributions and contingencies of U.Sh. 2.6 million. It is now roughly estimated that costs may have - 29 - exceeded U.Sh. 78 million, even without counting the farmers' contribution and excluding U.Sh. 5 million for two components (U.Sh. 4.4 million for PMB tobacco storage and U.Sh. 500,000 for consultants) . If implemented as envisaged at appraisal, the project would have cost almost U.Sh. 85 million, or 67% more than original estimates. Original and Revised Project Cost Estimates 1974 PCR Estimate % Appraisal Estimates Amount of SAR -----------------U.Sh. million------------- On-Farm Investment 21.4 11.2 11.2 52 (Farmers' contribution) (2.8) (n.d) 2.8 2/ - Cooperative & Administrative Investment 10.9 30.2 3/ 47.9 3/ 439 Cooperative Management 5.3 1.2 1.2 23 Project Management 8.4 15.1 15.1 180 Tobacco Storage (PMB) 4/ 4.4 - - Technical Assistance 4/ .5 - - Total 50.9 57.7 78.2 154 1/ Supervision estimates to determine expenditures reimbursable by IDA. Costs to project completion (1974-76) were rough estimates which did not envisage further increases of U.Sh. 17.7 million for construction. 2/ Assuming same value as at appraisal due to lack of data. 1/ Assumes that buildings were in fact completed at the 1975 cost estimates which cannot be ascertained. 4/ Cancelled components. 3.38 The above estimates must be interpreted without presumption as to the accuracy of underlying data because no final figures exist with which to compare actual costs with appraisal estimates. This is due to several causes: (a) Poor financial records and control of the Cooperative Unions which was due in large part to the chronic shortage of competent accountants; (b) Inadequate financial statements of Project management after 1972, especially with respect to the Government contribution; (c) Lack of information on the contribution of beneficiaries, which was included in project cost estimates; - 30 - (d) Deletion of certain components and activities and use of these funds for other purposes (e.g. PMB warehouses); (e) The peculiar method of classifying IDA financing into three broad categories without respect to purpose of expenditure by project component makes it impossible to determine the corresponding local costs of an item of expenditure; (f) Cost increases for leading items, i.e. fertilizers and other farm supplies, and building materials and equipment, which were not either documented or classified during the last three years of the project. Cost increases for project management (80% over appraisal estimates) are particularly disturbing since these were overhead expenditures that should have been controllable, in view of the persistent shortages of staff and operating equipment. Project Financing and Disbursements 3.39 Project financing was based on an IDA credit of US$4 million (equivalent to U.Sh. 28 million) , government budgetary contributions of US$1.17 million (U.Sh. 8.2 million) and farmers' contribution in labor and materials equivalent to US$2.1 million (U.Sh. 14.7 million) . The IDA credit was intended to finance 55% of project costs, representing the foreign exchange component (US$2.6 millior and including 30% (US$1.4 million or U.Sh. 9.8 million) of local currency expenditures. 3.40 The final sources of project financing is difficult to discern because of the poor financial records maintained by Project management, the cooperative unions and participating institutions (i.e. PMB, Ministry of Agriculture) and incomplete financial reporting to IDA. Overall, because of cost escalations, the Government contribution exceeded the original estimate of U.Sh. 8.2 million, but the amount is unknown. Presumably it is at least the difference between the IDA credit and cash expenditures for project completion estimated in 1974, i.e. U.Sh. 29 million. 3.41 The IDA contribution of US$4 million (equivalent to U.Sh. 28 million) was originally intended to finance 100% of foreign exchange requirements (US$18.5 million) and 30% of local currency expenditures (U.Sh. 9.8 million) , as follows: - 31 - Allocation and Disbursement of IDA Credit Category a Original Purposes Allocation Disbursed -----------Us$----------- I. Tobacco storage equipment, vehicles, bulk purchased construction materials, fertilizers and insecticides, and other imported items 1,000,000 1,534,348.27 II. Foreign exchange costs of expatriate staff and of consultant services 310,000 28,218.76 III. Other expenditures on the Project 2,450,000 2,437,432.97 IV. Unallocated 240,000 - Total 4,000,000 4,000,000.00 Category III of the IDA Credit financed salaries and local expenditures for investment purposes, but throughout the disbursement period, satisfactory statements of expenditures were frequently not submitted to IDA, so that several withdrawal requests were rejected. It is not known, therefore, how much was financed by IDA for operating expenses and investment, respectfully, under this Category. 3.42 Cost escalations under Category I and the elimination of Consultants under Category II persuaded IDA to approve reallocation of the proceeds to Categories II and IV to Category I in 1975. This was consistent with IDA's wish to fully disburse the credit by the closing date -- i.e. December 31, 1975 -- while ensuring that, under the nearly unmanageable course of events, at least the vital elements of the project such as farm inputs and curing barn building materials were available. 3.43 Disbursements preceeded very slowly, with only 17% of appraisal estimate disbursed on June 30, 1972, and 54% two years later, and 59% six months before the original closing date. On the original closing date, US$2.8 million had been disbursed, but commitments were sufficient to ensure that a three month extension of the closing date would be sufficient for full disbursement. The slippage during the early years was of course due to the project's shaky start-up; thereafter, however, slow disbursement was due to the following: (a) Cumbersome procedures in Uganda, whereby several agencies and officials cleared withdrawal applications which, when these were not accepted by IDA, had to be withdrawn and/or cancelled; (b) Poor accounting practices and some financial irregularities led to frequent cases of improper or totally deficient supporting documentation, some of which were never resolved; as late as 1977 -- i.e. one year after last disbursement -- IDA was still trying to obtain appropriate invoices and evidences of shipments for the - 32 - last 23 disbursements applications (Nos. 75-97) ; in fact smooth disbursement was the exception rather than the norm throughout projects implementation; (c) Very strict control in procurement and disbursement imposed by IDA in early 1973, in order to prevent diversion to other uses of project funds, materials and equipment. Thus, requests for withdrawals exceeding the equivalent of US$50,000 individually, and US$100,000 in the aggregate during the three-month period, required clearance and approval in Washington. 3.44 Since there seems to have been a distrust between IDA and Ugandan officials over the use of project funds (e.g. fertilizer purchases, the Arua tobacco storage facility) procurement difficulties were therefore translated into slow disbursements. Hence, after the departure of Asian contractors and suppliers from Uganda and the imposition of imports and foreign exchange restrictions, disbursements under Category I were changed from reimbursement to direct payment of foreign suppliers, by mutual agreement between IDA and Project management. While this enabled IDA to maintain some integrity in the procurement process, the new disbursement procedure was either not fully understood or reluctantly accepted by Ministry of Finance officials. 3.45 Until August 1975, four months before the agreed closing date, IDA felt that the credit could be disbursed by December 31,1975. The optimism was guarded, however, since it was subject to the finalization of procurement under Categories I and III expenditures. When this did not materialize on January 9, 1976, IDA agreed to the Government's request to extend the closing date to March 31, 1976. The final disbursement was made on April 8, 1976. Procurement 3.46 Procurement difficulties plagued the project throughout its implementation and delayed disbursements of the Credit. Initially, the slow and cumbersome process of the Central Tenders Board (CTB) delayed acquisition of vital materials, equipment, especially vehicles, and award of building contracts. Even when CTB procedures were waived and the Project Tender Committee was allowed to procure goods and services from 1973 onwards and the advisory Board of Trade granted "direct importer status" to the Project (thereby overcoming certain import licensing delays) , other problems emerged, i.e. Government did not have the funds with which to pay pending IDA reimbursements; suppliers insisted on advance payments and c.o.d.; certain inputs purchased in relatively small quantities, e.g. fertilizers, were in short supply in East Africa and could not be economically procured from distant suppliers; conflicting orders for the same supplies were occasionally submitted by different persons, in different amounts from suspect sources, pointing to irregularities, which attracted the attention of IDA officials and prompted it to scrutinize closely every purchase to minimize risks of misprocurement. In view of this, the full amount of applications for reimbursement was not always considered justified by IDA and several were disallowed, causing consternation to Project management and adding to its financial woes. - 33 - 3.47 It is evident from the records, however, that Project management did its very best under very difficult circumstances, and IDA acted swiftly whenever possible, to expedite procurement of vital items. In retrospect, however, the project would have undoubtedly benefitted by having a full-time Procurement Officer, who would have expedited matters, especially during the first two years. IV. Agricultural Impact Area Yields and Production 4.01 The simultaneous expansion of area and increase in yields on which success of the project depended never materialized during project implementation. Peak production was reached in 1972, with 3.2 million kg, or roughly 10% more than in 1969-70. This was acheved on a slightly larger area (10%) than in 1969-70. Because of the slow start of the project, there is no reason to attribute the 1972 results to the project, except perhaps, in West Nile, where some increase in area was sustained through 1973. Even by allowing for erratic yields due to weather factors, tobacco yields followed a downward trend in the three districts after 1972, including West Nile where conditions are more favorable for tobacco production. Planned and Actual Tobacco Area and Production District, Area Appraisal 1974 Actual as % and Production Target Actual of Appraisal West Nile Area (acres) 5,000 2,383 48 Production (kg) 2,250 2,106 94 Middle North Area 6,000 2,221 37 Production 2,412 379 16 Kigezi Area 800 620 78 Production 335 151 45 Totals Area 11,800 5,224 44 Production 4,997 2,636 53 4.02 It will probably never be known which factors were responsible for the precipitous decline in flue-cured tobacco production after 1972-73. Political and security risks, initially in Middle North in 1972, then in Kigezi and finally in West Nile, certainly were main deterrents for field work by project staff and intimidated cooperative managers and farmers. There were other factors, however, beginning with over optimism by IDA and Ugandan authorities at the outset of the Project that simultaneous increases in area and yields could be achieved over time while - 34 - several thousand new farmers, totally uninitiated in flue-cured tobacco production, could be recruited. There is no record to indicate the number of new farmers who were recruited, though some were, since the revolving fund for short term credit was utilized. However, instead of the U.Sh. 2.8 million required in the Credit Agreement, the revolving fund did not exceed U.Sh. 1.8 million, when the Government withdrew the funds in 1974, in violation of the Credit Agreement. The low prices paid to farmers for their tobacco with the consequential deterioration in their terms of trade during a period of rising input costs, are discussed below. Also evident causes of the decline are the poor tractor services of AMD for land preparation and hauling fuelwood, shortages of fuelwood, farmers' distrust of PMB tobacco classification and slow payment for their tobacco, delays in delivery of seed-bed material and building supplies for the curing barns, and the general breakdown of management and finance of the three Cooperative Unions. 4.03 The multitude of obstacles to increasing flue-cured tobacco production were identified by IDA missions and Project management at the earliest stages of project implementation. Since none of the earlier problems were resolved in spite of best efforts by Project management, the difficulties that arose after 1972 aggravated conditions for putting the project back "on course." The failure to increase flue-cured tobacco production was not a unique phenomenon in Uganda's agriculture. It affected all cash crops as farmers sought security in subsistence foodcrops. Producers of fire-cured tobacco suffered a similar, if not worse fate as flue-cured tobacco, since they did not benefit in any way from external financing and technical assistance. First, some fire-cured tobacco producers shifted to-the flue-cured process. Second, fire-cured tobacco area and production declined from 1.8 million kg in 1968 to 1 million kg in 1974, a 33% decline. By 1979, flue-cured tobacco production had dwindled to about 680,000 kg and fire-cured tobacco to 151,000 kg. It might be inferred, therefore, that the contribution of the project financed by IDA was to sustain flue-cured tobacco production at a level above that which would have been obtained without external financing. Tobacco Prices and Marketing 4.04 In current terms, prices paid to farmers for their flue-cured tobacco declined during the crucial years of project implementation to about 80% of the 1968 farmgate price, although both wholesale and Uganda export prices rose by about 15% at current prices (Annex VIII) . In real terms, the farmers' terms of trade and income suffered much more than is indicated by the widening margin between farmgate and export prices, which increased from about U.Sh. 6.0 in 1969 to about U.Sh. 9 in 1974 and more thereafter, while costs of inputs, materials and labor were rising in the wake of the severe inflationary shocks of 1973-74. Since the 1970-73 period was marked by world price stability for flue-cured tobacco, and a steady increase in the price paid for Uganda's output thanks to quality improvements, prices paid to farmers could have been increased at the first signs of resistance to increase output in 1972. Failure to take action in this direction, in spite of repeated urging by IDA to the Ministry of Finance may have been the single most important non-political factor to undermine the project, especially in West Nile where farmers seem predisposed to expand output. - 35 - 4.05 The rigid price structure stifled incentives for both farmers and the Cooperative Unions. For farmers, while the price received for flue-cured tobacco remained at an average of U.Sh. 4.80/kg during 1970-73 and was raised to U.Sh. 5.93/kg in 1974, cash costs of production increased from the SAR estimate of about U.Sh. 630/acre to U.Sh. 1,100-1,400/acres, plus a debt service of about U.Sh. 530/acre, with yields of 200 kg/acre in Middle North and 400 kg/acre in West Nile, net income decreased in both districts, as summarized below and in Annex VI. Appraisal Actual (October 1974) Estimate West Nile Middle North Price (U.Sh./kg) 4.73 5.5 5.5 Yield (kg/acre) 400 400 200 Gross Value of Output 1,892 2,200 1,100 Costs 636 1,432 1,142 Debt service 1/ 435 530 530 Net Income 821 238 (572) 1rTIncludes interest on seasonal loan. 4.06 Cost increases since appraisal affected mostly fertilizer and insecticides (from U.Sh. 130/acre to U.Sh. 400) and baling materials (from about U.Sh. 58 to U.Sh. 150/acre) , but other factor inputs as well. Fertilizer prices alone increased from U.Sh. 0.70/kg in 1971 to U.Sh. 4.75/kg in 1975. Consequently, with fixed prices, net farm income deteriorated to the point at which it did not even remunerate family labor (about U.Sh. 1,000/acre) , even for farmers who did not have a debt service obligation in Middle North. Hence, only well established and efficient farmers, mostly in West Nile, continued to produce tobacco. Others shifted to the production of other more remunerative crops, i.e. groundnuts, maize, which were not subject to price controls and assumed subsistence during a difficult period. 4.07 Overall, it seems that farmers tried to cope with higher input costs by reducing fertilizer applications. It was reported, for example, that the average fertilizer application in 1974 was 80 kg/acre, compared with the recommended rate of 150 kg/acre. There are allegations, however, that fertilizers purchased for the project were diverted before reaching farmers, which could also explain the lower fertilizer use. The better farmers apparently took defensive measures by improving leaf quality. This is evident from the marked increase in high quality flue-cured tobacco marketed, i.e. from 21% of production in 1969 to 56% in 1975, with a corresponding decline in lower grades, while the proportion of middle grades remained about the same (i.e. 27-28%) . In the absence of quantity expansion, however, the achievement of better grades was insufficient incentive for most farmers. 4.08 For the Cooperative Unions, the effects were catastrophic. They undertook investment and increased staff and operating costs on the premise of a larger volume of tobacco purchases and revenues from the fixed "cess" - 36 - of U.Sh. .40/kg. Proceeds of the "cess" were to be paid into a Tobacco Capital Account to service loans made to cooperatives (for buildings and fuelwood plantations) . Their operating costs were to be financed from the margin on loans made to farmers and an operating charge per kg of tobacco marketed (U.Sh. .46/kg in Kigezi and U.Sh. .26/kg in West Nile and Middle North) . Since the Cooperative Unions began the project in precarious financial conditions (i.e. excess of current liabilities over current assets due to uncollectable arrears from members) , the failure to increase volume in the early years and the decline in output after 1972, brought about their financial collapse. 4.09 Meanwhile, marketing margins of the PMB were improving, due largely to stable current world prices, although at constant prices world tobacco prices fell by 61% during 1968-73. In fact, the price received by Uganda for its flue-cured tobacco out-performed world prices (based on the unit value of India's flue-cured tobacco) even with the over-valued Ugandan currency. According to appraisal estimates, a margin of U.Sh. 3/kg would have been sufficient to cover all PMB cost for tobacco exports (fob Mombasa) , and U.Sh. 2/kg for domestic use. In fact, the PMB margin increased from an average U.Sh. 6.44/kg in 1969 to U.Sh. 8.24 in 1973 and U.Sh. 9.72 in 1974. Since PMB marketing services, especially classification and transport, deteriorated and it was chronically late in paying farmers for their tobacco, IDA repeatedly requested a reassessment of the price structure in light of its ample margins and the need to pay a higher price (at least U.Sh. 2/kg more) to farmers. This was never done during project implementation. 4.10 PMB was obviously attempting to extract as much revenue from a small amount of tobacco as it would have from the larger volume forecast at appraisal. Efforts by IDA to obtain reports of PMB's financial situation were futile. Meanwhile, tobacco exports, which had been forecast to increase from about 800,000 kg in 1968 to 3.1 million kg upon project completion, reached 1 million kg only in 1972 and 996,000 kg in 1976, and have dwindled ever since as follows: - 37 - Flue-Cured Tobacco Exports 1/ Volume Value ('000 kg) (million U.Sh.) 1968 645.3 6.5 1969 971.8 10.9 1970 807.6 8.8 1971 950.7 11.0 1972 1,001.7 11.6 1973 527.4 6.8 1974 712.5 11.0 1975 579.6 8.5 1976 996.0 17.2 1977 622.8 10.5 1978 281.4 6.0 1979 178.0 3.0 1/ As reported by the UNTC. 4.11 IDA eventually recommended that PMB responsibilities for tobacco marketing be transferred to the Uganda National Tobacco Corporation (UNTC), which succeeded BAT. UNTC assumed tobacco marketing and processing functions in 1978, two years after the closing date, at a time when flue-cured tobacco production had nearly collapsed. V. Economic Returns 5.01 It is not possible to calculate returns on a project which was not fully or satisfactorily implemented, while costs exceeded appraisal estimates and no tangible economic benefits were attained. It appears, moreover, that the economic rates of return estimated during appraisal, 17-27% was based on unqualified and over-optimistic assumptions, irrespective of the intervening political and institutional impediments to project implementation. The rapidity with which simultaneous increases in tobacco acreage and yields could be achieved was overstated, and the influence of price incentives relative to costs of producing tobacco and other crops was not considered. Hence, when upon perceiving serious problems IDA gave "second thoughts" to the project's economic feasibility in mid-1972, the economic rate of return was re-estimated to be between -4.8% to 2.4%, if the project continued as designed, or 12.5% under the best of circumstances. 5.02 In order to achieve the IRR of 12% recalculated in 1972, it would have entailed a two-year extension of project implementation, no cost increases, and an expansion of 2,400 acres in Middle North. Irrespective of political events in Middle North, this approach would have required a new implementation strategy, in which most project resources were allocated to Middle North, and only "minimum packages" to West Nile and Kigezi. Such a decision was not considered. - 38 - VI. Institutional Performance and Development A. Institutional Design 6.01 The project never coalesced into the well coordinated multi-institutional machinery intended by the appraisal mission and negotiating teams. Defective organizational design became obvious during the project's early years with the difficulties encountered in making the Ministry of Finance, PMB, the Ministry of Agriculture, the UCB and the three Cooperative Unions part of a team led by the Project Unit of the Ministry of Marketing and Cooperatives. 6.02 There is no evidence that the Project Committee ever did perform more than perfunctory functions. There are no records of its deliberations or decisions it may have made that influenced project implementation. When the Ministry of Marketing and Cooperatives were merged with the Ministry of Agriculture in 1971, the Committee might have exercised some influence and control, especially over Ministry of Agriculture services and PMB, but the opportunity to do so never arose as the Ministry of Agriculture was caught up in the political ordeal of the country at its outset. Hence, agricultural services and PMB behaved without being accountable to the Project Committee. Needless to say that the authority of the Project Director was thereby automatically reduced, except over the staff and activities originally assigned to him by the Ministry of Marketing and Cooperatives. 6.03 The Ministry of Agriculture, through which no resources were channelled, collaborated passively at first (till 1972) , erratically in 1972-73, and almost not at all after 1973 when the Ministry of Marketing and Cooperatives was re-established. Field staff seconded from the Ministry of Agriculture and paid by the Project however, seem to have performed their extension and forestry tasks the best way they could as long as equipment and security conditions allowed. The technical officers to be employed by the Ministry of Agriculture for tobacco research and field trials were never hired. 6.04 The PMB entered into the Project with no other obligation than to maintain "separate and identifiable accounts" of its tobacco operations in return for IDA financing of tobacco storage facilities and equipment. Its financial accounts were never known, the storage facilities were not built with IDA financing, and it did not even perform its services as well as the former BAT Company. It, with the Ministry of Agriculture, was the weakest and eventually fateful link in the project organization's breakdown. 6.05 The Ministry of Finance, which was represented on the Project Committee, at best acted as a buffer between the various project entities, on the one hand, and on the other, between a badly going project and IDA. It seems, in fact, that by 1973, it had lost interest in the project. As representative of the "Borrower," it never firmly imposed any financial discipline regarding the maintenance of accounts and enforcement of financial covenants under the Credit Agreement (PMB, Tobacco Capital Account, the revolving fund administered by UCB) . Finally, when preparation of a new Plan of Operation was deemed essential by IDA and - 39 - Project management in 1972, in order to attempt to restore economic viability to the Project, the Ministry of Finance took no action for almost one year, until December 1973, when it was too late. 6.06 The Project Unit and its Director, with the scope of its influence and authority over the Project circumscribed to its own administration and the activities of the three Cooperative Unions, nevertheless, tried to perform effectively under the circumstances. The Project Unit provided the only continuity in the Project as its Director and his deputy survived the sweeping tide of political changes and upheaval, at least through 1973, even though they were directly exposed to pressures and threat, and operated in a climate of mistrust in institutional and individual relations. The Unit was relatively effective only as long as it could retain competent deputy Project directors who also managed the two largest Cooperative Unions (West Nile and Middle North), and accountants. The Cooperative Unions, their constituent local societies and rank-and-file members were not in a position to significantly affect the course of events and decisions of the Project Committee and Project Unit. The Middle North Union, on which success of the Project depended the most, was also the most vulnerable since it was caught up in political turmoil from 1971 on; the Kigezi Union never performed satisfactorily; and the West Nile Union was initially immune from political events until 1973-74, in part because of its location and distance from Kampala, and in large part because it was the most experienced and self-sufficient in flue-cured tobacco production. 6.07 In retrospect, the institutional design was unrealistically complex for a fairly simple and technically straightforward project. As designed, success of the project depended on too many agencies, most of which did not have a permanent vested interest in its success or, as for the PMB, had other means of pursuing its interests (high margins and marketing of other crops) . Hence, the Project Unit and the Cooperative Unions were the only sustainable elements of the project organization. In contrast with the Smallholder Tea Project, that was executed by the Uganda Tea Growers' Corp. which was responsible for all aspects of production, processing and marketing, it was wishful thinking on the part of IDA and the Uganda Government that a new and untested organizational structure, in which responsibility was divided among 5 entities, could make great strides in a short period of 4-5 years. It would have been desirable to either insist on retaining BAT or, if this was not possible, to await the establishment of its successor, the UNTC. B. The Project Unit 6.08 It is incontestable that without the Project Unit (PU) , the Project would not have started at all and continued as long as it did. For a time the PU filled a void -- albeit not to the satisfaction of tobacco producers and IDA -- left by the withdrawal of BAT from flue-cured tobacco production. The effectiveness of the PU was limited initially by its newness and inexperience, but more importantly by implicit institutional constraints on staffing, financial and procurement functions. It thus lacked the means to credibly respond to the financial, managerial, technical and marketing needs of the Cooperative Unions. After 1972, the Project Director was categorically instructed to appoint and dismiss staff - 40 - according to the political whims of the military Government. All but routine activities of the Project Unit, including communications with and supervision by IDA, became a matter for political intervention. Control of the PU over resources (inputs, equipment and material) acquired for the Project was tenuous, as supplies were reported to have been diverted for other purposes or resold by Government officials. Staffing Key Posts 6.09 Recruiting and retaining competent material, financial and technical staff was already difficult during the project start-up period. It affected the Project Unit as well as the Cooperative Unions' need for managers, accountants and senior technicians. When political pressures intervened in the staffing decision of the Project Director after 1972, the situation deteriorated to the point that the Project had virtually no qualified managers and accountants after mid-1974. 6.10 Once the Project Director was appointed in 1970, a deputy project director was selected and shortly thereafter appointed. He was seconded from FAO, where he had been Tobacco Production and Marketing Officer in Uganda since 1968. He brought 12 years' experience and seemed ideally qualified for the post. However, recruitment of cooperative union managers posed problems related to the need to hire expatriates and provide them sufficient incentives and assurances of job security. This was achieved only sporadically in West Nile, where the first Cooperative Union Manager was appointed in January 1971 by the Minister of Agriculture, Forestry and Cooperatives. He was concurrently appointed as Assistant Project Director. He resigned as of September 7, 1971, after terms of his employment contract with the Government were unilaterally amended to reduce salary and allowances. According to the farmers, the manager had succeeded, during a single season, in preventing the decline in West Nile's tobacco production which had started the previous year after BAT pulled out from West Nile. To offset poor PMB services, he had secured the fuelwood necessary for the cooperative members and the tobacco transportation, sufficiently to inspire confidence of farmers. A successor was appointed in West Nile, but he stayed in the post only one year; after some hesitation he accepted the same position in Middle North, from which he resigned in mid-1974. Eventually, therefore, since job security was not achieved and all expatriates were compelled to leave -- one was deported in July 1974 and the other left immediately thereafter for political security -- the Project Director had no deputy or assistant director. 6.11 The staffing problem was yet more acute for accountants needed by the Project Unit and the Cooperative Unions. Recruitment and retention of competent accountants was hindered first by the low government salary, U.Sh. 16,000 per annum for the most senior, which was one-third of private sector salaries. Efforts to raise accountants' salaries at the urging of both IDA and the Project Director were fruitless. The shortage of competent accountants became more acute with the expulsion of Asians, both foreign and Ugandans, which influenced Project accountants. As a result, the Project and the Cooperatives did not retain competent accountants very long and obtained inferior budgeting and accounting services until 1974, and had no accountants after mid-1974. This is reflected in the poor quality and late financial reporting by the three Unions, the qualified reports of auditors, and the lack of project accounts. - 41 - 6.12 With the departure and resignation of senior Project and Cooperative Union staff, and the extended illness (due to unsupportable pressures) of the Project Director, the project was "managed" by local agricultural field staff by the end of 1974. A proposal by Government to UNDP to provide a tobacco production officer and economist apparently never materialized. Upon his return to his post, the original Project Director carried on his responsibilities under existing constraints to the best of his ability until the closing date. C. Reporting 6.13 No regular reporting requirements were specified in either the Credit Agreement or in the SAR. There is no record that IDA requested formal progress reports on project implementation during the first two years of the Project. Thereafter, the project was in a constant state of crisis. IDA was apprised of progress, therefore, mainly through its supervision missions and partial reports submitted by the Project Manager in connection with procurement, staffing and disbursements. In connection with the last full supervision mission in October 1974, Project management prepared a briefing document that could be considered an approximation of a progress report, under the circumstances prevailing at that time. Moreover, it appears that strict censorship was imposed on communication between Project management and IDA from 1972 on, except for routine matters of disbursements and procurement. This is evident from records of communications from the Ministry of Finance which fail to convey particularly serious aspects of a situation or issue, which was known to be quite serious from informal reports by project staff. D. Accounts and Audits Accounts 6.14 The clearest indication of the poor quality of accounting of the Project and Cooperative Unions is chronicled by the IDA supervision mission up to and including 1975 rather than the fragmentary financial statements (audited and unaudited) that are on record. No Project accounts after 1972-73 are on record and those of Cooperative Unions are "draft" (i.e. unaudited) after 1974 for West Nile, 1973 for Middle North, while no financial records exist for Kigezi after 1971. The lack of competent accounting staff is but one reason for the chronic delinquency in maintaining and reporting the financial conditions of the Project and the Cooperative Unions. As mentioned earlier, the Cooperative Unions were in weakened, if not delinquent, financial conditions when the project began. It was short-sighted, therefore, not to require an initial audit, to enable them to begin anew, and make a possible comparison of financial performance forecast at appraisal with the status quo ante, with the advantage of clear knowledge of the status of assets and liabilities (e.g. uncollectable receivables, unserviceable or inexistent asets, etc.) . Since this was not done and financial irregularities were reported throughout the Project, competent accountants alone would not have been sufficient. In fact, on the occasion that professional accountants were assigned to Cooperative Unions (e.g. Kigezi in 1971 and Middle North in 1974) , non-compliance with basic bookkeeping and sound accounting procedures were reported along with findings of unrecorded transaction, lack of controls, etc. - 42 - Audits 6.15 The Project's accounts were to be audited by the Auditor General of Uganda after each fiscal year ending on June 30. The audited accounts for FY 1970/71 and 1971/72 were finally sent to IDA (RMEA) in May 1973 with qualifications of the Auditor General as to the Tobacco Capital Account, which had to be referred to the Bank's Legal Department. RMEA staff moreover communicated its concerns to the Project Manager regarding deficiencies in the content of balance sheets and income statements. Accounts for FY 1973 had presumably been submitted, but had not been certified by the Auditor General as of mid-1975 and there is no record that these have ever been submitted to IDA. Recommendations by IDA that Project accounts be audited by a commercial firm were not heeded by the Ministry of Finance. Commercial auditors for the Cooperative Unions' accounts were replaced by another firm in 1972 at IDA's request, after their unacceptable audit practices were revealed in Middle North. The new firm audited the 1972 accounts of West Nile and Middle North, but qualified its findings. The 1973 Middle North accounts were reported to be chaotic (in May 1975), those for Kigezi not available after 1971, and those for West Nile were audited, with qualifications, for 1974. 6.16 In summary, it appears that the Project lacked responsible financial management during almost its entire duration. Financial reports that were occasionally submitted to IDA were partially or totally defective, and auditors' reports were received too late after the end of a fiscal exercise to take any corrective measure. There was clearly a breach of all financial covenants of the Credit Agreement and, on this basis alone, on several occasions, IDA had sufficient cause to cease disbursements since corrective measures were not taken by the Borrower. VII. Bank Performance Project Appraisal and Design 7.01 The Project was appraised and approved by IDA mainly on the premise that progress made during the 1950's and early 1960's in introducing flue-cured tobacco to Uganda, under BAT guidance, could be rapidly expanded and made more efficient through state entities. In strict technical terms, project objectives were reasonable, in view of the favorable market for the flue-cured tobacco and the receptiveness to gradual technical innovation shown by Ugandan farmers in the past. A Smallholder Tea Project (Credit 109-UG) , approved in 1967, was going reasonably well, with nothing more than the "growing pains" of a new project, IDA approached the Smallholder Tobacco Project, therefore, with the plausible conviction that the Tobacco Project, like the Tea Project, was technically sound, could be made institutionally feasible and was economically viable. 7.02 Technically the Project was sound with objectives that, even now, do not appear to be unattainable. The time needed to achieve increases in both area (4,700 acres) and average yields (200 kg/acre) would have been disputable in 1969 as it is now, with the benefit of hindsight, but the basic means for the expansion were provided, e.g. credit for curing barns - 43 - and fuelwood plantations. The appraisal mission could be faulted for failing to provide tractors, since the AMD services were unreliable, but this alone would not have been sufficient. Hence, a delay of even 5 years in achieving appraisal objectives would not have been economically ruinous. 7.03 The appraisal was mute, however, regarding economic risks and institutional uncertainties. 1/ The Project was appraised, of course, some time before the Bank routinely analyzed project risks as it does now. It turned out that the most elementary of risks, the decline of economic incentives, affected the project almost from the outset in 1971, with the results given in 1972 when the IRR was re-estimated at -4.8%. Had the effect of changes in relative prices on the supply response of tobacco farmers been estimated, the appraisal mission would have undoubtedly rethought the implementation strategy. It would at least have included provisions for setting prices paid to farmers with reference to export prices. It could also have probably focussed more intensively on the incentives needed by the less experienced Middle North and Kigezi producers, which were expected to provide 90% of the acreage expansion and 70% of the increase in flue-cured tobacco output. Other questionable economic assumptions -- i.e. zero costs for farm labor, and a mechanistic sensitivity analysis with notional increases in costs and decreases in benefits -- both of which are no longer widely accepted in the appraisal process, conveyed a more favorable forecast of the Project's economic returns than it would merit now under ideal institutional conditions. 7.04 Uncertainty regarding the institutional feasibility of the Project was not apparently explicitly questioned in the appraisal. Bank staff did not distinguish the difference between reasonably successful tobacco production under private initiative and under interested state management. The need for a Project Unit was incontestable then and now; the rest of the Project's institutional apparatus, with diffused responsibilities, and no central accountability, and limited authority granted to the Project Unit, however, was ill-suited for this and similar projects. Tinkering with the institutional structure approved by the Borrower and the Bank would not have changed the outcome. Only the prompt action of an entity to replace BAT, with a similar degree of autonomy and control over resources as that of the Smallholder Tea Project, would have given the Project a chance to survive the ordeal that it faced. Supervision and Implementation Reviews 7.05 Beginning with the protracted process leading to credit effectiveness, and the first pre- and post-effectiveness supervision missions, IDA became aware of the defects in project design and implementation strategy, and the weak implmentation capacity in Uganda. Eight supervision missions were conducted between Aug. 1970 and November 1/ Risk is different from uncertainty, since some form of "insurance" or compensation can be built into project design and appraisal to protect against some risks. Uncertainty, i.e. the unknown, is not subject to meaningful probabilistic estimates. - 44 - 1974 (plus one by Comptrollers) . Although they varied in their degree of optimism, the reports are unanimous with respect to their perception of problems and their recommendations on the salient issues, as follows: (a) Need for autonomy of the Project Unit and control over staff; (b) PMB finances and tobacco prices paid to farmers; (c) Unreasonable delays in procurement; (d) Unsatisfactory financial management and non-compliance with accounting and auditing covenants; (e) Poor services by the Ministry of Agriculture's extension and AMD services, with resulting delays in land preparation; (f) Bottlenecks in the supply of fuelwood; and (g) Shortage of supplies, building materials and,spare parts due to import bottlenecks and restrictions. 7.06 The unanimity of supervision missions on the foregoing issues was due in large part to the continuity in the staff composition of such missions. Except for the fourth and seventh supervisions, at least one mission member had participated in the previous mission. Nevertheless, all but two of the supervision missions found cause to accentuate some positive finding leading to conclusions that conditions would improve. One supervision mission which spent only two days in Entebbe, without being able to venture out to Kampala and the tobacco growing districts, euphorically reported the status of the project "surprisingly heartening." The mission nevertheless reported on all the above problems, which were obviously becoming more acute, including decisions by Government to expel expatriates and restrict travel to Uganda by foreigners (including World Bank staff) . Because of Government restrictions and security risks, full field supervision by IDA was not conducted from June 1972 until March 1974. 7.07 Aware of these problems, IDA consistently supported efforts of Project management to try to keep the Project on course. For example, it supported the decision to accept the resignation of the West Nile Cooperative Union manager in 1971, irrespective of the breach of the employment contract (para. 3.12) . When the project was at an impasse, in June 1972, it did not heed the recommendation to suspend disbursements made by one supervision mission and followed it with another mission to find ways to restore economic viability to the Project, by extending the area and implementation period in Middle North. This led to recommendations for a new Plan of Operations, which became meaningless (para. 3.20) . Although IDA was aware that PMB was "taxing" farmers to finance its bloated overhead, and that the project was in deep trouble, including sizeable cost overruns in 1973, IDA acceded to the request and led the Ugandan authorities to believe that it still might finance the PMB tobacco stores (para. 3.33) . When the Deputy Project Director was peremptorily deported in 1974, the record shows that IDA could only enquire as to the deportee's safety and whereabouts. - 45 - 7.08 In the broader context of Uganda's deteriorating economic and security condition, the Bank's higher management (i.e. VP's) was concerned with the fate of all IDA-financed projects in Uganda. Two high level assessments were made, based on extensive staff deliberations. The first review, in early 1973, concluded that as a minimum objective, the Project should be brought into production. It deliberated on the rationale for suspending disbursement in light of breach of covenants, but decided against this ultimate course because inter alia (i) it would be perceived by Uganda as a hostile act and jeopardize minimum project objectives; and (ii) without proper supervision there was insufficient evidence to make a convincing case for suspension of disbursement. The second review, one year later in 1974, again stressed the vulnerability of the Project to security risks but recommended continuation of disbursements by direct payment to suppliers against verifiable claims in order not to jeopardize project implementation. 7.09 IDA's position from 1973, therefore, was clearly to contain the "damage" suffered by the Project., while disengaging itself from non-essential commitments (i.e. the tobacco storage facilities, Project preparation, tobacco research) and concentrating on supplying inputs and materials for flue-cured tobacco production. This was essentially a "holding operation" intended to expedite disbursements and close of the Credit on schedule. Implications for Project Outcome 7.10 Without knowing the evolution of political events after 1972, there would have been good reason for IDA to insist not only on a revised Plan of Operations, as it did for the sake of trying to restore economic viability to the Project, but to streamline the Project's organizational structure and consolidate the authority of the Project Unit. There is no reason to believe, however, that a stronger Project Unit or the earlier creation of the National Tobacco Corp. would have better survived politically than the fragmented organization that tried to implement the Project. More importantly, the Bank's approach to carry on with the Project regardless of Borrower performance was ill-advised. By 1973, when Project failure was evident, there were sufficient grounds on which to suspend disbursements for breach of key financial covenants and on grounds that the Borrower could not comply with any part of Section 4.18 of the Credit Agreement regarding (i) general status of the Credit; (ii) conditions that interfere with the accomplishment of its purpose; and (iii) affording opportunity to World Bank staff to visit the country for purposes related to the Credit. VIII. Conclusions 8.01 The Project failed to achieve any of its objectives for several reasons rather than due to a single cause such as events following the military coup d'etat in 1971. However, even if it is possible to isolate individual causes and their direct effects, it is more difficult to be certain about the multiple and simultaneous effects of all abortive factors. For example, since four major factors -- low prices, poor PMB services, weak PU, and political interference -- contributed to failure, - 46 - would the elimination of any one or more causes have improved the chances of success? Which symptoms of failure, i.e. stagnation and decline of tobacco production, financial collapse of cooperatives, staff demoralization, procurement and disbursement delays, etc. would have been healed by increasing prices paid to farmers, without taking any other action? Would the project have succeeded under a stable political regime other than the one that threw the country into upheaval from 1971 to 1979? Possibly, but not over a short 4-5 year period during which farmers' terms of trade deteriorated. 8.02 Measured against appraisal expectations, the project failed because (a) the Borrower did not take corrective measures recommended by IDA regarding prices and PMB functions; (b) IDA failed to insist that its recommendations be adopted and did not perceive the need for altering project organization at an early date when its ineffectiveness was evident; and (c) growing political interference in nearly all aspects of project implementation with the ensuing institutional, financial and economic and human ramifications. 8.03 Although appraisal objectives were quantitatively realistic, it was unrealistic to seek their attainment during a short 4-5 year period. In formulating the implementation strategy, the preparation and appraisal mission failed to consider the key factors that contributed to earlier success with flue-cured tobacco production, i.e. willingness of a single commercial enterprise (BAT) to experiment and innovate and share risks with farmers, unfettered by the legal formalities of either an IDA credit agreement or bureaucratic regulations and discord. In other words, BAT and the farmers had a common interest in achieving market-oriented results. The appraisal mission could not possibly have translated the BAT-farmer partnership into a farmer-Government partnership, when Government consisted of discordant parties. Finally, in formulating the implementation strategy, it should have been clear from the appraisal objective that since 90% of the area expansion and 70% of the project's production objectives were to be derived from Middle North and Kigezi, efforts should be focused on these areas. This would have implied providing West Nile with little more than the minimum complement of resources needed to increase internal efficiency, e.g. eucalyptus woodlots, credit and improved marketing arrangements. The validity of such a strategy was confirmed during the first 2-3 years of the project, as it was West Nile that sustained its level of production, even under the stress to which project management was subjected. - 47 - Annex I U GANDA SMALLHOLDER TOBACCO PROJECT (CREDIT 212-UG) PROJECT COMPLETION REPORT Actual Timetable of Conditions of Effectiveness 1970 1971 1. Signature of Credit Agreement 7/24 2. Target date of effectiveness and Extensions 10/31 1st Ext 12/31 2nd Ext 3/1 3rd Ext 5/1 4th Ext 8/2 3. Actual date of effectiveness 7/28 4. Conditions and date satisfied a. Establishment of Project Committee 11/13 b. Appointment of Project Director 11/30 c. Plan of Operation 1/6 d. Allow relending of IDA fund by Cooperative Unions under External Loan Act of 1962 10/26 e. Execution of subsidiary loan agreements (10) with Cooperative Unions and PMB 7/28 f. Agreement with commercial banks to continue supplying credit to existing tobacco growers 11/13 g. Agreement with UCB to be disbursing agent for funds lent to Cooperative Unions 12/10 h. Purchase of BAT facilities in W Nile and M North districts by Cooperative Unions 11/13 i. Legal opinions of Government 7/28 - 48 - Annex II UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report Key Indicators Appraisal 1971 Plan 1973 1974 1974 as % Estimates of operation Actual Actual of Appraisal (year 4) (year 4) Tobacco Area (Acres) Total 11,800 10,654 8,098 8,100 44 New 4,708 3,304 - n.a Yield (kg/acre) 392 428 275 322 82 Production ('000 kg) Total l/ 4,631 4,561 2,230 2,635 57 Incremental 2,201 2,131 - 205 1/ - Fuelwood Plantings 7,968 15,146 7,536 7,736 97 (Acres) Curing Barns New 4,711 2,922 1,466 1,802 38 Converted 1,600 831 n.a n.a 1/ Approximation from unions and users / Figures based on the SAR assumption of 2.4 million kg production in base year; actual base year (1968-69) was 2.8 million kg. Therefore, there was in fact no sustained increase in production, and it decreased after 1974. UGANDA SMALLHOLDER TOBACCO PROJECT PROJECT COMPLETION REPORT (CREDIT 212-UG) ACTUAL AND PLANNED FLUE CURED TOBACCO AREA (ACRES) A) TOTAL ACREAGE WEST NILE MIDDLE NORTH KIGEZI TOTAL Actual Plannedl/ Revised2l Actual Plannedl/ Revised P Actual ActualPlanned Revised 1969 3,630 5,630 3,630 2,646 2,978 2,644 434 484 6,710 7,092 6,758 1970 5,190 4,000 4,036 2,646 3,360 2,646 509 550 8,345 7,910 7,242 1971 5,214 4,400 4,686 1,891 4,057 1,955 586 625 7,691 9,082 7,266 1972 5,054 5,000 5,054 2,615 4,811 2,517 600 725 8,269 10,536 1,296 1973 5,211 5,000 5,211 2,411 6,000 2,650 476 800 8,098 11,800 8,661 1974 2,383 5,000 5,693 2,221 6,000 4,161 620 800 5,244* 11,800 10,654 1975 5,000 6,053 6,000 5,328 800 9,452* 11,800 12,181 1976 5,000 6,053 6,000 6,530 800 7,890* 11,800 13,383 1977 5,000 6,053 6,000 6,530 800 6,700* 11,800 13,383 1978 5,000 6,053 6,000 6,530 800 5,790* 11,800 13,383 1979 5,000 6,053 6,000 6,530 800 4,100* 11,800 13,383 1980 5,000 6,053 6,000 6,530 800 11,800 13,383 B) NEW ACREAGE PLANNED 1970/71 370 382 66 818 1971/72 400 697 75 1,172 1972/73 600 754 100 1,454 1973/74 - 1189 75 1264 1,370 1,799 3,022 4,012 316 366 4,708 6,177 1/ SAR 2/ Targets were revised on at least two occasions; first, upon preparing the Plan of Operations, before credit effectiveness in 1971, and the second in 1973. 3/ No revisions made until 1973. A/ From project review and supervision in November 1974, until 1974. From NTC data thereafter 5/ Proposed revisions in late 1972 by IDA supervision mission. NOTE: Data after 1974 is very unrealiable in view of upheavals in Uganda and the transfer of responsibilities from the Produce Marketing Board to the National Tobacco Corporation (NTC) in 1977-1978. NTC apparently revised previous years' acreage, yield & production data. Acreage estimates of the NTC* seem to have been derived by dividing total production by yield estimates, thereby giving at best a speculative result. *aft r 1975 - 50 - Annex IV UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report West Nile Middle North Kigezi Total 2/ Actual Planned 2/ Actual Planned 2/ Actual Planned 2/ Actual Planned 1968 2,221 478 57 2,757 1969 2,232 589 128 2,949 1970 2,140 1,800 514 773 163 126 2,817 2,699 1971 2,328 1,980 472 1,168 183 182 2,983 3,330 1972 2,539 2,250 493 1,611 134 238 3,166 4,099 1973 1,593 2,250 503 2,087 137 294 2,233 4,631 1974 2,106 2,250 379 2,412 151 335 2,635 4,997 1975 2,014 2,250 455 2,544 118 347 2,586 5,141 1976 1,976 2,250 359 2,544 129 356 2,463 5,150 1977 1,590 2,250 202 2,544 12 360 1,805 5,310 1978 1,071 2,250 92 2,544 360 1,163 5,310 1979 607 2,250 11 2,700 54 360 672 5,310 1/ All figures rounded to nearest thousand. 2/ Planned data from SAR. Source: SAR (Annex 14) and NTC date published in 1981. - 51 - Annex V UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report Actual and Planned Flue-Cured Tobacco Yields (kg/acre) West Nile Middle North Kigezi Weighed Average Actual Planned Actual Planned Actual Planned Actual 3/ Planned 1969 571 222 1/ - 295 1/ - 415 1970 413 450 173 230 320 230 336 340 1971 498 450 223 230 312 300 414 341 1972 504 450 196 350 223 350 377 367 1973 305 450 191 400 288 400 252 389 1974 391* 450 168* 450 280* 450 322 392 1975 354* 450 200* 450 241* 450 273 423 1976 385* 450 179* 450 225* 450 313 436 1977 298* 450 152* 450 250* 450 264 436 1978 234* 450 98* 450 - 450 201 450 1979 - 450 450 450 164 450 1/ SAR figures were used. However, according to the NTC actual yields for 1968 (reference years) were 612, 191 and 199 respectively, and in 1969, West Nile yields were 612 kg/acre rather than 577 kg/acre. 2/ Estimated in November 1974 Supervision Report up to 1974. 3/ Actual data provided by NTC in July 1981 which is different from data obtained during supervision missions through 1974. For this reason, when NTC data is used, it is marked with an asterisk (*) and should be taken as "best estimates" of the NTC. 4/ From SAR. - 52 - UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report Farm Costs and Returns from Flue-Cured Tobacco in West Nile and Middle North Appraisal Actual (October 1974) Estimate Middle North West Nile Production Kg tobacco/ac (estimated average) 400 200 400 Income Sales U.Shs.: 4.73/kg 1,892 UShs:5.5 1,100 UShs:5.5 2,200 Crop Loan 600 800 800 Barn Loan - - - Total 2,492 1,900 3,000 Expenditures Purchases of wood fuel 60 15 30 Transport of wood fuel 100 125 250 Raising seedlings 50 80 80 Land preparation 110 125 125 Fertilizer 100 472 472 Insecticides 30 70 70 Insurance 60 Baling materials 50 75 150 Other materials 60 15 30 Barn repairs ) Flue-pipe replacements ) 76 165 165 Barn construction -_- - Total 636 1,142 1,432 Debt Service Barn Loan 375 450 450 Crop Loan principal 660 800 800 Crop Loan interest 80 80 Total 1,671 572 2,762 Net Income 821 (572) 238 UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report Actual and Planned Fuelwood Plantations (Acres) West Nile Middle North Kigezi Total Actually Actually Actually Actually Planned Planted Planned 1/ Planted Planned 1/ Planted Planned 1/ Planted SAR 1969 - n.a. - n.a. - n.a. - n.a. 1970 - n.a. - n.a. - n.a. - n.a. 1971 713 699 978 743 111 129 1,802 1,571 2,375 1972 2,000 1,395 1,379 1,379 491 131 3,888 2,905 2,137 1973 2,000 1,678 2,380 1,208 455 374 4,835 3,260 1,728 1974 1,785 n.a. 2,380 n.a. 457 n.a. 4,622 n.a. 1,728 Total 1/ 6,498 3,772 7,135 3,330 1,514 634 15,146 7,736 7,968 1/ Planned figures from Plan of Operations. Source: March 1974 supervision report, SAR and 1971 Plan of Operations data in SAR does not coincide with subsequent supervision reports and the 1971 Plan of Operations. m 1-4 Annex VIII - 54 - UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report Cumulative Disbursement Schedule Appraisal Calendar Forecast and Actual Year Extensions Amount as % of Appraisal -------------------------US$ '000-------------------------- 1971 June 30 1,250.0 1972 June 30 2,337.0 396.7 17 1973 June 30 3,086.0 1,762.0 57 1974 June 30 4,000.0 2,170.5 54 1975 June 30 4,000.0 2,341.1 59 December 31 4,000.0 2,855.6 59 1976 (Last est.) March 31 4,000.0 (Extended) 4,000.0 (4/8/76) Annex IX - 55 - UGANDA SMALLHOLDER TOBACCO PROJECT (Credit 212-UG) Project Completion Report International and Domestic Flue-Cured Tobacco Price Comparisons International Uganda Export Domestic Wholesale Farmers' Price 1/ Price 2/ Price Price (U.Shs.) Index (U.Shs.) Index U.Shs. Index U.Shs. Index 1968 13.7 100 9.99 100 6.32 100 5.83 100 1969 13.2 96 11.25 113 6.50 103 4.81 83 1970 13.8 101 10.97 110 6.87 109 4.79 82 1971 13.5 95 11.61 116 6.84 108 4.61 79 1972 13.4 98 11.56 116 7.15 113 4.97 85 1973 13.4 98 13.00 130 7.70 122 4.76 82 1974 18.0 131 15.55 156 8.97 142 5.93 102 1975 18.0 131 14.16 142 11.51 182 7.06 121 1976 16.0 117 17.22 172 12.26 194 7.25 124 1977 21.0 153 16.89 169 13.08 207 8.00 137 1978 14.0 102 21.39 214 - 9.16 157 1979 22.0 161 17.20 172 - 11.30 194 1/ India flue-cured export unit value, at current prices expressed in U.Shs. con- verted at official rates or exchange. (U.Shs. 7.14 per US$ to 1974, and there- after as per IMF Financial Statistics). 2/ From NTC export data. 3/ Unknown proposition of production sold at wholesale prices.  UG.ANDA SMALLHOLDER TOBACCO PROJECT D A ARUA. GULu W., N;Jý P.k- Prolet ar cz o M.TAL.h,~ Kig4 LIRA o Railways 3Ama0ch Mam rro ds Bata BAL N o bu yoNaaub POR L Hmo e kaorRgil coc KAMPALAE ORTAL~AK V-/CTORIA . TRO MUEENDE KM,g, B 0' ,J INJA /APL T. A-NNZEASNE Rz A D A 0R KIGAll