Malawi - Karonga Rural Development Project

Document of The World Bank FOR OFFICIAL USE ONLY FILE Cu0 PY Report No. 2576 PROJECT PERFORMANCE AUDIT REPORT MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I (CREDIT 282-MAI) June 29, 1979 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.  - i - FOR OFFICIAL USE ONLY Project Performance Audit Report MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I (Credit 282-MAI) TABLE OF CONTENTS Page Preface iii Basic Data Sheet iv Disbursement Table v Highlights vi PROJECT PERFORMANCE AUDIT MEMORANDUM I. Summary Background 1 The Project 2 Implementation 2 Agricultural Development 3 Other Development 4 Rate of Return 5 Financial Developments 5 II. Major Issues Inadequate Project Preparation 6 Agricultural Development and Marketing Corporation's Pricing Policies 8 Project Staffing 10 Conclusions 10 Annex I - Borrower Comments 12 PROJECT COMPLETION REPORT I. Background 23 II. Formulation 24 III. Implementation 25 IV. Institutional Performance and Development 33 V. Financial Performance 36 VI. Project Impact 41 VII. Rates of Return 51 VIII. Bank Performance 51 IX. Conclusions 52 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 IFC authorization. - 11 - TABLE OF CONTENTS CONT'D Page Appendices 1. Staffing Pattern 54 2 Seasonal and Medium-term Credit 55 3. Crop Purchases by Admarc 56 4. Illustrative Crop Budget - Irrigated Rice - I Acre 57 5. Completion Report 64 Map - iii - Project Performance Audit Report MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I (Credit 282-MAI) PREFACE This is a performance audit of the Karonga Rural Development Project - Phase I, in Malawi, for which Credit 282-MAI was approved in January 1972 in the sum of US$6.6 million. The closing date is June 30, 1979 and the credit was fully disbursed on April 30, 1979. The audit report consists of an audit memorandum prepared by the Operations Evaluation Department and a Project Completion Report (PCR) dated December 8, 1978. The Project Completion Report was prepared by the Eastern Africa Projects Department. The Audit Memorandum is based on a review of the Appraisal Report (No. PA-106a) dated December 30, 1971, the President's Report (P-1008) of January 6, 1972, the Credit Agreement dated January 26, 1972 and the PCR; correspondence with the Borrower and internal Bank memoranda on project issues as contained in relevant Bank files also have been reviewed and Bank staff associated with the project have been interviewed. A copy of the draft report was sent to the Borrower on February 16, 1979. The comments received from the Government (see Annex I) have been taken into account; noted corrections in the PPAM and PCR have been so incorporated. The audit finds that the PCR presents a comprehensive and objective account of the project's development and attainments. The audit memorandum focuses on several issues that merit further discussion because of their relevance to this and to other projects in Malawi and East Africa: (1) the failure of two project components - the irrigated rice scheme and lake transport, (2) the revamping of another component - livestock development, (3) staffing problems and (4) the impact of price policy on project results.  - iv - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET MALAWI KARONGA RURAL DEVELOPMENT PROJECT PHASE I (Credit 282-MAI) KEY PROJECT DATA Appraisal Actual or Item Expectation Current Estimate 1/ Total Project Cost (US$ million) 7.8 7. Underrun (%) 6 Credit Amount (US$ million) 6.6- Disbursed - 6.6 Cancelled - - Repaid to - - Outstanding to - 6.2 1/ Date Physical Components Completed 3/31/77 9/30/76- Proportion Completed by Above Date ()- 94 Proportion of Time Underrun (20 Economic Rate of Return (%) over 12 2 6 OTHER PROJECT DATA Original Actual or Item Plan Revisions Est. Actual First Mention in Files - 5/ /70 Government's Application - 10/ /70 Negotiations 11/ /71 11,12/ /71 Board Approval 1/18/72 1/18/72 Credit Agreement Date 1/26/72 1/26/72 Effectiveness Date 5/25/72, 8/25/72 8/25/726/ Closing Date 12/31/77- 12/31/78, 6/30/79 6/30/79 Borrower Republic of Malawi Executing Agency The Government of Malawi Fiscal Year of Borrower April 1 - March 31 Follow-on Project Name Karonga Rural Development Project Phase II Loan Number 1286 - MAI Amount (US$ million) 9.2 Loan Agreement Date 6/24/76 MISSION DATA-/ Sent Month, No. of No. of Date of Item b Year Weeks Persons Manweeks Report Appraisal 4/71 4 5 20 12/71 Total Supervision - Transport I Washington 9/72 .5 1 .5 10/72 Supervision - I RMEA0/ 12/72 n.a. 2 n.a. n.a. Supervision - (Livestock) II RMEA 3/73 .5 2 .5 3/73 Supervision - Transport II Washington 4/73 1 2 2 5/73 Supervision - III RMEA 9/73 .5 3 1.5 10/73 Supervision - Transport III Washington 11/73 1 1 1 1/74 Supervision - IV RMEA 5/74 .5 2 1 6/74 Supervision - Transport IV Washington 7/74 .5 1 .5 8/74 Supervision (Disbursements) Washington 10/74 .5 1 .5 11/74 Supervision - V RMEA 1/75 .5 2 1 2/75 Supervision - Transport V Washington 10/75 .5 1 .5 11/75 Supervision - VI RMEA 11/75 1 2 2 2/76 Supervision - VII RMEA 7/76 1 1 1 8/76 Supervision - Transport VI Washington 5/77 .5 2 1 6/77 Supervision - Transport VII Washington 10/77 .5 1 .5 Total 13.5 Completion 4/78 1.5 3 4.5 10/78 Total 18.0 COUNTRY EXCHANGE RATES Name of Currency (Abbreviation) Malawi Kwacha (MK) Year: Appraisal Year Average (1971) Exchange Rate: US$1 = MK .769 Intervening Years Average (1972-75) US$1 - MK .841 Completion Year Average (1976) US$l = MK .907 1/ Project reduced by eliminating lake jetty, 500 ac. irrigation scheme, a livestock holding ground, and 4 livestock marketing yards, amounting to about 10% of estimated project cost. 2/ Phase I was shortened by one year. 3/ Including exchange adjustment of US$ 10,000. 4/ Two rates not strictly comparable, PCR Appendix 5. 5/ As shown in Credit Agreement. 6/ Final disbursement date. 7/ Supervision missions (except Transport) sent by the Regional Mission in East Africa, Nairobi. 8/ PMEA until October 1973.  - v - MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I Comparison of Estimated and Actual Accumulated Disbursements US$'000 Equivalent IDA Fiscal Year Actual as % of and Semester Actual Estimate Estimate 1971/72 December 31, 1971 - June 30, 1972 - 88 - 1972/73 December 31, 1972 94 526 18 June 30, 1973 403* 1,314 31 1973/74 December 31, 1973 1,045* 2,101 50 June 30, 1974 1,592* 2,800 57 1974/75 December 31, 1974 2,517* 3,524 71 June 30, 1975 3,328*' 4,274 78 1975/76 December 31, 1975 4,085* 4,978 82 June 30, 1976 5,146* 5,636 91 1976/77 December 31, 1976 5,593* 6,116 91 June 30, 1977 6,191* 6,419 96 1977/78 December 31, 1977 6,204* 6,600 94 June 30, 1978 6,206* 6,600 94 1978/79 December 31, 1978 6,213* 6,600 94 April 30, 1979 1/ 6,600* 6,600 100 * Includes exchange adjustment. 1/ The closing date for the loan was changed to June 30, 1979.  - vi - Project Performance Audit Report MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I (Credit 282-MAI) HIGHLIGHTS Karonga was the fourth rural development project in Malawi supported by Bank financing. The previous three being the Lilongwe Agricultural Development Program, Phase I and Phase II, and the Shire Valley Agricultural Development Project. Since the Karonga - Phase I - Credit was approved, credits for Phase II of Shire Valley, Phase III of Lilongwe, Phase II for Karonga and the National Rural Development Project, Phase I have been approved by IDA. As in the case of the three prior rural development projects, Karonga was a multifaceted project with components covering irrigation, rainfed crops and livestock development; supported by improved credit, improved extension services, water supplies, health facilities and transport. Karonga was the first of the four to include a health component. Except for irrigated rice and lake service transport, both of which performed poorly, project components were successful in achieving their objectives and now are expected to yield rates of return above those estimated at appraisal. For the project as a whole however, the rate of return is re-estimated at 6 percent compared with 12 percent at appraisal. Due to experience gained during project implementation and to offset cost overruns and price escalation, irrigation, livestock, health and transport components were reduced from project appraisal levels. The pricing policies of the Agricultural Development and Marketing Corporation appear to have contributed to lower than expected crop yields. The following items may be of special interest to the Bank: - inadequate preparation and design of irrigation component (PPAM paras. 7, 20 and PCR paras. 3.04, 3.12 - 3.14, 8.01 and 8.02); - poorly designed barge that was unsuitable for Lake Malawi weather (PPAM paras. 23, 33 and PCR para. 3.07); - vii - - lake jetty built too low to accommodate highest water levels (PPAM para. 22 and PCR para. 3.06); - only rainfed rice and maize achieved production targets (PCR para. 6.05); - all crop yields were below targets (PCR para. 6.02); - demonstration effect of project improves performance of non-participants (PCR para. 6.03); and - danger of neglecting minority components in multi-sectoral projects (PPAM paras. 32 and 33). Project Performance Audit Memorandum MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I (Credit 282-MAI) SUMMARY Background 1. The Karonga Rural Development Project Phase I (KRDP) was designed and implemented to promote rural development in areas near Lake Malawi in the district of Karonga which lies in the northern part of Malawi. The total project was expected to cost US$7.8 million of which 85% would be financed by the IDA credit of US$6.6 million. The Government and participating farmers would finance the balance of US$1.2 million. Foreign exchange cost of the project was estimated at US$4.1 million. The project was reduced during implementation by eliminating several sub-components because of large increases in costs. Total cost of the reduced project amounted to US$7.3 million. The Credit, however, remained at US$6.6 million. 2. The Malawi Government with the assistance of the Permanent Mission in East Africa (PMEA) prepared the project and subsequently submitted a request for financing to the Bank in October 1970. An IDA mission appraised the project in April 1971 and recommended certain changes in the original project design which were accepted by the Government. The credit for the modified project was approved in January 1972. It became effective in August 1972, after a three months delay because the Government failed to present to the Bank on time a satisfactory plan to improve Lake Malawi transport services and because of delays in appointing.the project manager. Both were condi- tions of effectiveness. 3. The District of Karonga in the northern region was considered by agriculturalists to have good potential for growing a large variety of crops and raising cattle. This development had been constrained by lack of infrastructure and the area's isolation due to rugged mountain terrain. In early years, all weather roads through this area were not thought to be feasible. Lack of health services also was retarding development. The main objective of the project was to accelerate rural development mainly through increasing crop and livestock production which would be supported and augmented by improved extension, research, credit, health facilities and transportation. -2- The Project 4. The main components of the project were: a) development of irrigated and rainfed rice schemes; b) expansion of acreage of rainfed maize, cotton and groundnuts under improved production practices; c) development of livestock industry; d) seasonal and medium-term credit for farmers; e) crops and hydrology research and a feasibility study for a phase II project; f) construction of five rural health outposts, improvement in two hospitals and provision of staff and drugs for bilharzia control; and g) expanded lake terminal facilities including access roads and a 200-ton freighter for lake transport. To implement agricultural and health components of the project a manage- ment unit was set up within the Ministry of Agriculture and Natural Resources. The Ministry of Transport and Communications had the responsi- bility for the lake services component. in addition, a Karonga Rural Development Project Liaison Committee was to be established to coordinate interministerial activities of the project. Implementation 5. The project was delayed initially by the slowness of the Government in appointing a project manager, the lack of staff housing and difficulties in recruiting senior staff. It became apparent, too, in late 1972 that some project components had to be revised due to poor forecast of costs at appraisal and changes in the Government's overall health plan. More importantly, because of general price escalation and large cost overruns for irrigation works and lake transport, the Government decided, and IDA concurred, that to keep within the original total cost estimate it would be necessary to delete certain project components and accordingly, to reduce the implementation period of Phase I by one year. The following components were deleted: (a) 500 acre extension of the Wovwe irrigation scheme; (b) four cattle - 3 - markets and an 18,000 acre holding ground; (c) improvement of the Chilumba rural hospital; and (d) improvement of Chipoka jetty at the southern terminal. 6. Although the construction of buildings for staff housing, health outposts, etc. began slowly, it gained momentum and most of these facilities were eventually completed according to schedule. Road construction was carried out as planned, and apparently the mileage constructed exceeded the appraisal expectations, but because the appraisal report is not clear on the mileage planned the exact comparison cannot be made. 7. In other construction components, large cost overruns and delays and consequently, deletions occurred. After adjusting for inflation the Lufira irrigation scheme incurred a cost overrun of more than 250 per cent due to changes in design and underestimated unit cost. For this reason, inter alia the Wovwe irrigation scheme was dropped from the project. The vehicles, machinery and equipment category also experienced a significant cost overrun; much greater than explained by general inflation. 8. When the Karonga project was conceived, it was decided that the transport capacity to and from the project area was in- sufficient. Road transport was generally ruled out, because of the poor condition of existing roads and very difficult terrain which increased operating costs. Transport on Lake Malawi was judged the only suitable means to increase capacity. This was to take the form of construction of better jetties at each end of the lake transport route and procurement of a 200 ton barge, to increase the capacity of the existing fleet. There was a cost overrun of about 35% on procurement of the barge, 88% on the northern jetty. There would have been an estimated 120% overrun on the southern jetty, but because available funds were insufficient, the latter jetty was not built under the project. Agricultural Development 9. The improvement and expansion of crop production was mixed, although total acreage planted at full development has exceeded appraisal estimates. At appraisal about 1,500 acres of irrigated land were expected to be developed and planted to rice by over 800 farmers at full develop- ment. By 1976/77, only 800 acres of the 1,000 acres Lufira scheme had been developed for irrigation but project farmers planted only 195 acres of rice. Consequently, the Government allocated about 460 acres to Spearhead Enterprises, the commercial wing of the Malawi Young Pioneers. Because of cost overruns and other problems at Lufira the 500 acre scheme on the Wovwe river was dropped (PPAM para. 5). The planting of rainfed rice on the other hand, exceeded expectations by 16 percent. The number of farmers participating was more than double the appraisal estimate. Planting of maize and groundnuts exceeded expectations by 90 and 60 percent, respectively. On the other hand, cotton acreage was barely one-tenth as large as envisaged. Yields of all crops up to the present have been between 60 to 90 percent of those expected, cotton standing at the low end of the range and rainfed rice at the high end. The relatively low cotton yields achieved together with low prices likely explain why cotton plantings were much less than expected (see PPAM para. 25). 10. Under the livestock development program 24 dipping tanks have been built. As a result most animals are now within five miles of those tanks, and the percentage of animals dipped has increased from 55 to 80 percent in Karonga and from 21 to 85 percent at Chitipa. Because of this improvement in dipping, animal health has improved and the herd growth has increased from 2.4 percent to 6.7 percent while maintaining an off-take of 10 percent per annum. At appraisal it was expected that the off-take rate would rise to 12 percent. This rate was not achieved, however, most likely because more males are being retained in the herd for draft power. 1/ Some of this estimated improvement in livestock output, however, may be due to improved methods of collecting data. Other Development 11. Health facilities have been greatly expanded and improved under the project. Five rural health outposts have been constructed and three improved. A bilharzia control program has been introduced in three irrigated areas of the project. The capacity of the Karonga hospital has been increased 50 percent. Visits at this hospital and the health outposts have increased greatly. 12. The reduced lake transport part of the project was completed after considerable delay. A 380-ton (instead of the planned 200-ton) barge was purchased because the marginal cost of the nearly double size was only a few hundred dollars. When this barge was put into operation it cracked during heavy weather. It was out of operation for repairs for about one year. After much negotiation, it was agreed that, repair cost would be borne by the consultant, supplier and Government, with 1/ Current estimates of the livestock populations at the beginning of the project are 15 percent higher than those used for appraisal. - 5 - the Government paying the largest share. The appraisal does not provide a separate economic return for the lake transport component. The PCR in para. 6.16 concludes that the lake transport component was apparently premature, because traffic did not grow as expected. Meanwhile, the appraisal of the second project in early 1976 found an economic return of 18 percent for the reconstruction of the southern terminal. This return is based on the assumption that the terminal would become virtually useless without rehabilitation. 13. The use of credit by farmers too was less than appraisal estimates. Near the end of Phase I about 5,000 farmers were making use of credit. Seasonal loans made have been running at 50 percent of those expected at appraisal and medium term credit one-third that expected. This low credit activity may have been due to an increase from 10 to 12.5 percent in the interest rate charged by the project authority and the rising cost of annual inputs such as fertilizer, pesticides and seeds. The recovery rate for loans declined from over 70 percent early in the project to just over 50 perint in 1975/76, but as of September 1977 had risen to over 85 percent.- Rate of Return 14. The rate of return for the modified project is estimated at 6 percent compared with 12 percent for the original project. This low rate of return is due to the very poor performance of the irrigated rice component. If this component is excluded, the estimated rate of return is 33 percent. Financial Developments 15. Because of world-wide inflation and poor estimates total costs for machinery, equipment, buildings and other construction exceeded estimates. These rising costs necessitated changes in the project (PPAM para. 5) but tended to increase the rate of disbursements. After a slow start, disbursement caught up to the original schedule by December 1977. The closing date was extended first from December 31, 1977 to December 31, 1978 in order to carry out improvements in the Chilumba jetty. The Bank then extended the closing date to June 30, 1979 as US$397,000 was unallocated. The credit was fully disbursed on April 30, 1979. 16. Only minor procurement problems were encountered throughout the disbursement period and most contracting and procurement of goods and services were carried out in accordance with IDA Guidelines. 1/ See Borrower's comments (PPAM Annex I, B.2.10 - B.2.16). - 6 - II. MAJOR ISSUES Inadequate Project Preparation 17. The major set-backs and problems that have been associated with the project relate to inadequate project preparation and the Bank's lack of resolve to stand fast on its earliest positions. Lack of adequate preparation affected the irrigated rice, livestock and lake transport components of the project. The irrigated rice scheme and the lake trans- port components suffered from this; fortunately, from the production viewpoint, the livestock component did not. 18. Livestock. Four new markets and an 18,000 acre holding ground were dropped from the project in 1975 when it became apparent that the existing cattle markets were being under-utilized and that there would be no demand for the cattle holding facility. At appraisal it was envis- aged that the veterinary service would receive cattle at the holding ground for supplemental feeding during the dry season and further fatten- ing before shipping them to the south for consumption. Livestock raisers have chosen either to keep more male animals for draft power or sell any surplus for slaughter locally. Apparently, they have not found the Government scheme competitive with these alternatives. One preappraisal mission noted that information on livestock, livestock marketing and producer's willingness to participate in the project was lacking. Thus, inclusion of the livestock component involved a known risk that was not mentioned in the appraisal report. 19. The cattle program apparently has been effective and there has been a significant increase in the rate of growth of the livestock herd from about 2.4 percent to 6.7 percent. Current information on the estimated livestock herd in the project area indicate that it may have been 20 percent higher than thought at appraisal. However, there is still some question about t 1 accurateness of the estimation of livestock numbers (PPAM para. 10).- 20. Irrigated rice. Some Bank staff were skeptical about the feasibility of the irrigated rice component prior to appraisal. The consensus was that the component needed more study, citing the lack of soil surveys, hydrological information, insufficient water and potential problems with plant disease and flooding. These turned out to be the major factors that doomed the full development of the irrigated rice scheme as well as causing a significant cost overrun. Because soils were more porous than anticipated, canals had to be lined. Canals and weirs were damaged or destroyed by two serious floods and double cropping of rice turned out to be infeasible because of plant disease problems as well as lack of suitable short growing season varieties. At appraisal the Government 1/ For a fuller discussion on this point see Borrower's comments (PPAM Annex I, paras. A.2.1 - A.2.6). - 7 - thought irrigation was feasible and necessary in the project area and the Bank agreed to finance the component despite the fact that it was marginal and risky. In the appraisal report, thT,estimated rate of return for the rice irrigation component is 9 percent.- 21. Transport. The lake transport component was still in an early stage of preparation when it was appraised, and if it had been a separate transport project, it would not have been appraised until much later. -A transport economist who joined the appraisal mission, in essence identified the project and during the subsequent eight months before negotiations, the lake transport component was further defined. Still, at the time of loan signature, only preliminary cost data were available. If final engineering had been available prior to credit approval, much of the cost overrun could have been avoided because (a) construction quantities and cost estimates would have been more accurate, and (b) construction could have been substantially completed before major price increases due to the oil crisis occurred. 22. Once the northern jetty was built, it turned out to be too low in relation to prevailing water levels. Remaining funds from the credit were used to increase the height of the jetty by 2-1/2 feet. Considering the uncertainty of the maximum water level and the substantial variation on record (about 20 feet), it seems that serious consideration should have been given to a floating jetty as an alternative to the present solution. According to some sources, fluctuations in the water level are expected to be less extreme since completion of a dam which controls the outflow of the lake. However, the dam already existed before the project was appraised and since it is built for electricity generation, it may in dry years even contribute to a lowering of the lake's water level. Therefore, it appears that even now too little information is available to form a definite opinion on the best engineering solution for the jetty. 23. The 380-ton lake barge was procured at a cost of US$375,000. After five trips the barge had to be taken out of service due to severe cracking as a result of inadequate design (PPAM para. 12). Major structural revisions were required with an estimated cost of US$50,000, or 13 percent of the original cost. The barge appeared to be designed for river transport and could not withstand the strong winds and high waves that occur on Lake Malawi. In addition, some of the locally done welding when assembling the barge proved of poor quality. It has not been possible to determine whether the consultants or the supplier were responsible. 1/ Other relevant information and comments have been provided by the Borrower (PPAM Annex I, paras. A.2.7 - A.2.12). - 8 - 24. Very little information is available on the roads component, not even 7nough to determine its relative importance as part of the project.- Agricultural Development and Marketing Corporation's Pricing Policies (ADMARC) 25. There is considerable room for doubt concerning the efficiency of ADMARC's pricing policies, which has been brought out in the PCR paras. 6.09 - 6.12 but needs further elaboration. It has been the policy of ADMARC to heavily "tax" farmers who produce cotton and groundnuts, especially cotton, and to subsidize the production of rice and perhaps to a certain extent maize. This policy is illustrated by the data in the following table. If farmers had been paid prices around the economic value, at least on the average, for cotton and groundnut, it is likely that the acreages of these c5?ps would have been near those estimated at appraisal (PCR, Table 8).- Table 1: ADMARC Profit by Commodity-Percentage of average Price Paid to Producers (Percent) Year Cotton Groundnuts Rice Maize 1972 33 54 -15 - 5 1973 38 48 -10 - 3 1974 74 48 + 9 +75 1975 92 24 +36 + 8 1976 16 24 - 2 -49 1977 64 106 -46 -45 Source: Staff Working papers. 26. Perhaps, too, the yields would have been closer to the estimates (PCR, .Appendix 4) as farmers likely would have cultivated their crops more intensively by using more fertilizer, etc. The appraisal apparently assumed, that farmers would receive adequate incentive prices for the commodities involved (Appraisal Report Annex 5). 1/ See Borrower's comments (PPAM Annex I, paras. A.2.13 - A.2.17). 2/ See Borrower's comments (PPAM Annex I, paras. A.2.18 - A.2.20). - 9 - 27. This apparent lack of concern about adequate producer price incentives may be due to the continuing debate about the responsiveness of farmers to price incentives. Yet one study in Malawi 1/ demonstrated that even smallholders respond to improved price incentives. 28. ADMARC's policy seems to follow one prevalent in Africa which is to tax crops that are exported e.g. cotton and subsidize food crops e.g. maize. The policy is most likely related to the desire for food self-sufficiency. Supervision missions raised the issue of unsatisfactory producer prices on numerous occasions, not only for the Karonga but also for the other Malawian rural development projects (Shire, Lilongwe). The missions correctly identified the lack of proper incentives to producers especially in the light of much higher prices farmers could obtain in neighboring countries. In their discussions with Government officials the missions found that due to the highly political nature of the official pricing policy a high level dialogue between the Government of Malawi and the Bank would be required. RMEA informed Bank headquarters accordingly, but no action was taken by the Bank until recently. Discussion with the Government during the negotiations for the recent National Rural Development Phase I Project (Credit 857-MAI) result in a covenant for developing a national agricultural price policy- . 29. This experience suggests the Bank might find it worthwhile to give increased attention to country pricing policies in the course of lending wherever projects, justified on the basis of economic prices, may be adversely affected by prices actually paid to producers. 1/ J. C. Mills, "Price Responses of Malawi Small Holder Farmers - Fast, Slow or None?" Occasional Paper No. 2, University of Malawi, 1975. 2/ "In order that marketing and pricing policy can be fully considered, a statutory or official pricing committee would be established at the commencement of the project consisting of members from Treasury, MANR, ADMARC and the Economic Planning Division of the Office of the President and Cabinet. The committee would keep prices under review and submit a report to the Board of ADMARC for consideration by the Minister of Agriculture. Within 18 months of project commencement, Government would review marketing and pricing policies and prepare a report, in consultation with IDA, defining the marketing and pricing criteria and policies Government proposes to adopt. The pricing committee would review annually as part of its price and marketing analysis the production targets and achievements of NRDP. It would comment on the efficacy of its pricing policies in meeting these targets and the committee would communicate these comments through the Treasury to IDA. An assurance on these points was obtained at negotiations. ADMARC would appoint economists at its headquarters to deal with pricing, economic and marketing issues. To improve marketing and input supplies ADMARC would staff its headquarters, markets and stores with liaison personnel in sufficient numbers to meet the input and marketing needs of NRDP development areas. Assurance on these points were also obtained at negotiations." (Appraisal Report para. 6.04) - 10 - Project Staffing 30. Throughout project implementaion it was difficult to employ qualified staff, especially Malawian nationals. The effectiveness of the credit was delayed by 3 months primarily because a qualified project manager could not be found. The project manager hired was an expatriate. Part of the problem of employing staff was likely due to remoteness of the area, at least in the beginning of the project. Another reason was that positions in the Karonga project as well as other similar projects in Malawi were not in the system of Government Civil Service "established posts". Personnel in "non-established posts" do not have tenure or employment rights as in the Civil Service. When the Karonga project is completed the employees have no assurance of continuing employment with the Government. But the primary cause was the general lack of qualified staff in Malawi to meet the overall demand of the Bank's projects, other development projects, regular government service and private sector. 31. This problem could have been avoided or at least reduced if staffing had been looked at from the view point of the entire sector. The Bank ought to be making forecast of personnel needs of projects coming up for appraisal to determine if persons with adequate training will be available throughout the sector to meet both public and private needs. If a gap is likely to exist then a training program should be initiated and the project delayed until a sufficient number of trained personnel ar ?available. This problem, too, was addressed in the apprai- sal of NRDPI-- . However, as just noted a better procedure than making the staffing study a part of the project would have been to have the study undertaken before the project was appraised and used the results as an input into project design. Conclusions 32. When a project is aimed at one sector - in this case agriculture - and also includes a small component of another sector - in this case transport - the minority component may be neglected. Project preparation 1/ "Government would carry out by December 31, 1979, a study to examine and make recommendations on the present staffing levels and work programs of MANR to improve the agricultural extension services; the study would be carried out by consultants appointed in consultation with IDA and using Terms of Reference agreed to by IDA. Within 6 months of completion of the study Government would discuss with IDA the implementation of its conclusions" and "further expansion of the NRDP and other MANR development projects would be phased in such a manner that they would allow the full staffing of NRDP and other major projects to be fulfilled before new development undertakings were started." - 11 - for the lake transport component came too late and this has substan- tially contributed to the implementation problems. It is not clear whether more supervision could have helped project implementation. 33. The problems with the barge are surprising. It appears that between qualified consultants, experienced shipbuilders and World Bank supervision, it should have been possible to produce a small vessel which would have been suitable for lake navigation, especially since on several occasions the problems of high winds and waves had been mentioned in Bank project preparation reports. As a result of this experience, project staff familiar with this particular case are now inclined to involve Lloyd's during the design stage of vessels. 34. It appears desirable that in future appraisals of similar projects, road components should be clearly identified, giving mileages, cost data and standards. The report should also indicate whether the roads are an indispensible part of the project or to what extent transport could use existing facilities. 35. The project was successful in increasing crop and livestock output and improving health facilities. The economic rates of return on all agricultural components except irrigated rice are estimated to be high at this stage of the project. Lessons learned from the irrigated rice scheme have influenced the design of the Phase II project; it does not have an irrigated rice component. Moreover, the lack of farmer participation because of inadequate price incentives (in this as well as other projects in Malawi) has caused the Government to change its pricing policies and to increase its purchase prices of cotton and groundnuts,,although still not fully consistent with domestic and world market levels.- 1/ As it stands the audit does not agree with the Borrower's comment (PPAM Annex I, para. A.2.21) "Certainly farmer participation in the irrigation schemes (Lufira) had nothing to do with price." If this statement means that price was a positive influence because it was subsidized, yet, because of other factors, did not prevent failure of the component then the audit agrees. - 12- ANNEX I In reply please qouteNo . . . Telegrams: FINANCE, Lilongwe MINISTRY OF FINANCE Telephone: Lilongwe 31311 Communications should be addressed to: P.O. BOX 30049 Tho Secretary to the Treasury LILONGWE MALAVI 31st May, 1979. Mir. Shiv S. Kapur, Director, Operations tvaluation Department, World Bank, WAShiiiiGTUN D.C. Dear Sir, pXOJ,vCT PRFUR14RMANCE AUDIT RhPORT ON MALAQ'S KARONGA .HURA.L DEVELOkMiEN'P PiiOJLCT - PHASE I (CREDIT 282 - MAI Please refer to your letter of Yebruary 16, 1979 on the above subject for which you sought our comments. The delay in replying to your letters is very much regretted. On the basis of the original expectations and assumptions, the Audit report presents a balanced and fair appraisal of Phase I achievements, which in general are commendable. We also note in some cases yield figures from Evaluation Section sources are -mis-quoted. Regarding the cost overruns on the Lufira Irrigation Scheme the Bank should have anticipated these problems because of the inadequate data-base at Lufira, in which case a less ambitious pilot scheme would have been ideal in the first phase with contingencies for further developments thereafter. Similarly the Lake Transport Component was also inadequately appraised and despite an over- estimate of requirements, the original proposal for a 200 ton self- propelled barge was not adhered to. Subsequent cost overruns due to inadequate design and consultancy work have been borne largely by the Malawi Government and have necessitated the modification of associated development, notably abandoning the Chipoka Terminal during the first phase. Roads development proposals were also not clearly stated in the appraisal report, inspite of the general acknowledgement that infrastructure was going to be a major component of the project. It needed to be defined clearly as to who was going to construct the roads and maintain them. This resulted in the total neglect of the roads. Detailed comments on the various aspects of the audit report and completion are attached. Your, faithfully. ki !Magalasi. for: UkORiTARY TO THE TREASURY - 13 - A. AUDIT REPORT 1. Summary and Highlights: 1.1 Paragraph 2, 4th sentence is not consistant with paragraph 5 of the same summary. Paragraph 2, for instance attributes the delay for the project to failure by Government in submitting to the Bank a satisfactory plan on time to improve Lake Malawi transport services? and delays in appointing the project manager. Paragraph 5, on the other hand states that delays were due to slowness of the Government in appointing a project manager, lack of staff housing, and difficulties in recruiting senior staff. 1.2 Paragraph 3 in the summary of the Audit report and 1.02 of the background to completion report, the phrase that development in the area of Karonga were constrained by the area's isolation from the rest of the country by a virtually "impassable mountain range" appears to be an overstatement. It is correct to say that communication by land were difficult due to rugged relief in the early stages before all-weather roads were proved feasible but it is not correct to say that the mountain ranges are impassable. 1.3 Paragraph 4, of the summary under the Project and 4.01 of completion report seems to portray the impression that the Project Management Unit -,,as established to implenent both agricultural and health components of the project. It needs to be made clear that the Project Management under the diroction of Ministry of Agriculture and Natural Resources, was responsible for only the agricultural component. The Health Component, like the Lake Services Component under the Ministry of Transport and Communications, was the responsibility of the Ministry of Health. 1.4 Paragraph 9, sentence number 4 of summary of Agricultural Development, the name Spear iterprises should read Spearhead Enterprises. 1.5 The first paragraph on page V of Bighlights needs to be re-written as follows :- "Karonga is the fourth rurl development project to be initiated in Malawi, and the third to be supported by the Bank financing; the other two Bank financed projects being the Lilongwe Land Development Programme (Phase I and II), and the Shire Valley Agricultural Development Project. The Central Region Lakeshore Development Project was financed under German Technical Assistance. Since the Karonga ...... approved by I. D. A." 1.6 The second paragraph should include improved extension services, water supplies after improved credit. 2. Major Issues 2.1 Livestock: It is noted in both the Audit report (paragraphs 18, and 19 pages 9 and 10) and Completion report (paragraph 6.14, page 46) that despite a herd increase from 2.4% to 6.5% (against the expected 3% per annum), off-take has not risen from 10% to the projected 12%. - 14 - Paragraph 18 of the Audit report also states that information on livestock, livestock marketing, and producers' willingness to partici.ate in the project was lacking. It is necessary to take into account the uncertainty in livestock numbers prior to project, and the scanty or inavailability of reliable statistics have obviously contributed in leading to the misinterpretation of the statement. Prior to projects, av able figures showed off-take rates of 12.2 % in 1970 . It is th3rafor at the statements quoted are too generalised, as they are based on phase I trend, rAther than on the annual increments. 2.2 The mean pre-Project % increase in livestock numbers was a reasonably uniform rate of 2.4. Throughout Phase I, mean % increase was 6.7 essentially as quoted. Incremental increase varies over Phase I as follows :- YEAR I INCREASE 1972-73 3.4 1973-74 11s4 1974-75 10.7 1975-76 1.2 2.3 From a relatively uniform rate of increase at a round 2.4%, census figures show a steep rise between 1973 and 1975 from 32344 to 39862 head. Thereafter the mean % annual increase is 2.9o although in a rising trend.. It remains uncertain whether the sharp increase in the middle of Phase I is as a result of the dipping programme or whether it is due to improved data collection. The Completion report has mis-quoted the trend, however, the fact that off-take h-.s not risen may suL-,est that improved data collection may have been mainly responsible for the apparent increase. 2.4 The alternative argument is th.it, while animal health was !,romoted, marketing was not. A current Evaluation Section livestock survey sug ests that livestock ownrs tend not to display commercil attitudes, and while they are eager to increase their herd sizes, partly through improved animal health, they do not have corresponding intentions to increase off-t%ke. Traditional attitudes prevail, and apparently increasingly so with increasing herd size. 2.5 It is felt that as a marketing component was lacking during Phase I, this has, in part, been resoonsible for off-take remaining at pre-Project levels. The extent of village marketin': is still unknown. 2.6 In comparing livestock figures in Karonga and Chitipa districts it should also be noted that in Karon-a district, population densities (in terms of both human and livestock) are high causing more pressure on land. And yet throughout phase I, little progress was made in finding a livestock system which would integrate with the cropping pattern on the Lakeshore areas. Lufira Irrigated Rice Scheme 2.7 In the Appraisal Report, it wAs acknowledged that there was a lack of hydrological data, but no shortage of settlers !as anticipated (Appraical Report P.19, para.6.09). Not only was the physical appraisal inadequate, but farmers response to a system involving a radical change in attitudes and life style 1:as not considered. The subsequently poor farmer response impinges on the re-consiO-eration of proposals to exact a rental charge in irrigated (and. rainfjd) rice schemes. - 15 - 2.8 In as far as the physical implementation aspects are concerned, it is certainly agreed that nearly everything went wrong with it. The problems were technical, administrative, logistical, including natural misfortunes. Floods of uncom.-ion magnititudes destroyed much of the headworks, impounding a weir in March, 1976. These had to be reconstructed, and hence raising the costs of construction further. The problems associated with the Lufira, diverted attention of resources (in terms of human and capital) towards Lufira Construction. This had obviously a bearing on the project's impact on other components. 2.9 It has been stated in paragraph 3.14 page 25 of the Project Completion report that the final cost of astablishing the Lufira Irrigation Scheme was approximately VK1034 per acre, when the appraisal missions had estimated the development costs at iT185/acre. This figure quoted by the Completion report is of course high if one compares it with the Appraisal report figure. If one is to include overheads but no charge by the Irrigation Branch, the right figure is MK373 per acre. According to present day prices, the costs of develo;:ing an Irrigation Scheme comparable to Lufira are estimated to be not less than MK1600 per acre. We feel therefore that last paragraph under Irrigated rice (Completion report, page 25, para.3.14) beginning with : "In retrospect the combination of poor farmers response ..... does not appear feasible" is a pessimistic and subjective view or generalised opinion. 2.10 Para. 20, on page 10 of the Audit report should read Irrigated rice and not..;riated prime &Farfors.reTpoe%t tbe costly. Irrigation Scheme, it is said in paragraph 3.13, oage 25 of Completion report that the (farmers) response to the costly Lufir, Irrigation Scheme -as disappointing. One of the reasons .. outlined for this disappointing response is that they prefer3d growing the less labour intensive Faya rice. This in part is probably true but it should also be noted that to the rural farmer, irrigated scheies entailed an intensive and a highly disciplined form of agriculture. Because most resources were diverted toward Lufira, the extension service iias unable to inform and prepare local settlers for this type of intensive agriculture. And also the lack of consumer goods in Karonga, at the time, were not adequate to provide an incentive to the farmers who wished to produce. At some stage, a lot of farmers who produced high yielding varieties were known to have re;ceived more than adequate income that they did not want to continue producing because the compensatory consumer goods were not available. 2.llIt is understood that at .resent, the performance of local settlers in growing irrigated rice is good and local farmer response (which was disappointing at first) has been gradually building up. Soon br later this is likely to exert pressure on the land given to Spearhead nterprises. This is an encouraging indication on the part of local settlers, and goes to show that given the right advice by &tension and Agricultural services, local settlers are not adverse to change. 2.12 In so far as the water resource is concerned, there is enough water in the Lufira to irrigate the full acreage during the summer season, however, the -creage which could be brought under rice cultivation during the winter period may vary from 250 to 500 acres*This suggests diversification of the winter crop to maize, wheat, vegetables (beans) and other upland croos. The Spearhead Enterprisas grew 25 .,.cres of maize in 1978, utilising mainly residual moisture with one or two light supplementary irrigation. Such additional occupations should provide extra food ani incomes to farmers. - 16 - Transport - Roads Construction 2.13Paragraph 24, page 12 of the Audit Report and p-ragraph 3.05 page 23 of the Project Completion report only highlights the confusion that existed at the time this particular component was to be implemented. If "according to Bank Staff it was planned that the project unit would build and maintain the roads in order not to overstretch or over- extend the resources of the Ministry of 'orks (MOW)", no maintenance funds were included in the Phase I for this purpose. The statement that the project was going to maintain roads is therefore doubtful (paragraph 24 of the Audit report, page 12). 214 What was even worse, the physical construction programme for roads was less straightforward in Appraisal Report, and the programme was affected by the rapid escalation of costs and the concentration of resources at Lufira. Until the end of Phase I, there was still confusion about how many miles of roads were built, where and to what standard. According to original proposals 34 miles of roads were to be constructed in rice schemes, 37 miles dryland, 83 miles for livestock, and 18 miles all to be constructed by Ministry of Works and Supplies, including Chilumba spur. This programme was modified it Appraisal to 16 miles at Lufira, 7 miles at Wovwe, 37 miles dryland, and 58 miles for Livestock, a total of 119 miles (paragraph 2.01, item (b). The major works for MIinistry of Works was removed. Omitted from the plan was also the expected road length in the rainfed rice schemes. 2.15Total specified length of roads at 119 miles, had varying costs ranging from MK200 - MK1800 per mile for upgrading and new roads respectively. A programme which 1as arringed for 1975 for about 120 miles, incurred cost overruns : gravelled roads cost DK4500 per mile, ungravelled at DK500. 2.16During Appraisal agreement was roached that all roads would be maintained by hinistrv of Works and this was of course on the assumption of satisfactory construction of roads according to Tinistry of Works standards. Ministry of Works take over road which only conform to specific standards. Most of the roads constructed during Phase I were of low standard and Ministry of Works were not consulted during their construction, and therefore not obliged to take them over. The low standard of roads, most likely also affected input flow and extension services. 2.17Lack of co-ordination at the time between the Project and Ministry of Works resulted into poor maintenance or total ne-lct on all of Project roads. The reconstruction costs referred to in the Audit Report (P.12, para.24) might have been avoided if initial road construction had been carried out with full consultation with Ministry of Works. ADMARC'S Pricing Policy 2.18Paragraph 25, page 12 of the Audit Report and paragraph 6.09 to 6.12 pages 44-45 of the Project Completion Report makes mention of ADMARC price policy penalysing cotton and groundnuts farmers in Karonga Project area (also table 1, page 13 of the Audit Report). This has been debated and discussed in many circles particularly in situations of cotton, which unlike groundnuts, farmers hive no opt4P but to sell to ADPARC. It will be noted that ADMARC increased the price of goundnuts for 1978/79 season by 50%. It remains to be seen if farmers will produce and sell more groundnuts this year or the year after. If farmers are highly dependent on ADMARC's price incentives, we should expect more groundnuts to come for.ard, and this should be evidence enough to seek policy directives. This should not be interpreted as a sign of doubt that Malawian f,irmers are not responsive to orice incentives. - 17 - 2.19 It should also be noted that ADMARC trades in a variety of crops. Based on present day World Market prices, ADIIARC incurrs losses on maize, rice and other food crops (in fact 1rodipor prices for rice are above World Market prices). The profitre 7 2. on other crops like tobacco, cotton, groundnuts is taken to Cross-subsidise the losses on rice, maize, This obviously undercuts producer prices, which are considered by ADMARC overboard instead of by individual crops. It will be uneconomic to ADMARC also if they were to uay producers unrealistic prices because the international market is subject to wide fluctuations - a common feature to primary commodities. Hence ADMARC is bound to move cautiously in order tot to disrupt Government policy of stabilizing prices to smallholders, and at the same time maintaining its dual role of development. 2.20Studies are continuing however in the field of pricing, under the National Crop Pricing Co 'mittee, to which Ministry of Agriculture and Natural Resources, Treasury, &onomic Planning Division, and ADAIC are members. 2.2lLast sentence in respect of paragraph 35, page 17, of the conclusion of Audit Report is misleading and not clear. Certainly farmer participation in irrigated schemes (Lufira) had nothing to do with price. In case of dryland crops, see para. 2 (b) below. B. COMPL-ETION REPORT 1. Formulation : Paragraph 2.01, the word region should be substituted by district. 1.1 The footnote at the bottom of paragraph 2.01 of the Completion Report states that "In 1970 the Karonga area contributed about 35% of total marketed rice production in Malawi. With the project, it w-,s expected to maintain this sharo." In spite of the setb.ick with the Irrigation Component, this level w:,s maintained except for 1972. The following figures show this trend Rice Production in Short Tons 1972 1973 1974 1975 1976 Karonga District 6324 6993 8291 5952 10071 All Malawi 21859 19400 23475 15354 27275 Percent of Total 290 36% 35% 39% 37% Also the last santence of the same footnote, the last part ' Cattle production of 510,000 head' to reAd ' Cattle Loulation of 510,000 head. 2. Implementation : Revision of Targets: 2.1 Paragraph 3.02, last sentenco. The project completion date for Karonga Project Phase I, w!s 31st t'arch, 1976 and not Septamber 30th 1976. Phase II of the Project, in fict started about September, 1976, after six months delay. Crop Development - Scheme Approach (Dryland Cro.s - Cotton and Groundnutsl: 2.2 It may be worth noting that the proposals concerning general develoomant of groundnuts, cotton, maize 'Jas expected with incresed seed distribution and improved extension. - 18 - The approach to be used -e not sp3cific in the Project Appraisal and it was taken for gr;nted that this would involve traditional extension methods. As correctly indicated in para,raph 3.10 the Project initiated a new approach of dryland schemes, similar in management and organisation, to rainfed rice schemes. This approach which is unique in Malawi, was in our view implemented satisf.otorily and credit is due to Project effort. As a result of the scheme, expansion of areas was rapid and exceeded appraisal targets (Crop area in total increased for maize, cotton, groundnuts). Maize was the dominant crop. 2.3 The main rea.son for the low rate of expansion in groundnuts and cotton is this th-t although the development of these crops were to follow a scheme approach, which had the support of this Ministry, the World Bank Supervision mission and local community on the understanding that no extra costs were to be incurre., thQimppovt4xnity--by- t projct.7to' fally db@el4he 4hai .wbk.ot :presting and al1ocligohe labies 20t. INAVy 'taska!'Tor bush olearance were completely left to the farmer himself, and farmers lost inter.st. 2.4 The advancement of credit to scheme farmers w s one way of encouraging farmers to follow scheme approach, but certainly credit was not denied to non-scheme farmers. 2.5 Table 2 of Completion Report, paragruph 3.16 : The actual figure for the area planted to maize, 7roundnuts and cotton in 1972/73 ind 1973/74 should be 260 and 1529 acres (Table 3, Annex 2 of The Implementation and Effects of Phase I, Karonga Rural Development Project) and not 114 and 1237 acres aT indicated in that table 2(refer also to para,.-ra:,h 2.3.4 of the same Project Report on the Implementation and .ffects of Phase I - Karonga, April, 1974. The breakdown by crops for 1972/73 was 120 acres maize, 120 acres cotton and 20 acres groundnuts. A breakdown for 1973/74 is not available. In the same paragraph 316, the planned figures of 3500 acres for maize and cotton, 800 acres -roundnuts '-ot agree with those in table 2 of the same page 27, of the Completion RepoMfas 2.6 The discrepancy between cotton yiLlds quotd in the various reports is due to the common prictice of leaving cotton unpicked in the field and losses incurred between the grower and the market, estimated t 1 of production (@valuation Section Sup ,liment to Working Paper 1)77/7). 2.7 These losses hive been taken into account in the Completion report, i.e. are based on marketed cotton rather than actual yields. 2.8 &ceot where they are dependent on Admarc seed, groundnuts farmers commonly sell their produce in the lucr;.tive local markots. Health (Paragraph 3.18 on page 2 of the Project Completion Report). 2.9 The recommended posts of a Public Health Officer and Public Health Nurce to Karonga -ere complied to by Government by postipg to the area. a Public Health InqPgg-obf T.0. grade and -Hoalth AsAidtant bf tA grAde in August, 1975. Credit and Input Use (Paragraph 3.19 of page 28 of the Completion Report). 2.10It is felt that credit should have been quoted more fully in the Audit Report and highlighted as a major issue. The credit mechanism is a fundamental Project instrument in promoting stated objectives. Despite the high level of farmers response, seasonal loans ,ore 50% of appraisal and medium term loans were only one third of the expected (Audit Report, page 7 para. 13 and paragraph 3.19 of Completion Report). - 19 - The effectiveness of a credit facility hinges on farmer confidence, based on an understanding of the benefits and management implications, through extension and training, and faith in an efficiently oper.ating system. 2.11Although in the Karonga Project production proposals relied heavily on farmers using a range of improved seed and fertilizers, and an expansion of Ox-cultivation, the Credit Section of the Project was without a Senior Officer throughout the phase as has been correctly noted. The credit unit being unitially under Extension, was later transferred to Finance, with the arrival of the Financial Controller. Responsibilities were split between credit and extension, the former's role being basically administrative - keeping records, authorise disbursements, and collect bad debts - while that for extension was to provide supervision over farmers inputs. 2.12There was also initial uncertainity as to who, between ADMARC and the Project should distribute inputs to farmers in the Project area. Although ADARC agreed to supply and distribute inputs, this service was discontinued between 1973/74 and 1974/75 because ADRAIC did not have adequate funds. Project lorries had therefore to be deployed from construction work and transport for the purpose. This hiatus continued into Phase II of the Project (in 1977, when ADMARC agreed to resume distribution of these inputs). Disruptions in distributions were not obviously conducive, to the effecient running the credit system. 2.13Seasonal inputs were to be available for both rice and drvland farmers, medium term credit to cotton sprayars, and oxen and ploughing implementr for irrigated schemes only. During implementAtion thise restrictions were not adhered to. The range of medium term inputs were expanded to include ridgers, ox-arts, etc. and these were made available to all scheme farmers and not only for cotton crops as indicated in paragraph 3.10 of the Completion Report. 2.14As correctly indicated in Completion Report, confusion also arose in regards to input mixes. There was no clear credit policy to control input mixes and there were reductions in seed rates for rainfed rice and maize, but increases in rainfed rice fertili-ers. Insecticides and fungicides were not used at all on irrigated rice schemes. Although such changes are inevitable as data about new area builds up but the situation led up to some confusion between farmers and extension field staff. Hence production was most likely alfected, as farmers were able to take seed and fartilizars in whatever quantities or ratios they desired. A package apiroach for the two rice crops, maize and cotton 1.?as tried in 1975/76 but dropped because it was unpopular withArmers. 2.150f equal importance were the effects of service and handling charges for credit. During the initial stages when ADMARC distributed inputs the project levied a service charge of 11% in addition to the 10% handling charge by ADMARC - a total of 21%. Although the other 11% service chargo was covered by inputs bought in bulk from ADFARC which attracted a 10% discount, this was discontinued in 1974, with increased fertilizer prices. This resulted in the restoration of the fall service charge in 1974/75, which had an adverse eff.ot on farmer adoption until this was dropped with I.D.A. approval. The rise in fertilizer prices and increased service charges to 21% in our opinion would appear to have had more effect on low credit activity and not so much the rise in intere2t charges from 10% to 12FI, (refer to paragraph 13 of Summary of the Audit Report). - 20 - 2.16For dryland crops, the uptake of groundnut seed exceeded appraisal estimates for first three years but was then cut to zero in order to combat the border trade with neighbouring countries. Compliance with Covenants: 2.17Paragraph 3.22 and 4.02 of the Complktion Report raises two issues concerning the formation of Liaison Committee to co-ordinate the activities of the Ministries involved with the Project, and the levying of develo!.ent charges to participating farmers on irrigation schemes, rainfed rice schemes, and othar areas. In case of Liaison Committee to co-ordinate interministarial activities of the projects, this was infact formed in 1972, as a Projects Liaison Committee for all major projects i.e. to cater not only for Karonga project but also Shire Valley and the Lilongwo Land Development. The Committee had five to six m3etings ,the last one being on 11th December,0f974 or 27th March, 1975. The Committees were not continued because it was felt that such Committees were duplicating the efforts of the Project Monitoring Committee although the latter also ceased operating. Representation on Projects Liaison Committees were Ministry of Agriculture and Natural Resources, Project Managers, Ministry of Works and Supplies, Ministry of Health, Ministry of Local Government, ADMARC, Ministry of Community Development and Social Welfare, Development Division, Town Planning Department and Treasury. 2.18The decision not to charge levy was a result of policy changes which affected not only farmers contributions but also cattle grazing and dipping fees. It has baen felt by Government that service charges or levies would be a disincentive to production or farmer participation and hence these are politically unanceptable. The proposed National Irrigation Study proposed under the Shire Valley Project(Phase III) will certainly re-examine the issue. 2.19 In.1he last sentence of the same paragraph, q.b : 79Psqation of levies in rice scheme areas has been sug asted (Pago 30, para. 3.22 it will also.be;enotd :.thataes wx first- QA,lined Ja *ha Apprais-al Report, (Annex 10): MK7.00 per acre rainfed rice MK14.00 per acre irrigated rice plus : .20 per 160 lb. bag of rice sold to Admarc. 2.20 Whil(i.t tmip baconside-red ..to be taqhniql)y feasible to implament these- now,thenet returns quoted in the Appraisal report are felt to be unrealistic, although the inclusion of actual production figures in the analysis also does indicate that such levie- would be viable in relation to net teturns froc other crops. An analysis (Ovaluation Section Workin,7 Papfer 1976/4, " crop budgets for rice, maize, groundnuts and cotton, for the season 1975/6")beirs this out. It is felt that a further more detailed analysis would be required, particularly with regard to the disincentive effect on irrigated rice farmers who have already displayed a seasonal preference for crops outwith irrigated areas. Certainly, the imposition of a levy at this stage, is likely to be met with resistance. Marketing 2.21 Paragraph 4.06 of the Completion report. The Malimba variety, like Manipint,-q is not a Confectionary nut. It is an oil nut. - 21 - Finance: (Paragraph 5.02 - Finanoial performance, page 33 of Completion Report). 2.22An examination of the I.D.A. Credit disbursement for each category shows expenditure increases mobt particularly affected vehicles and machinery, operating costs ind maintenance and the construction element of the project. Funds were re-allocated from other Components to meet these shortfalls. 2.23 Paragraph 15 of the Summary on Financial development, and 5.02 of the Completion Report states that U.S. #397,000 is undisbursed or unallocated, and that a decision on the use of these funds has to be made by Government. It will be noted that credit No. 282 MAI is now fully disbursed. The Application No. 186 dated May 19th, 1978 was cancelled for the balance of $10053.95 and substituted by a fresh application for the same amount under Loan number 1286 MAI (re World Bank letter dated 25th April, 1979, signed by Dietrich Von Busse of RMEA). The reasons for the balance were communicated to Bank. Prices 2.24 (Paragraph 5.06 and table 7 on page 37 of the Completion Report) Producer prices for cotton and groundnuts should read as follows$- t /lb Crop 1972/73 1973/74 1974/75 1975/76 Maize 1.25 1.25 1.75 2.25 Cotton 6.0 7.0 8.0 8.5 Groundnuts 6.5; 5.0 7.5; 6.5 8.0; 6.5 8.5; 7.0 Paddy Rice 3.33 3.33 4.0 4.5 2.25 Cutton prices refer to grade 1, the topmost grade, while those for groundnuts refer to Chalimbana (GDA), and Malimba/Mampintar (Shelled). Paddy rice prices are also those of grade 1. 2.26 Paragraph 6.09 of the Complation Report : ADMIRC does set producer prices but not farmgate prices in the true sense of the word. 2.27 Farm Budgets and Croping Patterns: (Paragraphs 6.05 and 6.06 of Completion Report). 2.28The appraisal reports farm budgets (Annex 10) are not matched by the majority of Project farmers, as the yields were inflated and assumed high standards of management. In 1975/76, 16% of irrigated rice (scheme) farmers, 57% of rainfed rice (scheme).farmers and 63% of maize (scheme) farmers used no fertilizer (Rraluation Section yields studies 1975/76). 2.29 Crop acreage have exceeded targets in all crops except cotton and irrigated rice, (Complotion Report, Page 27, para. 3.16, table 2), and a greater number of farmers than anticipated took part. The expansion drive has not been matched by management improvements. Seemingly the supposed response of farm.3rs to improved methods of managements has been over-estimated. This may be in part due to a characteristically conservative tttitude in the agricultural section, but may also reflect on the effectiveness of the Project Phase I credit facility. - 22 - Inflation 2.30 (ImpAct on project costs indicated on the Summary of the Audit Report) 2.31 It should be noted that inflation in the general level of prices subsided to 3.3% by 1976 in developed countries as measured by the GDP deflator index in United States dollars. It was highest between 1970 and 1975 - ranging between 10% and 14%. The Karonga-Chitipa Rural Development Project was implemented in this period. The index of intetnational inflation (a meagure of changes in purchasing power of developing countries' exports of primary commodities imports of manufactured goods) rose even faster than the GDP deflator index after 1973 because of the extra-ordinary strong increases in international demand for machinary and equipment. The arguments that cost increases in the project rose faster than the general rise explained by inflation, is not clear. Planning Division MINISTRY OF AGEICULTURE AND NATURAL RE3OU hCff - 23 - MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I Project Completion Report I. BACKGROUND 1.01 The Karonga Rural Development Project, Phase I (KRDP I), was designed to assist agricultural development in the Karonga area in Northern Malawi, and was part of the Government's strategy for improving agriculture through an integrated regional development approach. The Project was financed by IDA, and was the fourth IDA credit in Malawi for such development, the other three having been for projects in the Central and Southern Regions (Credit 113-MAI, 1968 (LLDP, Phase I); Credit 114-MAI (SVADP, Phase I) 1968; Credit 244-MAI, 1971 (LLDP, Phase II)). 1.02 The Northern Region was considered to have good development potent- ial for a large variety of crops, but was isolated by a virtually impassable mountain range. Infrastructure was very limited. Lack of prop:r road links with the rest of the country necessitated continued reliance on lake trans- portation. Health facilities were inadequate, and diseases such as bilharzia and malaria were serious problems. 1.03 Accordingly, in 1970, the Malawi Government, assisted by the Perma- nent Mission in East Africa (PMEA), prepared and submitted a proposal for a 5-year integrated project, which was appraised by an IDA mission in April 1971. The Project proposal submitted by the Government included the construction of irrigation works on about 2,700 acres, involving an extension by 700 acres of an existing irrigation scheme, and construction of a new scheme to cover 2,000 acres; the purchase of four additional lake vessels and construction of a new district hospital. Because of inadequate hydrological data, however, the appraisal mission recommended that only an initial 1,500 acres of rice (500 acres of extension to the existing Wovwe scheme and 1,000 acres of a new scheme) be developed under irrigation with provision for expansion at a later stage. The appraisal mission also noted that improved port facil- ities would allow the addition of a single vessel to provide adequate in- creased shipping capacity, and that improvements in the existing district hospital would be sufficient to meet the Project-related health problems. The other parts of the Project-rainfed rice, dryland crops and livestock production, health and infrastructural components - were included more or less as proposed by Government. 1.04 Total project costs were estimated at US$7.8 milion, and the approved credit for US$6.6 million was to finance 85% of the project costs, covering foreign exchange costs of US$4.1 million and 69% of local currency costs. The balance of US$1.2 million on local costs was to be provided by the Government. The Credit Agreement became effective in August 1972 and the closing date was December 31, 1977. - 24 - II. FORMULATION 2.01 The objectives of Phase I were to accelerate agricultural develop- ment in the Karonga district and to improve infrastructure and health facil- ities in the district. In addition to providing benefits to participating farmers, the production targets of the project were in keeping with national policy. The intention was to prevent the region from becoming deficient in its main food crop, maize, to increase rice production in order to keep pace with a rapidly increasing domestic demand and still expand exports of rice, and to initiate the development of the North into an important source of beef for the country. 1/ The agricultural component of KRDP I was designed to cover about 10% of the arable land in the Northern Region (15,300 acres out of 157,000 acres), to involve a total of 4,600 farmers out of 18,000 farm families and to provide experience for future development of the Region. Specifically, the project components were: (a) establishment of eight rice schemes: two irrigated double-cropping schemes on about 1,500 acres, involving the settlement of about 830 farmers; and six organized rainfed single-cropping schemes on about 6,000 acres and involving the settlement of about 2,000 farmers. (b) improvement of dryland crop production of maize, cotton afid groundnuts on about 7,800 acres, involving about 1,800 farmers, through provision of inputs on credit and introduction of improved husbandry practices through extension services; (c) livestock development through construction of dipping tanks, markets and an 18,000-acre holding ground, and improvement of existing stock routes; (d) Provision of seasonal and medium-term credit to project farmers; (e) further studies on crops and hydrology, including a feasibility study for a second-phase project; (f) construction of five rural health outposts, improvement of Karonga District and Chilumba hospitals, and provision of staff and drugs for bilharzia control within the rice schemes; and 1/ In 1970 the Karonga area contributed about 35% of total marketed rice production in Malawi. With the project it was expected to maintain this share. The Northern Region had about 10% of the country's estimated cattle population of 510,000 head. - 25 - (g) provision of improved lake terminal facilities, at Chilumba in the north and at either Monkey Bay or Chipoka in the south, 1/ and the addition of a 200-ton self-propelled vessel for the lake service. (h) Improvement and construction of 119 miles of roads. 2.02 The improved husbandry practices to be introduced under the project were to increase yields by varying degrees, from 10% for irrigated double- cropped rice, to as much as 230% for cotton (see Table 9 for estimated increase in yields). Returns to the farmer were estimated to be about MK 83 (US$108) 2/ for a 4.5 acre farm of cotton, groundnuts and maize, and MK 279 (US$363) 2/ for 3 acres of double-cropped irrigated rice (cultivated with oxen). At the time of appraisal, the average farm family cash income was estimated at MK 6 (US$8). III. IMPLEMENTATION Effectiveness and Start-up 3.01 At the time of the Credit Agreement in January 1972, the project was expected to become effective in May 1972, but two of the thrt conditions of effectiveness, i.e., submission of a satisfactory plan for improvement of the lake service and official appointment of the Project Manager, were not met in time. As a result, an extension of 90 days was granted and the loan became effective in August 1972. Due to lack of staff housing and senior staff, the project's physical start took place only in January, 1973. After that, however, the execution of the project quickly gained momentum. Revision of targets 3.02 Soon after the start of Phase I it became apparent that a revision of the appraisal plan was necessary. By the end of 1972, the proposed cattle markets and the 18,000-acre holding ground were being reconsidered because cattle sales at existing markets were not as great as anticipated. In 1973 the health component was revised to fit in with the national development plan for health services. In addition, a combination of price escalation on vehicles and equipment and large cost over-runs for construction of irrigation works required a revision of the project. Accordingly, and to keep within the original appraisal cost estimate, the Government decided, with IDA agreement, to shorten Phase I by one year and to delete the following components: 1/ The choice of the southern terminal was not made at the time of nego- titations, but the Government wanted the site to be Chipoka, and this was finally agreed to by IDA in July 1972 (para 3.08). 2/ In 1970 terms. For comparison with actual returns in 1975 terms see Appendix 4. - 26- (a) 500-acre extension to the Wovwe irrigation scheme; (b) 4 cattle markets and 18,000-acre holding ground; (c) improvement of Chilumba rural hospital; (d) improvement of Chipoka jetty at the southern terminal. The Project completion date was advanced to September 30, 1976. There were no revisions in the rates of return. Physical Construction 3.03 Buildings. After a delayed start-up (see para. 3.01), physical implementation of buildings gained momentum, and proceeded practically as scheduled. Staff houses, health centers, workshops and offices were con- structed as planned, and costs of construction were close to appraisal estimates. 3.04 Irrigation Works. Construction of the Lufira irrigation scheme was considerably hampered by a number of difficulties, such as absence of senior technical staff, division of responsibilities for construction bo-ween the Irrigation Department and Project Management Unit, need for changes in design and consequent escalation of unit construction costs. As a result, construction work was not completed until 1975/76, two years later than planned. Cost overruns were enormous (MK 827,000 actual costs compared to MK 185,000 estimated). 3.05 Road Construction. Road construction by the project apparently exceeded appraisal estimates by about 10%, although there has been confusion on the exact number of miles and location of the roads (Table 1). In addi- tion, the question of maintenance of the roads is still unresolved, although at negotiations it was agreed that the Ministry of Works would assume respon- sibility. The Ministry has been reluctant to take over roads which it considers substandard, and as a result, up to the end of Phase I, the project maintained its own roads. The Project Staff is now planning to upgrade the roads to a standard acceptable for maintenance by MOW, and any further road construction will be done in consultation with an engineer from MOW. This problem of road maintenance could have been avoided if, at the outset, there had been consultation between MOW and MANR. Table 1: ROAD MILEAGE As at end of Appraisal September 1976 (Project Completion) Gravelled 31 20.5 Ungravelled 88 111.7 Total 119 132.2 - 27 - 3.06 Port construction and lake transportation. During the Project period, the facilities at Chilumba port were improved by construction of a jetty and handling facilities, but at considerable cost overruns (esti- mated MK 306,800 at appraisal compared to MK 530,000 actual costs). These were due mainly to poor base cost estimates, which did not take into account the remoteness of the site, and to the escalation in the costs of transporta- tion, materials and equipment. Construction of Chilumba jetty was completed in August 1973. In 1977/78, however, higher than anticipated water levels required the raising of the jetty by about an additional three feet. 1/ At the end of Phase I, therefore, the jetty was in service, but undergoing reconstruction. 3.07 A 380-ton barge was purchased during Phase I. The rationalefor the choice of a barge almost twice the capacity of the original estimate was that the marginal cost of the additional capacity was negligible. Shortly after the barge was put into operation on the lake, in March 1976, its construction proved to be inadequate to withstand the weather conditions, and it cracked. The responsibility and subsequent cost of repairs was borne jointly by the consultants who approved the original design, the company that built the barge, and the Government of Malawi. 2/ At the completion of Phase I, in September 1976, the barge was awaiting repairs, and was back in 1peration by August 1977. 3.08 By July 1972 Chipoka had been chosen as the site for the southern terminal because the Government was interested in linking the use of the port to the Salima - Balaka railroad, which was to be rehabilitated. Although designs were prepared and bids were received by 1974, construction was post- poned in order to coordinate it with a nearby pulp paper project (Viphya scheme) under consideration, which would make use of the port. At the same time, it was realized that due to cost overruns on the Chilumba jetty and the anticipated cost overruns for Chipoka (the lowest bidder had submitted a bid for MK 884,000 compared to MK 365,000 estimated at appraisal), there were insufficient funds to construct the Chipoka terminal. It was decided to postpone further consideration of Chipoka until Phase II of the project. Crop Development The Scheme Approach 3.09 For crop development, a scheme approach was used, involving the demarcation of areas within which there was a reorganization and consolidation of land holdings. Reallocation of land and settlement of disputes was done by Scheme Committees, composed of farmers' representatives, village headmen and local political dignitaries. The Committee also assisted extension efforts. 1/ An additional MK 46,000 are to be allocated for this. 2/ The barge had three deficiencies: inadequate structural strength, poor trim and unsatisfactory welding. - 28 - 3.10 The advantages of the scheme approach are that it appears to be an effective means of land redistribution, of establishing farms with more homogeneous soils, and of allowing for more efficient use of extension serv- ices. The disadvantages are that it is a time-consuming process, and the selection of areas for inclusion in the schemes may involve a bias toward more fertile areas, and, therefore, the exclusion of less productive and possibly poorer areas. For dryland crops in particular, it is not clear that the scheme approach is justified since these schemes are not crop specific, and therefore some advantage of intensive extension work to encourage uniform cultivation techniques may be lost. The scheme approach was not visualized for dryland crops at appraisal, but was implemented by the Project Manager with the approval of an early Supervision Mission. It was stressed, however, that farmers outside the scheme areas interested in the credit scheme should also be serviced by the Project. It appears, however, that only in the case of the purchases of insecticide for cotton crops have credit facilities been made available to non-scheme farmers. Irrigated Rice 3.11 The initial proposals for irrigation of 2,700 acres of double- cropped rice at Lufira and Wowe were based on studies undertaken by the Agricultural Development Service of the World Bank (1967), a sutey of irriga- tion potential by Lawrence (1967) and limited experience on a few small schemes in the area. Because the hydrological data available were limited, the appraisal mission reduced the area to 1,000 acres at Lufira and a 500- acre extension to the existing Wovwe scheme, with provision in both areas for future expansion. 3.12 The technical planning of the Lufira scheme was the responsibility of the Irrigation Department in Lilongwe, while implementation was done by the Project Management in Karonga. It was hampered from the beginning by the absence of senior staff and by problems of design. The senior Field Develop- ment Engineer was not in post until November 1973. Changes in design were required and concrete lining, which was not originally thought necessary, was needed along 25% of the main canals to prevent seepage. The increase in design costs was compounded by large escalations in construction costs, and as a result, it was decided not to start the 500-acre extension at Wovwe. Facilities at Lufira were not ready until 1975/76, and the first irrigated crop of 265 acres was planted in November, the 'beginning of the wet season. In March 1976, the headworks and impounding weir were destroyed by flood. Repairs were completed for the 1976/77 wet season, by which time it was determined that there were 800 acres (as opposed to 'the estimated 1,000) available for cultivation within the scheme. 3.13 The farmers response to this costly irrigation scheme was dis- appointing. They preferred to continue growing rainfed rice (var. Faya) rather than the more labor and management intensive irrigated rice (var. Blue Bonnet) even though the latter variety received a higher premium. Thus, during the 1976/77 season, only 195 acres of the available 800 acres were cultivated. The Government, therefore, allocated 457 acres to Spearhead - 29 - Enterprises, the commercial wing of the Malawi Youth Organization,leaving 343 acres for individual farmers. The area allocated to Spearhead Enterprises does not come under the direct management of KRDP, although the Malawi Young Pioneers should be seen as important contributors to the production of the scheme. However, as techniques become established and acceptable to small- holders and demand for this area increases, it is not clear whether Spearhead Enterprises will release the 457 acres allocated to them. In addition, water control and management poses certain problems, since dry season river flows have been lower than expected at appraisal. It is now anticipated that sufficient water to irrigate properly the full 800 acres will be available only one out of every three years, and that once every five years there will be sufficient water to irrigate only 260 acres at the optimum rate. Clearly, strong policy guidelines from MANR are required to ensure fair distribution and efficient use of water, and the Bank should draw this to the attention of the Ministry. 3.14 The Lufira scheme was an important component of the Project, and considerable resources were devoted to it. The cost of designing and build- ing the irrigation works and canals at Lufira was MK 827,000 or MK 1,034/ acre, (about US$1,050/acre) compared to MK 185/ac 1/ estimated at appraisal. In retrospect, the combination of poor farmers' response and une-Kpectedly low water levels indicates that the Bank should have been more cautious in accepting the Government's enthusiasm for irrigation schemes based on too little experience and inadequate data. In addition, the Bank's technical assistance and advice at the design stage of the Lufira scheme could have been useful. Even now, provision of such assistance could be of use in reviewing the existing design, reexamining hydrological data and in evaluating how to maximize use of the facilities and land in the future, now that double- cropping with rice every year does not appear feasible. Rainfed Rice 3.15 At the time of appraisal, individual farmers were growing rain- fed rice using the Faya variety with few if any water control measures. There was also a Government-sponsored pilot project of about 60 acres. Yields of about 1.1 tons/acre were being obtained with use of some fertilizers. Rice schemes totalling 6,000 acres, involving 2,000 farmers, were proposed at appraisal and by the end of Phase I, more than 7,000 acres, involving 4,200 farmers, were being planted. (For details by year, see Table 2). Although expansion of the rice schemes was rapid, the conservation works which had been planned 2/ were not constructed. Implementation of these works would have provided invaluable information and experience in controlled use of rainwater in rice cultivation. This lapse on the part of Project Management 1/ Including price and physical contingencies. Excludes farmers' contribution, roads and other overheads (i.e., field staff housing). 2/ They were to consist of simple earthworks to control variable water conditions in flood plains during the rainy seasons. - 30 - may be partly attributed to shortage of staff, but the major factor was the overwhelming importance and priority accorded to the Lufira scheme, and this lapse occurred in spite of the greater economic benefits estimated at appraisal accruing to rainfed schemes. Dryland crops 3.16 The total crop acreage by the end of Phase I for maize, cotton and groundnuts exceeded appraisal targets, but the acreages of individual crops are different from that envisaged at appraisal. In 1975/76, 4,500 acres of maize, 220 acres of cotton and 970 acres of groundnuts had been planted, compared to 3,500 acres each for maize and cotton and 800 acres for groudnuts estimated at appraisal (see Table 2 for details by year). The diversion from appraisal estimates was most dramatic for cotton, and this is apparently due to the low returns per acre and per manday relative to the other crops (see Table 12 for details on comparative returns and para 5.06 for discussion of pricing). Table 2: SCHEME AREAS BY CROP, PLANNED AND ACTUAL (acres) 1/ 1972/73 1973/74 1974/75 175/76 Project Year I 2 3 4 Planned Actual Planned Actual Planned Actual Planned Actual Rainfed. Rice 750 - 2,250 1,514 3,750 5,261 5,500 7,065 Maize 120 54 600 320 1,200 2,128 2,400 4,496 Cotton 120 28 600 627 1,200 471 2,400 221 Groundnuts 20 32 200 290 400 319 600 974 Totals 1,010 114 3,650 2,751 6,550 8,179 10,900 12,756 Numbers of Farmers Rice 250 - 750 1,051 1,250 3,010 1,835 4,217 Dryland crops 60 60 300 449 600 2,180 1,200 1,825 1/ Source - Evaluation Unit, KRDP. See Borrower's comments (PPAM Annex I, B.2.5). - 31 - Livestock Development 3.17 The Project has met the construction targets of the animal health component of the Project. Between 1973 and 1976, 24 dips were built. Most animals are now within 5 miles of a dip and dipping attendance has increased from 55% to 80%, although not all animals are presented for dipping regularly at weekly intervals. Table 3: CATTLE DIPPING PROGRAM Karonga Chitipa 1972 1977 1971 1977 No. of % No. of No. of % No. of % Animals dipped Animals dipped Animals dipped Animals dipped Within 5 miles of dip 19,122 80 38,000 85 5,543 81 36,783 85 Within 5 to 10 miles of dip 5,117 37 2,800 20 6,398 11 - - Subtotal 24,239 71 40,800 80 11,941 43 36,783 85 Over 10 miles to dip 6,997 - 200 - 12,411 - - - Total 31,236 55 41,000 80 24,352 21 3 85 Health 3.18 Although the health program only started towards the end of 1974, most of the buildings were completed by mid-1976. Pit latrines were not constructed as planned since the sdil conditions and high water table made it impractical. The bilharzia control program was introduced in the irrigated areas and later extended to adjoining areas as well. In the initial stages, staff shortages hindered implementation of the program, but all Project staff, except for two clinical officers, were in post at the end of the Phase. Supervision missions recommended the posting of a Public Health Officer and Public Health nurse to Karonga, but the Government has not been able to provide the posts. Credit and Input Use 3.19 Seasonal and medium-term inputs were available to the farmers on credit. The use of credit by the farmers was considerably below appraisal estimates. Although by the end of Phase I almost 5,000 farmers were making - 32 - use of seasonal credit, the total value was only MK 61,000, compared to an estimated MK 134,000 at appraisal (Table 4). There are a number of reasons for this: (a) The price of fertilizer rose more rapidly than did producer prices (see para 5.06), and some farmers were unconvinced of the benefits from its application. After an initially rapid uptake of fertilizer on credit, it remained at about the same level throughout the Project period (Appendix 2); (b) Seed rates for rainfed rice and maize were reduced by about one third of the initially recommended levels; (c) Fungicides and insecticides, which had been included at appraisal, were not used on irrigated rice; (d) Use of credit may have initially been discouraged in 1974/75 by an increase in the credit charge from 10% to 21% on seasonal inputs, and induced farmers to purchase them on cash directly from ADMARC. 1/ Use of certain inputs exceeded appraisal estimates (see Appendix 2). Demand for oxen was almost twice that anticipated, while purchases of oxcarts (50 purcha.ed) on credit were not included at appraisal. 3.20 Credit recovery (Table 4) suffered during Phase I, partly because the post of Chief Credit Accountant was vacant throughout the period and because of staffing problems at lower levels. With the arrival of the senior officer after the end of Phase I, credit recovery began to improve and stood at 87% for seasonal credit by September 1977. 3.21 Credit was extended only to individual farmers, and in an effort both to cut administrative costs and to improve credit recovery, an incentive is now offered for group credit in the form of a lower credit charge. l/ 1/ The credit charges included a handling charge of 10%. The handling charge was dropped with IDA agreement. The service charge remained at 10% for 1974/75, rising to 12.5% in 75/76. In 1976/77, the credit charges were brought in line with national policy: 15% for individuals; 10% for self-accounting groups. - 33 - Table 4: SEASONAL AND MEDIUM-TERM CREDIT ACTUAL AND PLANNED 1972/73 1973/74 1974/75 1975/76 Seasonal credit (MK) - Actual 913 24,540 60,522 60,670 - Planned 20,748 55,633 93,265 133,675 Medium-term credit (MK) - Actual 1,004 7,400 12,330 7,898 - Planned 9,650 16,850 19,250 24,000 Seasonal Credit recovery (%) - September 70 52 61 49 - December 78 88 70 64 Note: At September 1977, recovery on 1974/75 and 1975/76 seasonal credit was 87%. Compliance with Covenants 3.22 All covenants were compiled with, with two exceptions. 1/ The first was failure to maintain the Liaison Committee which was to be active throughout Phase I. This committee was to co-ordinate the activities of the Ministries involved with the-Project. As noted in para. 4.02, compliance with this covenant might have avoided subsequent implementation problems. Second, development charges were to be levied of M 14/ac for participation in the irrigation scheme and of MR 7/ac for rainfed rice, in addition to a charge of MK .20/160 lb. bag of rice sold to ADMARC. Government felt that because net incomes of rice farmers were low and farmers in other (non-IDA) schemes were not charged a fee, it should waive these fees. There is no available documentation that IDA ever agreed to this, but the fees were never levied. In view of the estimated net return for rice cultivation (para 6.07), however, it is likely that the decision was based mostly on political grounds. No scheme charges were levied in Phase II of the Project. Because it appears that rice farmers have substantial returns from rice cultivation under the Project, these levies should be reconsidered. IV. INSTITUTIONAL PERFORMANCE AND DEVELOPMENT Organization 4.01 The organizational structure was the same as at appraisal, except for minor changes. The project was set up as a semi-autonomous unit within the 1/ Covenants concerning the livestock component were not complied with because the component was substantially changed during Phase I. See Borrower's comments (PPAM Annex I, paras. B. 2.17 - B.2.19). - 34 - Ministry of Agriculture and National Resources (MANR), and was responsible for carrying out the agricultural and health components, the latter in close liaison with the Ministry of Health. The Ministry of Transport and Communica- tions was responsible for the Lake Service component. MANR was to make all policy decisions in such matters as management, staffing, administrative and accounting systems, leaving day-to-day operations and implementation to the Project Manager. 4.02 Two committees were to be established to assist the implementation of the Project. The Project Liaison Committee was to be an inter-ministerial committee, composed of senior officers whose primary function was to coordin- ate activities related to the project. The Project Monitoring Committee was to be set up within the project at Karonga to monitor and evaluate its pro- gress. The Project Liaison Committee failed to play an active role, and problems with construction and maintenance of roads, construction of the jetty, and with the self-propelled barge might have been identified and sorted out earlier if the Liaison Committee had functioned as intended. The Monitoring Committee meet several times through the year, but its usefulness appears to have been limited. Staffing 4.03 Staffing difficulties were experienced throughout Phase I. Eleven senior officers were to be recruited at the start of the project, but by early 1974 only 7 of these posts were filled (Appendix 1). There were delays by Government in the creation of junior posts, and 30% of the posts created remained vacant for most of the project period. The difficulty in filling the posts in Karonga was due to its isolation and lack of social and material incentives, as well as to the growing competition for trained staff among Government departments, other development projects and the private sector. Supporting Services 4.04 In general, supporting services 1/ suffered from the absence of many senior personnel in the first few years of Phase I, and in some cases again at the end of Phase I (See Appendix I). In addition, terms of reference for the Evaluation and the Land Husbandry Units were not specific. Coordina- tion between Units, particularly Research and Extension, appears to have been poor and as a result, recommendations for husbandry practices were not always consistent. The Land Husbandry Unit devoted most of its time to dermarcation of scheme and farm boundaries. There was very little emphasis on land-use surveys or soil surveys, which would have been invaluable for the Project for developing more specific recommendations on husbandry practices. Marketing 4.05 The Project Unit was to be responsible for the sale and distribution to farmers of all input supplies, while the Agricultural Development and 1/ Includes administration, finance, land husbandry, credit, hydrology, extension and research. - 35 - Marketing Corporation (ADMARC), was to be responsible for the purchase from farmers and for the marketing of all crops. 1/ Except for maize, which was to be sold within the Northern Region, the other crops (groundnuts, cotton and rice) were to be exported. The increase in beef was to be marketed by the Cold Storage Company and sold mainly in the Central and Southern Regions of Malawi. 4.06 With respect to inputs, ADMARC agreed to take over their distribu- tion during the 1972/73 and 1973/74 crop season. But by the following year, ADMARC had decided that distribution was too costly, and the Project Unit was forced to take over distribution of inputs for the remainder of Phase I. In 1977, ADMARC agreed to resume distribution. For most of the inputs, supplies were timely and adequate. The exceptions to this were late arrival of fertilizer in the region in the 1974/75 crop season, due apparently to border closings and insufficient lake transport, and supply of groundnuts seeds, due to a change in policy by ADMARC as of 1975/76. The supply of seed of the Malimba variety (an oil nut) was discontinued because the farmers were unwilling to sell their groundnut output to ADMARC at prices that were one half to one third of prices at the local market and in Tanzania. As a result, ADMARC decided to sell only Manipintar seed, an "oil" variety, which however was less easily sold locally and therefore, less interesting to the farmer. 4.07 Marketing of output, with the exception of groundnuts and to some extent maize and beef, posed no particular problems. Paddy was purchased by ADMARC and processed by NOIL. 2/ As of July 1973, ADMARC became responsible for the sale of milled rice, and the country continued to export, with the Republic of South Africa purchasing increasing quantities at a preferential tariff. 3/ Karonga district continued to supply 30% to 40% of national sales, which increased over this period, as estimated at appraisal (Appendix 3). All of the cotton and some of the maize produced by the Project was marketed by ADMARC 4/, while the problems of marketing of groundnuts by ADMARC are discussed in paragraph 6.08. 1/ Rice was to be purchased by ADMARC and sold to National Oil Industry Ltd. (NOIL), for processing and exporting. 2/ ADMARC has 50% ownership in NOIL, and thus has effective control of its operations. NOIL mills the paddy for ADMARC at a fixed fee, and up to the end of Phase I, had sufficient capacity to mill the paddy produced by Karonga district, mostly at its mill at Chilumba, although for a time some paddy was shipped south to the mill at Nkhotakota. With recent construction of a mill at Kambwe and a parboiling plant at Chilumba, NOIL should have sufficient capacity to mill all of the paddy from Karonga district in the Northern Region. 3/ South Africa has an import quota on which the preferential tariff is paid. The quota has exceeded the amount Malawi was able to supply and this situation is expected to continue. 4/ Maize is consumed on the farm and sold outside of official channels. In 1975/76 the market price was up to 50% higher than the official price. - 36 - 4.08 Beef marketing did not develop as foreseen at appraisal due to a combination of factors. Although herd sizes increased, male animals were kept or sold as oxen for ploughing, and offtake rates were lower than expected (10% as opposed to 12% estimated at appraisal). Second, demand for beef within the Karonga district was evidently higher than assumed at appraisal, and there was considerable marketing of beef outside official channels, and some beef was marketed over the border in Zambia. As a result, the expected availability of cattle for shipment to the Central and Southern Regions did not materialize. V. FINANCIAL PERFORMANCE Costs and Disbursements 5.01 Table 5 presents the available information on project costs. The most overspent categories are vehicles, machinery and equipment, and con- struction. Cost overruns on the first category are due mainly to rapid escalations in unit costs. In some cases they more than doubledwithin three years, while appraisal estimates allowed for only a 6% per annum price increase (on all costs except salaries). The most important elLment in cost overruns in the construction category was the cost of the Lufira irrigation works. At appraisal it was estimated that the costs of construction of irrigation structures at Lufira excluding overheads, roads and field develop- ment, would be about MK 185,000 (including price and physical contingencies) or US$240,500 whereas actual costs for the works were MK 827,000, or almost US$1 million. Costs of Lufira irrigation proved so great that it was not considered worthwhile to pursue detailed feasibility studies of further irrigation projects in the Region. As a result, provisions for hiring of consultants for this purpose were never used, and so this category remained almost entirely unspent. 5.02 The credit amount of US$6.6 million was to cover 85% of estimated total Project costs. After a slow start in the first two years of the Project, disbursements gradually caught up with estimates, so that at the end of March 1978, total disbursements were US$ 6.2 million, or 94% of expected total disbursement (PPAM, Annex 1). However, although total actual disbursements are close to those estimated, there is considerable variation within categories (Table 6), In addition, the discrepancy between costs of health construction (Table 5) and disbursements is due to mistakes in disbursement requests. 5.03 Because of unexpected price increases and high construction costs, the Government of Malawi requested in 1974, and IDA agreed, to rephase the Project over a four-year period, ending in March 1976. As a result, suffi- cient funds remained to continue Phase I until the start of Phase II, in October 1976. Given the concern expressed that the remaining funds would not cover a five-year development phase, it remains unclear why there is now a saving of us$397,000. 1/ Part of the explanation may be in the movement 1/ See Borrower's comments (PPAM Annex 1, B. 2.23). - 37 - of the exchange rate, during which the Malawi Kwacha depreciated against the US Dollar. Hence 85% of these Kwacha amounts eligible for withdrawal from the credit did not fully use up the US dollars available. The proposed use of the savings has not been determined, although Government has requested, and IDA agreed in principle, to use part of the savings for the raising of the Chilumba jetty and rehabilitation of the 380-ton barge 1/ (see pqra- graph 3.06). MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I Table 5: Total Costs (US$ '000 Equivalent) ActL-l % Estimated Actual of Estimated Vehicles, machinery and equipment 470 734 156 Construction: Lufira irrigation 1/ 214 996 465 Wowe irrigation 107 - - Health 320 325 102 Other 1.006 1.194 119 Total Construction 1,647 2,515 Personnel cost 1,647 1,399 85 Seasonal and medium term loans 294 246 84 Second phase feasibility study 412 4 1 Operating and maintenance 1,176 1,375 117 Lake transport: Chilumba port 340 640 188 Chipoka port 404 - - Barge 315 375 119 Total lake transport 1J059 1.015 TOTAL 7,765 7,301 1/ Excluding farmers' contribution, roads and field staff housing. 1/ These costs amount to about US$100,000. The Government intends to submit a further proposal for the use of the remaining US$397,000 on Project-related items. Table 6: Disbursements - Actual and Estimated (US$'000 Equivalent) Total Credit Total Disbursement as of Category Allocated Disbursement Deviation % of Allocation (DCA, Schedule 1) I. Vehicles, machinery, equipment 400 624 -224 156 II. Construction of: a. Buildings, housing, offices 1,150 2,015 -865 175 infrastructure, irrigation and livestock facilities b. Health facilities 250 123 127 49 III. Personnel costs 1,400 1,189 211 85 IV. Seasonal and medium-term loans 250 209 41 84 V. Second phase feasibility study 350 3 347 1 VI. Operating and maintenance costs 1,000 1,169 -169 117 VII. Transportation 900 863 37 96 VIII. Unallocated 900 - 900 - Exchange Adjustment - 10 - - Total 6,600 6,206 1/ 394 94 I/ Total is greater than sum of parts due to rounding. - 39 - Procurement 5.04 There were no major problems with procurements and, except for minor exceptions, contracting was carried out in accordance with IDA guidelines. Accounting, Audits and Reporting 5.05 In general, and with the exception of the early years in which the senior accountant was not in post and accounts were therefore not produced, there were no major problems with accounting or auditing. Reporting was satisfactory. Prices 5.06 The prices paid by farmers for inputs changed because of two factors: increases in the official supply prices of the inputs and changes in the credit charges on seasonal items. As of 1975/76 credit charges were reduced and established in line with national policy. The supply prices of inputs were f xed by ADMARC. A comparison of estimated and actual input prices (Table'8), exclusive of credit charges, shows that actual prices of most seasonal inputs were below estimates, with the important exception of fertilizer. Because of the amounts recommended for application, this price increase would have had the largest impact on farmers' expenditure if they had applied the recommended amounts. 1/ For medium-term inputs as well, actual prices were considerably below appraisal estimates. The effect this would have had on farmers incentives to participate in the project compared to estimated incentives at appraisal depends in part, of course, on the relative output prices. For three of the four crops dealt with here, cotton, groundnuts and rice, the output prices (shown in Table 8) were some- what lower than expected at appraisal, 2/ while for maize the price was somewhat higher than expected. The combined impact on farmers' returns of changes in input and output prices compared to appraisal estimates is dis- cussed in para 6.06. I/ Survey data indicate that farmers used one bag of sulphate of ammonia per acre, instead of the recommended two bags. 2/ Since Malimba/Manipintar were the groundnut varieties grown in Karonga District, the actual price of 7.0 t/lb should be compared to the esti- mated price for Chalimbana of 7.63 t/lb. - 40 - Table 7-: INPUT AND OUTPUT PRICES: ESTIMATED AND ACTUAL (t/1b.) Appraisal Actual Official Prices Estimated Prices 1/ 1972/73 1973/74 1974/75 1975/76 1975/76 Seasonal Inputs Faya Seed 4.0 4.0 4.0 5.3 Blue Bonnet Seed 3.3 4.0 6.5 5.3 Maize Seed 5.0 9.0 5.0 8.0 Groundnut Seed (Manipintar, unshelled) 4.0 4.5 4.5 8.0 Sulphate of Ammonia 2.5 6.0 6.0 4.1 Compound 20:20:0 3.1 8.0 8.0 6.6 Insecticides (MK/acre) 8.2 11.7 11.7 19.4 Medium-Term Inputs Cotton Sprayer MK/Unit 30.0 30.0 30.0 64.3 (Knap sack) Oxen MK/Pair 140-180 1/ 161.4 Plough MK/Unit 21.3 21.3 23.7 ) Chain MK/Unit 4.0 4.0 4.6 ) 80.7 Yoke MK/Unit - - - ) Farm Cart MK/Unit 160.0 160-245 235.0 - Crops Maize 1.25 1.25 1.75 2.25 1.62 3/ Cotton 4/ 6.00 7.00 8.00 8.50 7.84 Groundnuts 5/ 6.4;5.0 7.50; 6.5 8.0;6.5 8.5;7.0 6.45 Paddy rice 6/ 3.3 3.33 4.00 4.50 3.97 1/ Estimated by inflating input and output prices in the appraisal report, which were expressed in constant 1970 prices, by the GDP price deflator for 5-1/2 years, to end - 1975 prices. 2/ Depending on age, weight, and extent of training of animals. 3/ Appraisal report gave estimated price to farmer, but did not specify the year. It has been assumed that the expected price was expressed in 1970 terms for 1975. 4/ Grade not specified for estimated price. Actual prices are for grade A. 5/ Estimated price for Chalimbana variety only. Actuals for Chalimbana, and Malimba/Manipintar, respectively. 6/ Grade I. - 41 - VI. PROJECT IMPACT 6.01 The Project was aimed primarily at increasing crop and livestock production, through improved husbandry techniques and input supplies and thereby increasing the incomes of farmers. Project impact, therefore, can be gauged by the extent to which such increases in production occurred, and by the effect of the increases on farmers income. 1/ Crop Yields 6.02 Yields estimates for Phase I are available for two years only. As noted in paragraph 3.09, the Project was implemented in scheme areas, and the Table below shows the comparison among scheme, non-scheme and appraisal- estimated vields. Table8 PROJECT YIELDS, ACTUAL AND ESTIMATED (pounds per acre) 1974/75 1975/76 Appraisal Estimate Irrigated Rice - Scheme (2 crops) - - 6,6G- - Non-scheme 800 Rainfed Rice - Scheme 2,500 2,400 2,750 - Non-scheme 1,300 2,100 800 Maize - Scheme 1,750 1,650 2,000 - Non-scheme 850 950 800 Groundnuts - Scheme 600 400 800 - Non-scheme 350 250 300 Cotton - Scheme 550 600 1,000 - Non-scheme 400 350 300 6.03 In general, the scheme areas have higher yields than the non-scheme areas, thereby indicating Project impact, but the yields achieved fall short of appraisal targets. Rice and maize yields of individual farmers, however, indicate potential for achieving (or even exceeding) appraisal targets, which could be realized through a wider adoption of improved practices. Fertilizer use was envisaged as a key factor in improving yields, but actual use during 1/ The assessment on crop production is made on the basis of survey data collected by the Evaluation Unit of the Project, and on livestock by the data supplied by the Veterinary Department. - 42 - the phase was far short of appraisal estimates. Only about 20% of the maize area in schemes received fertilizers. The.rise in .fertilizer prices and. ADMARC's difficulties with traUsport and delivery are possible explanations for the low rate of use, but at the same time yield responses to fertilizer in many of the scheme areas have not been consistent.and have tended to cloud extension recommendations. Clarification of appropriate .levels of fertilizer application in conjunction with other husbandry practices through well-planned, coordinated trials could resolve this prob;em.. In addition,.in rice growing areas in particular, there is considerable variation in soil:and topography. ,There is need, therefore, to develop .husbandry recommendations appropriate to specific conditions and areas. Total and Incremental Production 6.04 Incremental production is.shown in Table 9. Irrigated rice produc- tion has been excluded since the scheme is only now coming into operation. By the last year of Phase I, in 1975/76, estimated incremental production of maize exceeded appraisal forecasts, while production of rainfed rice was about 90% of the forecasted level. In both cases, this was due primarily to increased acreages. By contrast, the performance of cotton and groundnuts was poor. As noted earlier, the expected expansion in cotton acreage did rit occur, while yields were*only about 60% of forecasted levels. Interest in cotton appears to be increasing, however, as there was a sixfold expansion of cotton in 1976/77, to 1,430 acres, and probably reflects the relative in- creases in cotton prices paid by ADMARC. I/ Groundnut production was lower than estimated mainly because yields were one half to three quarters of appraised levels. Cropping Pattern 6.05 The Project has also had an effect on farm size and cropping pat- terns. Farm survey data for 1975/76 shows that the average scheme holding. size was- larger-than the men-1cheme holdini abli 10. 'he diTleren ein holding size appears to have been brought about as a result of the scheme approach whereby land has been redistributed from a few traditional large land holders to a large number of smallholders. In addition, the cropping patterns are different between the two groups, where Project farmers tend to devote a larger proportion of their holding to their main crop, maize or rice, than do the non-Project farmers. 2/ 1/ In 1977178 cotton prices increased by 17%-30%, depending on the grade, while prices for other commodities remained the same, or increased by less. 2/ See Borrower's comments (PPAM Annex I, B. 2.29). - 43 - Table 9: INCREMENTAL PRODUCTION (Short tons) Rainfed Rice 1/ 1972/73 1973/74 1974/75 1975/76 Acreage na 1,514 5,261 7,065 With Project Prod. - 1,893 6,708 8,478 Without Project Prod. 2/ - 568 1,710 3,709 Incremental Prod.. - 1,325 4,998 4,769 Appraisal estimate 708 2,123 3,538 5,189 Maize Acreage 54 320 2,128 4,496 With Project Prod. na na 1,862 3,709 Without Project Prod. na - 904 2,136 Incremental Prod. na - 958 1,573 Appraisal estimate 31 153 306 612 Cotton Acreage 28 627 471 221 With Project Prod. na na 130 66 Without Project Prod. - - 94 39 Incremental Prod. - - 36 27 Appraisal estimate 36 227 456 911 Groundnuts Acreage 32 290 319 974 With Project Prod. na na 96 195 Without Project Prod. - - 56 122 Incremental Prod. - - 40 73 Appraisal estimate 0 54 112 172 I/ For 1973/74, scheme and non-scheme estimates for rice are 2,500 and 1,500 lb/ac.. respectively. All other yields are given in Table 9. 2) Without Project Rice Pro4uction assumes that only 50Z of the area was cropped. KALAWI KARONGA RURAL DEVELOPHENT PROJECT ( PHASE 1) Completion Report Table 1()Scheme - Non Scheme Holding Comparison 1975076 STRATUM 1 1j STRATUM 2 1/ STRATUM 3 I/ STRATUM 4 1/ Scheme Non Scheme Scheme Non Scheme scheme Non Scheme Scheme Non Scheme Number of Holdings 74 19 29 52 6 28 23 83 Mean Rainfall (inches) 108.4 47.2 34.0 43.3 Cropa Rice 3.14 (76) 1.51 (54) 1.68 (62) 0.49 (38) - (-) 0.02 (1) 0.59 (12) 0.20 (7) (acres) Maize 0.32 (8) 0.54 (19) 0.32 (12) 0.17 (13) 5.95 (88) 0.86 (43) 1.78 (35) 0.35 (13) Maize Mix 0.07 (2) 0.10 (4) - (-) 0.15 (12) - (-) 0.35 (17) 0.15 (3) 0.35 (13) Groundnuts 0.15 (4) 0.12 (4) 0 (-) - (-) 0.37 (6) 0.02 (1) 0.27 (5) 0.05 (12) Cotton 0.02 (0.6) - (-) 0.12 (5) - (-) 0.37 (6) 0.22 (11) 0.72 (14) 0.12 (5) Cassava 0.32 (8) 0.35 (12) 0.52 (19) 0.44 (35) 0.09 (1) 0.37 (18) 1.60 (31) 1.48 (56) Other 0.12 (3) 0.20 (7) 0.07 (3) 0.02 (2) - (-) 0.17 (9) - (-) 0.12 (5) Total Crop Area 4.15 (100) 2.82 (100) 2.72 (100) 1.28 (100) 6.79 (100) 2.02 (100) 5.11 (100) 2.67 (100) Fallow 0.09 0.07 0.07 0.05 - 0.19 0.15 0.05 Holding Size 4.25 2.89 2.79 1.33 6.79 2.22 5.26 2.72 (-) Percentage SOURCE: Evaluation Unit 1/ Stratum 1 and 2 - District area north of an east, west line through Karonga Airport. Stratum 3 and 4 - District area south of an east, west line through Karonga Airport. - 45 - Returns to farmers 6.06 A summary of illustrative crop budgets (Appendix 4, Tables 1-7) based on actual prices, outputs and inputs in the last year of the Project, and compared to the budgets estimated at appraisal, are presented in Table 12. The comparison between appraisal estimates and actual performance at official ADMARC prices shows that actual returns from both irrigated and rainfed rice exceed those from the dryland crops as estimated, although actual returns were below, and in several cases far below, appraisal estimates. The net return on rainfed rice and on maize were close to expected levels, mainly because actual yields approached those estimated. In the case of maize, output price was a bit higher than expected, while seed price was much lower, so the actual per acre return was about 10% above estimated return. In the case of rainfed rice, both input and output prices were lower than expected, but input prices were relatively less, (except for fertilizer, which the farmers responded to by using the less expensive kind), thus the net return was close to appraisal. The discrepancy in net returns shown for irrigated rice is due to the comparison between double-cropping that was thought to be feasible at the time of appraisal and the single-cropping used as an estimate of actual returns, that being a more realistic assessment of the situation (double-cropping is feasible in only one out of every three years). Groundnuts and cotton returns are much lower than expected mainl because yields in both cases were only 50-60% of appraisal targets. For grounduts the input costs were about 15-20% higher than estimated, while the official output prices was less than 10% higher. For cotton, input prices were much lower than expected (about 50% of estimated prices), but this was not suffi- cient to make up for the combined effect of lower than expected yields. 6.07 Although compared to appraisal estimates the returns to farmers are not as high as anticipated, a comparison between Project participants and non-participants (Table 12) shows that Project farmers had substantially higher returns to all of the crops involved, except for cotton. As an indicator of the impact of the Project on farmers' incomes, this comparison is an understatement of the benefit from the Project, because there has un- doubtedly been a spillover effect of improved husbandry practices to non- scheme areas. As a result, one can expect that non-participants' (or without Project, non-scheme) returns are higher than they would have in fact been if there had been no Project at all. On a per acre basis, returns to project participants are 23-85% higher than returns to non-participants, with the important exception of the return to cotton during loan repayment. In 1975/76 the net returns to Project farmers on a 4.0 acre holding in a rainfed rice area have been estimated at MK 305 compared to an estimated MK 195 for non- Project farmers (Table 11 and Appendix 4, Table 6 for details). For a dryland crop farm of 4.0 acres, the net returns to Project farmers is estima- ted at MK 96, compared to MK 77 for non-Project farmers (Table 12 and Appendix 4, Table 7 for details). Although on a per manday basis the differences in net returns are not as great (Table 12) and although the returns are somewhat less than expected at appraisal, it seems clear that Project participants, numbering about 6,500 by 1975/76, were benefitting substantially from the Project. Table 11: COMPARATIVE NET RETURNS PER ACRE AND PER MANDIY Estimated at At Official 1975/76 ADMARC Prices At 1977 Market Prices 11 Appraisal With Project Without Project With Project Per Acre Per Acre Per Manday Per Acre Per Manday Per Acre Per Manday (MK) (MK) (MK) (MK) (MK)0 (MK) (MK) Irrigated Rice 219.9 2/ 89.2 2/ .60 2/ 76.7 3/ 1.28 3/ 317.4 1.10 Rainfed Rice 99.6 94.6 1.35 76.7 1.2E 140.6 2.01 Maize 22.7 24.8 .50 17.8 .45 33.9 .65 Cotton 63.1 4/ 32.2 4/ .36 4/ 25.7 .37 30.7 5/ 19.4 5/ .22 5/ Groundnuts 54.0 23.1 .27 12.5 .16 108.1 1.27 Rainfed Rice Farm of 4 Acres 304.5 1.18 195.4 .99 Dryland Crops Farm of 4 Acrep 95.8 .46 76.6 .43 1/ Average prices in 1977 at Karonga Town: ADMARC Prices 1976/77 Paddy t/lb 6.4 4.5 Maize t/lb 3.6 2.25 Groundnuts t/lb (shelled) 3l.7 7.0 These 1977 prices were used to revalue the crop budgets from 1975/76, and are therefore not strictly comparable to the relative returns available to farmers in either 1975/76 or 1976/77. 2/ Appraisal estimate for double-cropped rice; actual estimate for single-cropped rice. 3/ For rainfed rice. 4/ Before loan repayment (or, assuming that life of equipment greater than period of loan repayment). 5/ After loan repayment - 47 - 6.08 There is a large disparity in returns per acre and per manday between crops. In particular, the returns from cotton after loan repayment are lower both on a per acre and a per manday basis than for the other crops. This probably accounts for the lack of farmers' interest in cotton cultivation. On the other hand, the expansion of maize acreage would be expected because of the relatively favorable returns per acre and especially per manday compared to other dryland crops. The actual situation on the returns to groundnuts is unclear because of the possibility of local marketing at a price considerably higher than the official ADMARC price (see Table 12). Farmers who are able to market groundnuts locally realize a much higher return per acre and per manday than for the other dryland crops, but groundnut acreage has not expanded. I/ It may be that demand for groundnuts is uneven throughout the year and that marketing in Tanzania involves middlemen who take part of the margin. In addition, ADMARC was reluctant to sell groundnut seed of the Manipintar/Malimba variety and this may have initially limited the planting acreage (in 1975/76) of farmers who had depended on ADMARC's seed supply. Finally, groundnuts is traditionally a woman's crop, and thus the higher returns available on the local market accruing to women would not be an incentive to the men as decision-makers to plant more of their acreage to groundnuts. Discussion of Official Pricing of Crops 6.09 Producer prices, which in Malawi are set by ADMARC, should serve several purposes. They should offer a sufficient return to farmers to act as an incentive to production as well as to ensure them, to the extent pos- sible, an adequate income. At the same time the producer prices should reflect the economic value to the country of the crops, and thus should be kept in line with export/import parity prices.8 With respect to cotton, the incentive offered by ADMARC's producer price has apparently been less than for other crops, and cotton production has not increased not only in Karonga, but also in other parts of the country, at least partly in response to this (cotton acreage in some areas has not increased over the pait five years). Because ADMARC has expressed concern with decreasing cotton production in another more important cotton-producing area of the country, the issue of "appropriate" cotton pricing and incentives that emerges in the Karonga district is an important one. 6.10 Far the 1977/78 season, ADMARC increased producer prices by 17-33% for the three grades of cotton. It will be important, in the following season, to examine acreage planted to cotton as a measure of farmers' response. If farmers have been responsive to the price increase, ADMARC might consider further increases. Based on the current export price, and the medium-term outlook (through 1985) for cotton prices, and based on the profits realized 1/ According to survey data, groundnut area for the district as a whole decreased dramatically from 8,000 acres in 1968/69 to 2,000 acres in 1976/77, including mixed stands. 2/ Using an acceptable shadow exchange rate. - 48 - by ADMARC on cotton 1/, there appears to be room for further increases in the producer price. In addition, ADMARC might consider changing the grading procedure for cotton to limit it to two grades, as in most countries. ADMARC's claim that it receives a premium on grade A cotton should be sub- stantiated. It may be that simplifying the grading system would involve no loss to ADMARC but a gain to the farmers in that less labor would be required at the same time that the average weighted price for cotton might increase. 6.11 The groundnut pricing is a more difficult issue because of the particular situation of Karonga district, where ADMARC's official price does not offer a good incentive to producers of groundnuts, although local market and Tanzanian prices do. While the current export parity price indicates considerable room for increases in farmgate prices, and ADMARC has realized a good profit, 2/ it would not be possible for ADMARC to increase its farm- gate price to a level competitive with local prices without incurring a heavy loss. In addition, the medium term prospect for groundnut oil indicates a decrease in export prices, in current terms. With respect to the groundnut situation in Karonga, therefore, it is not clear what effect an increase in ADMARC's producer price would have on production sold to ADMARC, or whether such an increase should be made. 6.12 One final comment is in order on ADMARC's producer price for rice, which clearly offers a good incentive to the farmer to cultivate the crop, as well as providing him with an adequate income. The economic parity farmgate price has.been less than actual farmgate price and ADMARC has incurred a loss on its rice account in four of the past six years (1971/72- 1976/77), which indicates that there has been a subsidy to rice growers. While this may be a conscious policy, it should be examined in light of the distribution of income. Dryland crop farmers, whether Project participants or not, have considerably lower incomes than rice farmers and with little or no subsidy. 3/ If it is the country's policy to subsidize subsistence or partly-subsistent farmers, ADMARC might consider redistributing the subsidy toward dryland crop farmers in the form of higher producer prices for maize, cotton and possibly groundnuts. 1/ Net profits on cotton as a percentage of the producer price for cotton have ranged from 16% to 92% in the past six years. 2/ Net profits on groundnuts as a percentage of the producer price have ranged from 24% to 106% in the past six years. 3/ Producers of maize may receive some subsidy, as ADMARC occasionally incurs a loss on maize marketing although if the farmgate price is considered in light of an import parity price (rather than export parity price, which is considerably lower), there is no subsidy. - 49 - Technological Change 6.13 The establishment of credit provisions to farmers was included in the Project to encourage and facilitate use of inputs considered necessary to improve farming practices. One indication of the extent of technological change is the use of this credit, although the extent to which credit was used is not an entirely satisfactory measure for appraising the input use, since farmers may have preferred to purchase the inputs on a cash basis. Infor- mation available from ADMARC's cash sales of inputs over these years is fragmentary, and there is no information on movements of stocks. It appears, however, that on certain items, such as maize seed, insecticides and in the early years of the Project, fertilizer, cash sales may have accounted for 30% to 50% of total sales. In addition, in the last two years of Phase I, ADMARC had cash sales of about 100 ploughs per season, compared to 25 ploughs per season purchased on credit. A second indication of the degree of success of the Project in terms of introducing changes in farming technology comes from survey information, collected by the Evaluation Unit, comparing schema-farmers to non-schema farmers. These data, available only for the final year of Phase I, indicate that for certain husbandry practices, such as use of improved seed, and application of fertilizer, scheme pa:ticipants had significantly higher adoption rates of improved husbandry practices than did non-participants. 1/ Thus, while credit use on both seasonal and medium-term items was considerably below appraisal estimates, there are indications that the Project has had an impact on introducing improved husbandry practices. Livestock 6.14 The effect of the livestock program has been to increase herd growth from 2.5% per annum to 6.7% per annum (appraisal under-estimated without Project rate at 0.5%). Some of this increase is probably due to a better collection ,of statistics, while the remainder is due to reduced mortality from improved dipping facilities and the retention of a greater number of oxen and bulls for ploughing, particularly in Karonga. 1/ The response of scheme farmers to husbandry practices involving changes in labor patterns was not so marked: weeding and planting dates were not significantly different between the two groups. These results on input use may be somewhat biased because inputs, although available on a cash basis through ADMARC, may have been less accessible to non-scheme farmers than to scheme-farmers, since the Project distributed the inputs to scheme farmers for most of Phase I. - 50 - Table 12: Herd Composition and Growth 1963-1977 Karonga Percentage of Percentage of Karonga Chitipa Combined Herd, Cows and Herd, Bulls and Cattle Nos. Cattle Nos. Herd Heifers Oxen 1963 25,146 - - 1968 28,066 27,132 55,198 60% (63) 17% (10) 1973 32,470 29,497 61,967 58% (66) 20% (12) 1977 41,183 36,783 78,045 57% (63) 24% (16) Growth rates 1963-73 2.5% 1.7% 2.4% - - 1973-77 6.7% 6.1% 6.5% At appraisal, it was anticipated that offtake of animals would rise from 10% to 12%. This has not taken place, because male animals are eithe- retained for draft purposes in Karonga or are slaughtered early for meat. A small grazing scheme has been started in Karonga district, but there has been difficulty in getting farmers to participate despite the fact that no fee is levied. Health Services 6.15 The Project has significantly improved health services in the Karonga District. Of the eight health facilities existing in 1971, the Project has improved three of them and added a further five. In addition, it has introduced a bilharzia control program in three irrigation scheme areas, i.e. Lufira, Wovwe, Hara and Wililo, in the hills above Lufira. In terms of increased medical attention, improvements to the Karonga hospital have in- creased capacity by nearly 50%. The new Nyongwe Center deals with 8,000 to 10,000 patient visits monthly. Each new health post deals with 60 to 70 patients daily and over 200 to 300 people attend the monthly or fortnightly under-5 health clinics. Port Construction and Lake Transportation 6.16 Although the Chilumba jetty was in service for the latter part of Phase I, the 380-ton barge was not available for use until after Phase I and construction of the southern port was postponed until Phase II. As a result, the impact of the improved lake transport facilites on the movement of cargo and traffic was considerably less than anticipated. In addition, however, increased traffic from the project was used as a justification for the rehabilitation of the lake service, but the volume of traffic envis- aged at appraisal has not yet developed. In view of this it appears that including the lake transport component of the project in Phase I was premature. - 51 - VII. RATES OF RETURN 7.01 Economic rates of return have been estimated for the Project as well as for separate components. The treatment of costs and benefits is discussed in detail in Appendix 5. 7.02 The estimated rate of return for the entire Project is 6% (Table 13), compared to 12% at appraisal. Excluding the irrigation component, for which the rate of return was pegative infinity, because costs exceed benefits throughout the life of the component, the estimated rate of return is 33%. It was difficult to determine rates of return for the dryland crops and rainfed rice components because the allocation of Project overhead costs is subjective and both separately and combined, the components are very sensitive to their allocation. A reasonable estimate, which allocated overheads in a proportion similar to that in the appraisal report, gives a rate of return for combined rainfed rice and dryland crops of 24%. This compares with an appraisal estimate of 16% for rainfed rice and 21% for dryland crops. 1/ Table 13: ESTIMATED ECONOMIC RATE OF RETURN Total Total Project Rainfed rice and Project excluding irrigation Dryland crops 1/ Livestock 6% 33% 24% 35% 7.03 In conclusion, the Project, exclusive of irrigation, has a very high rate of return, which reflects substantial increases in production (of rice, maize and beef) from relatively low-cost improvements of use of fertil- izer, improved seed and a more extensive dipping program. From the estimated crop budgets, the benefits to the farmers are apparent, while on the basis of the economic analysis, the Project is also of considerable value to the economy. The irrigated rice scheme, however, is the one component which has been singularly unsuccessful. VIII. BANK PERFORMANCE 8.01 The Bank's overall performance has been creditable. It designed and supported a Project which provided for infrastructure, crop development and 1/ Exclusive of health and transportation costs. - 52 - improved health facilities and also provided a strong stimulus to the overall development of the district. In retrospect, however, it tended to be both under-and over-optimistic at the planning stage. The increased production of rainfed rice and maize has been achieved with lesser amounts of inputs and fewer staff, whereas the introduction of irrigated rice in Lufira has met with difficulties. The Bank obviously had doubts with regard to the irrigation component but gave in to pressures, a decision which turned out to be very costly. 8.02 The supervision of the Project was satisfactory. However, it would have been of great assistance to the project if the supervision missions had given even greater emphasis to the problems connected with road construction and maintenance and irrigation works with a view to getting them resolved by the ministries concerned. IX. CONCLUSIONS Conclusions based on the foregoing chapters are as follows: (a) The Project brought about some land reform, provided inputs, extension advice and generally provided a stimulus to the development of the Karonga district. (b) It was successful in providing the planned physical infrastructure and increasing crop production in rainfed rice and maize. Health facilities have been improved in the Karouga district. The livestock program has been successful in improving the health of the cattle and in increasing the growth rate of the herd. Failure to achieve targets was due either to technical problems, e.g., irriga- ted rice, or to lack of price incentives, e.g., cotton. (c) Phase I was completed one year ahead of the appraisal schedule without seriously affecting the original targets. Further, the targets were achieved with fewer staff than envisaged and in the case of crop production, with a lower level of input use. Future phases however, would tend to be on less suitable land and could demand a greater extension and input use. (d) Farmers have accepted improved husbandry practices which involve minimum risk or change in labor-use pattern, e.g., improved seed. However, adoption of other practices will probably come from the rapidly expanding use of oxen for cultivation. (e) The problems of implementing the irrigation component highlight the danger of accepting proposals based on inadequate data. However, now that additional information is available, an evaluation of the Lufira scheme under the National Irrigation Study could prove to be useful. - 53 - (f) It is important to enunciate clearly the roles and responsibili- ties of the various components in the organizational structure. In Karonga the Land Husbandry Unit could have played a more useful role had its terms of reference been clearly defined. The Research Unit tended to give greater emphasis to national trials.at the expense of Project-specific trials. This over-emphasis was carried over into Phase II, but is now being corrected. (g) There is a need for improved coordination and communication between MANR and MOW. The question of road standards and maintenance ought to have been sorted out at the outset of the Project. This particular problem has remained with the Project through Phase II. However, definite steps are now being taken by the two Ministries concerned to resolve this issue. (h) Some of the lessons learned in Phase I were used in the plan- ning and implementation of Phase II. The roles of various sections in the organizational structure are more clearly specified. The research program is geared more to prolect needs. As a result of the Lufira experience no 'irrigation component was included. (i) Prospects for the future may lie in the increased production of rainfed rice, maize and cotton. This would require intensified research to develop recommendations which are location-specific. &¶ L. f6 'W * Karioa Rural Devliop:nr,nt Project, Ph se I Appendix I Com.plettic.n R,!n)OrL Stalfin, Paccerrn 1972 ' 1973 ' 1974 ' 1975 ' 1976 • 1977 1. PoraectAaniag.er J.J. V__en S.E.C. Shur.ba 2.Assts_ant P c t. anager Sh u-l lvul a Huwt la 3.Ftnanctal CntrollÉE Sm i th S trsworthl 4.Sr.Antnr l_Ibtsblndry Officer J.elc.aurhl1n tr, J.F. l1wenechanya (i») Ntmitz (PO) $S.Sta,lement and Ext. officer Mvula Ndolo 6.Sr. Field Mtvist on E_ngneer V.K. Dug-al. G. Owend (Pø) tIphanda (ro) Kaloloktuya (PO) Pelle (PO) 7-5r. Agrieultural Research Officer B.C. Smith% (59) • Maluwaya (PO) Ml r .kuso (0) 8.Chief Credit Accountant 9.AgUricuu_tural Economit S.J. Polland C.D. Poata 10.Land Hushandry Officer J.F. Blyth l1.Hydrologtst D.R. Archer E.2. Latet Appendix 2 MALAWI Karonga Rural Development Project - Phase I Completion Report Seasonal and Medium-term Credit 1972/73 1973/74 1974/75 1975/76 Actual Planned /1 Seasonal Credit Sales Sulphate of Ammonia (ST) 2.4 )145.2 172.0 203.2 )589 )3,491 Compound 20-20-0 (ST) 2.4 ) 42.4 21.4 ) ) Maize seeds (ST) 0.5 3.5 6.5 3.5 14.0 117.3 Ground seeds (ST) 1.3 11.6 26.5 --- 39.4 70.7 Paddy seeds (ST) --- 35.5 148.8 195.1 379.4 1,344.5 Packages insecticides 2,28b 67,368 60,923 53,231 133,803 na ! Total seasonal credit (MK) 913 24,540 60,522 60,670 14'3 /2 28,930 /2 Number of farmers 43 2,050 4,073 4,866 4,366 4,581 Average seasonal credit (MK) 21.2 12 14.9 12.5 12.5 35.5 Ln LI Medium Credit Sales Oxen (pair) --- 75 40 6 121 57 Chains (No.) --- na 35 34 69 57 Plows (No.) --- na 27 25 52 57 Sprayers (No.) 30 58 188 21 247 1,750 Ox carts (No.) --- 11 18 22 5L Ridgers (No.) --- na 16 --- 16 --- Total medium-term credit (MK) 1,004 7,400 12,330 7,8911 93,43?91,750 /3 Number of farmers 31 122 194 51 51 550 Average medium-term credit (MK) 32.4 60.7 68.6 154.9 /1 As at year 5 of project, note that all planned totals in MK are expressed in constnnt 1K'70 terms. In current terns, estinated totals, allowing for 6% per annum price increase, would be 20-35" vi her. /2 Incremental credit. /3 Includes credit for MYP settlers, totalling MX 13,200 (constant 1970 terms) over 5 yenrs. - 56 - MALAI Appendix 3 KARONGA RURAL DEVELOPMENT PPOJECT(PHAS17 1) Completion Report Crop Purchases by ADMARC 1/ (Short tons by risc- Year) 1972/73 1973/74 197,175 1975/76 Maize (1) Karonga oistrict 5.8 0.1 1.2 - (2) National Total 73,537 66,246 72,237 32,145 Proportion (1)*(2) (%) 0 0 0 0 Groundnuts (1) Karonga District 8.2 0.02 12.7 2.7 (2) National Total 43,197 32,955 31,724 36,165 Proportion(l) +(2)(%) 0 0 0 0 Cotton (1) Karonga District 183.2 248.6 327.3 344.0 (2) National Total 24,060 17,863 23,679 19,632 Proportion(1)-(2)(%) 0.1 1.4 1.4 1.8 Paddy (1) Karonga District 6,374.0 6,992.0 6,626.2 6,300.0 (2) National Total 21,859 19,399 23,475 16,213 Proportion(1)+(2)(%) 29.3 36.0 28.2 38.9 1/ Figures on Karonga district purchases are from the Evaluation Unit at KRDP and figures on national purchases are from the Ministry of Agriculture and Natural Resources. Appendix 4 - 57 - Table 1 MALAWI KARONGA RURAL DEVELOPMENT PROJECT (PPASE I) Completion Report Illustrative Crop Budget - Irrigated Rice - 1 Acre (Current Terms) Actual Estimate Appraisal Estimate (single crop)l/ (double crop)1/ 1975/76 1975/76 Yield (lbs) 2,670 6,586 Gross Return (MK) 120.2 308.0 Costs (.K) 3/ Seed 3.9 7.6 Fertilizer 19.8 34.8 Insecticides - 7.6 Field bagging 2.8 7.n Tools 2.0 - Scheme service charge 4/ - 22.2 Credit charge 5/ 2.5 8.9 Total costs 31.0 89.0 Net Income (MK) 89.2 219.9 Mandays Required 6/ 149 289 Net Income Per Manday (MR) .60 .76 1/ Appraisal assumed double cropping. At Lufira there is sufficient water to double crop only once in three years. Actual estimate of returns has therefore been calculated for a single crop. 2/ Average yields from 1976/77 crop season at Lufira. 3/ See footnote 2/, Appendix 3, Table 2. Use of inputs is as follows: Actual Appraisal 1975/76 Blue Bonnet Seed 60 lb/ac/crop 60 lb/ac/crop Fertilizer 20:20:0 - 100 lb/ac/crop S/A 150 kg/ac 200 lb/ac/crop Fungicides - MK 3.8/ac/crop 4/ At appraisal, scheme charges of MK 14/ac and MK 0.2/bag were foreseen These were never levied. 5/ See footnote 4/ in Table 2. 6/ From appraisal report. - 58 - Appendix 4 Table 2 MALAW4I KARONGA RURAL DEVELOPMENT PROJECT (PHASE I) Completion Report Illustrative Crop Budget - Rainfed Rice - 1 Acre (Current Terms) Actual Estimate Appraisal Estimate 1974/75 1975/76 1975/76 Yield (lbs) 2,550 2,420 2,750 Gross Return (MK) 1/ 102.0 108.9 129.0 Costs (MK) 2/ Seed 2.4 2.4 6.3 Fertilizer 6.6 6.6 7.8 Field bagging 2.7 2.5 3.3 Tools 2.0 2.0 - Scheme service charge 3/ - - 10.4 Credit charge 4/ .7 .8 1.6 Total Costs 14.4 14.3 29.4 Net Income (MK) 87.6 94.6 99.6 Mandays Required 5/ 70 70 Net Income per manday (MK) 1.25 1.35 1/ Price per pound taken from Table 6. 2/ Input prices for 1974/75 are the same as for 1975/76, and are found in Table 5, except for field bagging, noted below. Actual input quantities are taken from the results of surveys done by the Evaluation Unit and are presented below: Actual Appraisal 1974/75 1975/76 Faya Seed 60 lb 60 lb 100 lb Fertilizer 20:20:0 - - 100 lb S/A 50 kg 50 kg Field bagging (160 lb bags ) 50t/bag 50t/bag 19.1 t/bag It is assumed that farmers purchase bags once every three years. 3/ At appraisal, scheme charges of MK 7/ac and MK 0.2/bag were foreseen. These were never levied. 4/ Actual credit charges: lnT in 1974/73 and 12.5% in 1975/76 on fertilizer and insecticide (for cotton). At appraisal 21% in charges was estimated. 5/ From Evaluation Unit. - 59 - Appendix 4 Table MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I Completion Report Illustrative Crop Budget - Maize - 1 Acre (Current Terms) Appraisal Actual Estimate Estimate 1974/75 1975/76 1975/76 Yield (1bs) 1,750 1,630 2,000 Gross return (MK) 1/ 30.6 36.7 38.2 Costs (MK) 2/ Seed 2.0 1.1 2.9 Fertilizer 6.6 6.6 9.6 Field bagging 1.5 1.4 - Tools 2.0 2.0 1.0 Credit charge 3/ 0.7 0.8 2.0 Total costs 12.8 11.9 15.5 Net income (MK) 17.8 24.8 22.7 Mandays required 4/ 52 50 Net income per manday (MK) 0.34 0.50 1/ See footnote 1/, Table 2. 2/ See footnote 2/, Table 2 on input prices, except for the price of maize seed in 1974/75, which was 9.0 t/1b for synthetic. Actual Appraisal 1974/75 1975/76 1975/76 Seed 22 lb - 22 lb 30 lb Fertilizer (S/A) 50 kg 50 kg 200 lb Field bagging (200 lb/bag) 50 t/bag 50 t/bag ---- NOTE: It is assumed that farmers purchase bags once every three years, and that the entire crop is marketed. 3/ See footnote 4/ in Table 2. 4/ From NRDP appraisal report. - 60 - Appendix 4 Table 4 MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I Completion Report Illustrative Crop Budget - Groundnuts-l Acre (Current Terms) Appraisal Actual Estimate Estimate 1974/75 1975/76 1975/76 Yield (lb) 600 410 800 Gross return (MK) 1/ 39.0 28.7 61.0 Costs (MK) 2/ Seed 2.7 2.7 6.7 Field bagging 1.3 0.9 --- Tools 2.0 2.0 0.3 Total costs 6.0 5.6 7.2 Net income (K) 33.0 23.1 54.0 Mandays required 3/ 95 85 Net income per manday (MK) 0.35 0.27 1/ See footnote 1/, Table 2. 2/ See footnote 2/, Table 2. Input quantities for seed were 70 lb/ac unshelled at appraisal, and 60 lb/ac unshelled, actual. 3/ See footnote 4/, Table 3. - 61 - Appendix 4 Table 5 MALAWI KARONGA RURAL DEVELOPHENT PROJECT (PHASE I) Completion Report Illustrative Crop Budget - Cotton - 1 Acre (Current terms) Actual Estimate Appraisal Esti-ate 1974/75 1975/76 1975/76 Yield (1bs) 540 600 1,000 Gross Return (MK) 1/ 41.0 48.6 92.7 Costs (..K) 2/ Insecticides 11.7 11.7 22.9 Field Bagging 1.1 1.2 ---- Tools, Sprayer, Maintenance 2.0 2.0 1.9 Credit Charges 3/ 1.2 1.5 4.8 Total Costs 16.0 16.4 29.6 Net Income Before Loan RepaymenL 25.0 32.2 63.1 Loan Repayment 4/ 12.8 12.8 32.4 Net Income After Loan Repayment 12.2 19.4 30.7 Mandays required 5/ 90 90 Net Income per Manday .14 .22 1/ Output prices are weighted averages of the official prices for 3 grades of cotton; 1974/75:. 7.1 t/1b; 1975/76: 8.1.t/lb. 2/ See footnote 2/, Table 2. It is assumed that farmers purchase 60 lb bags at 35t/bag every three years. Insecticide use was as follows: Actual: 1974/75 - 1975/76 Estimated (NK) (MK) Sevin 30 pkts @ 2t/pkt 6.0 DDT 34 pkts @ 14t/pkt 4.76 Dimethoate 14 pkts @ 7t/pkt .98 Total 11.74 22.9 3/ See footnote 4/ in Table 2. 4/ For sprayer, at 10% mark up and 8% p.a. in 3 equal annual installments. 5/ From Evaluation Unit. - 62 - ^ppenaix . Table 6 MALAWI KARONGA RURAL DEVELOPMENT PROJECT (PHASE I) Completion Report ILLUSTRATIVE FARM BUDGET: RAINFED RICE AREA AT OFFICIAL 1975/76 ADMARC PRICES With Project Without Project Value of Value of Crop ArealProduction Costs Net Returns Area 1/ Production Costs Net Return (acres) (MK) (MK) (MK) (MK) (MK) (MK) (MK) Rice 3.0 326.7 42.9 283.8 2.2 183.2 14.5 16E.7 Maize .4 14.9 5.0 9.9 .9 19.2 3.2 16.0 Groundnut .2 5.7 1.1 4.6 .2 3.5 1.0 2.5 Cassava 2/ .3 6.8 .6 6.2 .4 9.0 .8 8.2 Fallow .1 - - - .1 - TOTAL 4A0 354.1 49.6 304.5 4.0 214.9 19.5 195.4 Mandays 259 198 Net Income Per 1.18 .99 Manday (MK) 1/ Crop acreage derived from survey results 2/ Information on cassava from Evaluation Unit, KRDP. MALAWI APPENDIX 4' KARONGA RURAL DEVELOPMENT PROJECT (PHASE I) Table 7 Completion Report ILLUSTRATIVE FARM BUDGET: DRYLAND CROP AREA AT OFFICIAL 1975/76 ADMARC PRICES With Project Without Project Value of Value of Crop Area 1/ Production Costs Net Return Area 1/ Production Costs Net Return Maize 2.0 73.4 23.8 49.6 1.4 29.9 5.0 24.9 Groundnuts .2 5.7 1.1 4.6 .2 3.5 1.0 2.5 Cotton .4 19.4 6.6 2/ 12.8 2/ .4 9.0 .8 8.2 Cassava 1.4 31.5 2.8 28.7 2.0 45.0 4.0 41.0 TOTAL 4.0 130.0 34.3 95.7 4.0 87.4 10.8 76.6 Mandays 209 178 Net Income Per .46 .43 Manday 1/ Crop acreage derived from survey results. 2/ After loan repayment - 64 - APPENDIX 5 Page 1 MALAWI KARONGA RURAL DEVELOPMENT PROJECT - PHASE I Completion Report Economic Rate of Return In estimating the economic rate of return the following assumptions were made: Project Life is taken at 25 years, as in the appraisal report. Costs Taxes are excluded at a rate of 40% on the 25% beAefit margin for construction of houses by contrators (as in appraisal). It was estimated that 25% of the total cost of construction was for housing. Farm Labor: As at appraisal, incremental farm labor was costed at zero. Since an argument can be made that farm labor has an opportunity cost close to the minimum legal wage (MK .25/day during 1973-74 period, or MK .13/day to MK .15/day in 1970 terms), an approximate effect on the rate of return was calculated costing this incremental labor. Total costs on all component excluding irrigation would increase by about 2%, and the rate of return for the Project, excluding irrigationwould decrease from 33% to 31% Transport component: Cost of the lake service was excluded, as at appraisal. Health component: Costs of construction and maintenance of the health component were included at appraisal. Overheads were distributed among the Project components as follows: 10% each for livestock and dryland costs, 50% for irrigated rice and 30% for rainfed rice. At appraisal the allocation was 10% each to livestock and dryland crops, 40% each to rainfed and irrigated rice. After 1975/76, - 65 - APPENDIX 5 Page 2 in view of the reduced importance of the irrigation scheme in terms of area (800 acres, of which only about 330 were directly under Project supervision, compared to an estimated 1,500 acres at appraisal), but continued high costs of maintenance, overheads were redistributed as follows: 10% for livestock, 15% for dryland crops, 35% for rainfed rice and 40% for irrigated rice. Cost of extension workers is included for the 473 acres of irrigated rice allocated to Spearhead Enterprises. (3 extension workers at MK 750 each/yr.). All costs are expressed in constant 1970 terms, and have been deflated by the index of international inflation. Benefits Incremental production is based on actual performance for Phase I of the Project (from information supplied by the Evaluation Unit). For yields after 1976/77 and for acreage after 1975/76 the following assumptions were made: Yields for with and without Project situations we:e taken as the average over 3 years (1974/75 - 1976/77) for scheme and non-scheme, respectively. In most cases the yields did not show a trend in either direction for the 3 years, so the average yields were considered the most reliable, although average scheme yields were generally below and some average non-scheme yields were above appraisal estimates. The latter may reflect spill over effects of the Project's introduction of improved crop husbandry. Acreage was assumed to remain at the same level as at the end of Phase I, except that for dryland crops the proportion of total acreage planted to maize, cotton and groundnuts was assumed to be the same as in 1976/77 (one year after Phase I). This reflected a shift out of groundnuts and, to a smaller extent, out of maize into cotton which was probably in response to a change in relative prices offered by ADMARC between 1975/76 and 1976/77. There was a further price change in 1977/78, which made cotton still more attractive, but the most recent information on planted acreage was from 1976/77 and this was therefore the year used as the basis for the projections. For dryland crops, it was assumed that the same cropping pattern would have existed in the absence of the Project. For rice schemes (irrigated and rainfed) it was estimated that 50% of the area had been planted with rice before the Project, and this percentage was used as the without- Project average during Phase I. From 1976/77 onward, it is assumed that the acreage would have expanded by 1% per year without the Project, in response to popu- lation pressures. Population growth has been about 2.9% p.a. in the Karonga district over the past ten years but estimates of total cultivated area over the 1968/69 - 1976/77 period show an increase of about 1% per year. Irrigated rice production includes estimated production of Malawi Young Pioneers, with yields assumed to be the same as those for Project pro- duction. It is further assumed that in one out of every three years there - 66 - APPENDIX 5 Page 3 will be sufficient water to plant only 50% of the area planted in the other two years. Prices: Actual import and export parity prices for 1972/73 - 1976/77 are taken on the basis of figures supplied by ADMARC and IBRD. After 1976/77 the prices are the same as those used for the appraisal of National Rural Development Program. All prices are deflated to 1970 terms by the index of international inflation. DEM. REPUBL1 u r e'eopthe O TANZANIA evlo90' co5PUuo, h ANGOLA ZAMWA_ SZAMBIA KARONGA , SOUTH .REPUBLICRD WEST J o WS BOTSWANA 2 AFRICAJ CHILUMBA REPUBLIC WAZILAND AFRICA L SOTO Chiweta Rvmpi Caq Mzuzu NKHATA. \ AKEI Nkhot'. Ko0. Kcasungu Dowa Senga.. Cotur Pp Mchinji CHIPOKA L.LONW ONK .. EBAY .. Deda a3 ,nf TR NA cA.LA O cheu Novo Freixo Bolåka O9I I RLWONDE Kosupe 0 2 40 60 0 y0 ZOMBA TTE ET9rE Mwanzo MALA WI T'LCimu. .Bl NTYRE LIMBE Chikw~w Mloii ROADSCo -- --ROADS UNDER CONSTRUCTION0 RALlWAYS - - -INTERNATIONAL BOUNDARY Cim . .REGIONAL BOUNDARY ONCOING IDA AGRICULTURAL PROJECTS iNoj 0 20 40 60 80 100 SEPTEMBER 1971 1BRD 359 4R