Papua New Guinea - Agricultural Development Project

Document of The World Bank FOR OFFICIAL USE ONLY RETURN TO \JL ~Report No.1877 ONE PROJECT PERFORMANCE AUDIT REPORT PAPUA NEW GUINEA (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (PART) (CREDIT 175-PNG, PART) February 9, 1978 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. CURRENCY EQUIVALENTS Currency Unit = Kina (K) US$1.00 - KO.81 KI - US$1..23 K1 million = US$1,233,660 See Annex Table 4 for variations through life of project WEIGHTS AND MEASURES 1 millimeter (mm) - 0.039 inch (in) 1 meter (m) - 1,000 mm = 3.2808 feet = 1.0936 yards 1 hectare (ha) - 0.01 sq km = 2.47 acres 1 kilogram (kg) M 2.2046 pounds (lb) 1 metric ton (t) = 1,000 kg = 0.9842 long ton = 2,205 lb ABBREVIATIONS DASF - Department of Agriculture Stock and Fisheries DPI - Department of Primary Industries GOA - Government of the Commonwealth of Australia GOPNG - Government of Papua New Guinea IBRD - International Bank for Reconstruction and Development IDA - International Development Association NBPOD - New Britain Oil Palm Development Ltd. NGGA - New Guinea Graziers Association PCR - Project Completion Report PNG - (Territory of) Papua New Guinea PNGDB - Papua New Guinea Development Bank PPAR - Project Performance Audit Report THE GOVERNMENT OF PAPUA NEW GUINEA FISCAL YEAR July 1 - June 30 FOR OFFICIAL USE ONLY Project Performance Audit Report PAPUA NEW GUINEA (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (Part) (Credit 175-PNG, Part) Table of Contents Page Preface Basic Data Sheet Highlights PROJECT PERFORMANCE AUDIT MEMORANDUM Project Summary 1 OED Comments Institutional Shortfall 5 Expatriate Record Keeping 5 Technology Transfer 6 PROJECT COMPLETION REPORT 1. Introduction A 1 2. Background A 2 3. Project Formulation A 3 4. The Coconut Component A 4 5. The Cattle Component A 6 6. Project Cost A 19 7. Economic Impact A 24 8. Institutional Performance and Development A 28 9. The Bank's Performance A 30 Map This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization  Project Performance Audit Report PAPUA NEW GUINEA (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (Part) (Credit 175-PNG, Part) Preface This report presents the results of an audit of accomplishments under those parts of Credit 175-PNG, signed in January 1970 for US$5 million, allocated to activities other than the oil palm settlement project at Hoskins. The oil palm project was discussed in PPA Report No. 1400 of December 28, 1976, which covered Credit 137-PNG and related funding under Credit 175-PNG. The latter was closed, fully disbursed, in June 1976. Of the credit, US$3.4 million was disbursed for the oil palm settlement and US$1.6 million for other activities, which in implementation were limited to the livestock sector. The report consists of an Audit Memorandum and the Project Completion Report. Under OED's abbreviated auditing process, its audit was confined to a review of the PCR, Appraisal Report and project supervision reports, and discussions with Bank staff. The PCR provides a comprehensive analysis of developments of the non-settlement components of the project. The Memorandum consists of a modified version of the summary given in the PCR and a few addi- tional remarks by OED. 4 PROJECT PERFORMANCE AUDIT BASIC DATA SHEET PAPUA NEW GUINEA: (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (PART) (Credit 175-PNG) KEY PROJECT DATA Item Appraisal Actual or Expectation Current Estimate Total Project Cost (US$ million) 8.8 8.5 Underrun or Overrun (%) n.a. /1 Credit Amount (US$ million) 5.5 7 Disbursed )- 5.5 Cancelled )- n.a. Repaid to )- n.a. Outstanding to )- 5.5 Date Physical Components Completed 6/74 6/76 Proportion Completed by Above Date (%) 100 72 Proportion of Time Underrun or Overrun (%) - 33 Economic Rate of Return (%) 11 21 OTHER PROJECT DATA Original Actual or Item Plan Revisions Est. Actual First Mention in Files or Timetable - 10/66 Government's Application - 10/66 Negotiations - Board Approval early/68 1/14/70 Credit Agreement Date mid/68 1/30/70 Effectiveness Date 5/29/70 6/03/70 Closing Date 6/30/74 6/30/75 6/30/76 6/16/76.L3- Borrower Administration of the Territory of PNG Executing Agency - Fiscal Year of Borrower July 1 - June 30 Follow-on Project Name Smallholder Livestock Credit Credit Number 348-PNG Amount (US$ million) 5.0 million Credit Agreement Date 1/04/73 Mission Data Sent Month, No. of No. of Date of Item by Year Weeks Persons Manweeks Report Identification Bank 3/67 - - - 8/67 Preappraisal Bank 6/67 - - - Appraisal /4 Bank 1-2/68 & 5-6/69 1 5-2 2 12/69 Total 2 Supervision I Bank 6/70 2 3 6 7/70 Supervision II Bank 5-6/71 2 2 4 7/71 Supervision III Bank 5-6/72 2 1 2 7/72 Supervision IV Bank 5/73 2 2 4 8/73 Supervision V Bank 3-4/74 2 1 2 5/74 Supervision VI Bank 8-9/75 1 4 4 11/75 Supervision VII Bank 7/76 1 2 2 8/76 Completion Bank 1-2/77 3 2 6 7/77 Total COUNTRY EXCHANGE RATES Name of Currency (Abbreviation) Australian Dollar/since 4/75 Kina (A$/K) Year: Appraisal Year Average Exchange Rate: US$1 = 0.84 A$ Intervening Years Average US$1 = 0.81 A$ Completion Year Average US$1 - 0.92 AS A$l - K 0.882 /1 Funds from the cattle/coconut component were reallocated to the smallholder oil palm component. /2 Plus exchange adjustment of $0.519 million. 7_? Final disbursement date. /4 Supervision Mission I for Credit 137-PNG served as Appraisal Mission for Credit 175-PNG. /5 Part of the supervision time was allocated to the smallholder oil palm subproject.  Project Performance Audit Report PAPUA NEW GUINEA (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (Part) (Credit 175-PNG, Part) Highlights The objectives-of the cattle and coconut subprojects, both intended for expatriate farmers, were not achieved as planned. Ranch development was scaled down and the coconut subproject never materialized, mainly because the date of independence was advanced much earlier than expected at appraisal. Support of the expatriate cattle sector nevertheless proved valuable to the new nation: project ranches produced the breeder stock required for small- holder livestock development. The project also provided some training for indigenous cattlemen. The re-estimated economic return of the reduced proj- ect exceeds appraisal expectations. The following issues may be of particular interest: Lack of data even on expatriate managed farms (PPAR, para. 20 and PCR, paras. 5.19, 6.01, 7.08, 7.12, 8.04, 9.06-9.07) Successful training of indigenous cattlemen (PCR, paras. 5.20-5.21, 7.07, 8.06) Effect on project of changing land ownership policy (PCR, paras. 4.02-4.04, 5.02, 5.13, 9.12) Inadequate project accounts and audit (PCR, paras. 8.07, 8.09, 9.05) Inadequate technology transfer among regions through Bank staff (PPAR, paras. 21-26).  Project Performance Audit Memorandum PAPUA NEW GUINEA (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (Part) (Credit 175-PNG, Part) SUMMARY AND CONCLUSIONS OF THE PCR 1. The project, a follow-on to Credit 137-PNG, was to finance coconut estates, private and government cattle ranches, oil palm smallholdings and related infrastructure. The Credit Agreement, for US$5.0 million, was signed on January 30, 1970; the credit was fully disbursed by June 1976, two years later than the appraisal estimate. A Project Performance Audit Report has been issued on Credit 137-PNG and the oil palm component of Credit 175-PNG. This Project Completion Report (PCR) therefore discusses only the coconut and cattle components, to which $0.3 million and $1.7 million respectively were allocated in the Credit Agreement. 2. The development strategy for the agricultural sector of Papua New Guinea (PNG) evolved from Bank survey mission recommendations in 1963 and 1967. PNG's economic development plan for 1968/69 - 1972/73 followed in many respects the various recommendations of these Bank missions The principal objective was to expand production with the maximum participation of nationals. The project was identified as an omnibus agricultural credit project by the 1967 economic mission, to support export crop diversification (oil palm), expansion (coconut and rubber), and import substitution (cattle). 3. It was recognized that expansion of participation by nationals in some of these subsectors would be slow, and that where expatriate participa- tion was already established (e.g. beef, coconuts, rubber), increased produc- tivity should initially be fostered through this group while building up the agricultural services needed to expand and promote participation by nationals. Credits 137-PNG and 175-PNG were formulated to correspond with these objec- tives and the realities of the sector at that time. Therefore they supported participation by nationals in export-oriented oil palm development, while supporting expatriate participation in the beef and coconut subsectors. The Bank has supported smallholder livestock development (Credit 348-PNG) as a follow-on to the cattle component of the project. Breeding herds developed on ranches supported by Credit 175 supplied 60% of the stock for the smallholder project. 4. It must be recognized in reviewing the Credit that the project, and indeed the underlying development strategy embodied in the 1968-1973 develop- ment plan, was evolved at a time when independence for PNG was a certainty, but appeared to be 15 to 25 years away. It was only with the change of government in Australia in -1970 that a definite program for self-government leading to independence was evolved. The project spans the period from PNG as an administered territory to independence; consequently some of the judgments and decisions made during appraisal were no longer appropriate during implementation. - 2 - 5. Perhaps the best example is the judgment made to fund coconut new planting and replanting on expatriate estates. The Papua New Guinea Development Bank (PNGDB) received no loan applications, even though estate owners were aware that funding was available. An appraisal judgment made in 1968 (before the announcement of the timetable for independence) that private estate owners would be willing to take on the risk of a long-gestation (12-15 years) crop, proved invalid once the move to independence was announced in 1970. The lack of investment in the expatriate coconut estates contrasts with the expatriate cattle ranch owners' willingness, during the same period, to continue to invest. 6. The coconut estates were, in the main, located in a high population pressure area in East New Britain, while the cattle ranches were being developed in a low density area - the Markham Valley. In the coconut areas, the population is traditionally politically more activist and consequently was opposed to further expatriate development, unlike the population in the cattle development areas. The contrast between the situation of expatriates in the coconut and cattle subsectors is best illustrated by the fact that from 1975 coconut ranches have been bought for national investment by GOPNG. Cattle ranchers, on the other hand, have been assured that, for the present, there is no intention to conduct a similar program in the Markham Valley. 7. The change in political sensitivities did have an impact on the cattle component of this project: land set aside in 1969/70 by the Govern- ment of Australia (GOA) for expatriate ranch development was not leased out, in spite of applications from willing investors. These decisions by the Government of PNG (GOPNG) were made in 1973/74 when clearly it was no longer politically appropriate to lease large tracts of land to expatriates for long-term development. 8. For the private sector ranchers, the project assisted in the funding of 6,500 ha of pasture improvement on 13 existing ranches, the establishment of 9,900 ha of grazing on two new ranches, and the import of 2,460 breeder cattle. On two government ranches 6,000 ha of improved pasture and the import of 200 breeders was funded by the project. 9. Private ranch and herd development reached about 60% of appraisal estimates. The shortfall from appraisal estimates in ranch and herd develop- ment can be attributed to a lack of "suitable borrowers" among a very limited number of existing and potential expatriate cattle ranch operators, and in later years, unwillingness by GOA/GOPNG to allocate large tracts of land to expatriates. Consequently, total investments were lower than expected. On individual ranches participating in this program, in general, the levels of existing herds and breeders were higher than expected at appraisal, thereby lowering the level of investment needed to bring each ranch to its full capacity. However, individual ranches brought their herds up to stocking levels expected at appraisal. 10. The overall shortfall from appraisal estimates in herd development is also reflected in the growth of the national herd. However, this lower herd level has not adversely affected PNG, as fresh beef consumption has been - 3 - lower than projected at appraisal. It is possible that if appraisal estimates for both project ranches and the national herd had been achieved, the market would have been oversupplied, as beef consumption appears to be relatively inelastic to downward price movements. For smallholders there has also been an over supply of breeder cattle since that program developed at a slower rate than expected. 11. In reviewing the overall development of the cattle subsector it is clear that the development strategy was appropriate and successful, and that the cattle component of this project contributed to that development. From the evidence available to the PCR mission, it appears that certain ranches would have developed without IDA financing, whereas others clearly would not. A general conclusion on additionality or substitution is therefore not possible. 12. GOPNG officials indicated to the PCR mission that they considered the cattle component to be successful, and to have contributed to the develop- ment of the livestock sector. The PCR mission was impressed with the quality of ranch development and management on both the private and Government ranches. While there is no supporting quantitative evidence the PCR mission feels that the visible success of the cattle ranches could have acted as an inducement to nationals to participate in the follow-on smallholder program. 13. The project supported the establishment of three training schools for potential participants in the smallholder program. By the end of 1976 over 2,000 had received training, and the facilities have been used also to train about 300 rural development and livestock officers. DPI estimates that about 50% of the trainees are now actively working with smallholder cattle either as direct beneficiaries of the smallholder project, or employees of beneficiaries. This aspect of the project is successful and more than fulfills appraisal objectives. 14. From the limited cost data available, total reported costs of the private ranch component funded through this Credit amounted to 55% of appraisal estimates, in line with the lower level of physical development. While expen- ditures on the Government ranches appear to be 200% of appraisal estimates, it is unclear how reported expenditures correspond to those expected at appraisal. No total cost estimate can-be made, because cost reporting is incomplete, espe- cially for private ranches. 15. Lack of detailed monitoring of the cattle component, and sparse data collection by the project authorities, allowed the PCR mission to conduct financial analysis on only three private ranches - returns range from 12% to more than 50%. Each ranch has a' unique development history not typical of the other 12 funded under the project. The PCR mission, in the absence of reliable data, estimates the revised economic rate of return for private ranches to be in the 17-25% range, as compared to the appraisal estimate of 11%. 16. The major issue that this project and Credit 348-PNG have highlighted is the need to clarify the functions of PNGDB and the Department of Primary Industries (DPI) in the credit/extension services offered by GOPNG. The ongoing discussion within GOPNG on placing-,extension staff in PNGDB, or alternatively expanding the loan promotion and evaluation duties of existing DPI field staff, should be speedily resolved before further major programs are launched. These decisions should be made in the broader context of achieving effective promotion of both agriculture and industry sectoral development programs, within the constraints of a limited supply of extension and banking staff. 17. As this project was implemented under the particular circumstances of an evolving monetized economy and in a country about to become independent there are few transferable lessons. The fact that independence was considered as a relatively distant event at the time of appraisal, is important in assess- ing appraisal judgments. With hindsight, for example, it is easy to point to the absence of appraisal consideration of the need for extensive training for livestock extension officers, and especially training in project preparation suitable for loan financing through local financial institutions. This was remedied in the follow-on smallholder project, when appraisal was conducted with a clear understanding of the timetable for independence. - 5 - OED Comments Institutional Shortfall 18. The PCR completes an institutional story that was left unfinished in the PPAR on the smallholder oil palm settlement project (Report No. 1400, dated December 28, 1976). The PPAR showed that the Bank had hoped to tie the settlement project proposal to other production programs in an omnibus credit project, thus giving focus to otherwise fragmented efforts to build up the Papua New Guinea Development Bank. But the effort forced a delay in signing agreements on the first settlement credit (Credit 137-PNG), and contributed to the delay in signing Credit 175-PNG as well. The delays did not affect the Territorial Administration's timetable for settlement and livestock activities, which went ahead on schedule, so that in both cases the dates of effectiveness of the two credits lagged far behind the imple- mentation dates. 19. The administrative inconveniences of these delays would have been justified if the institution building objective had been achieved. However, the PCR makes it clear that, in the absence of coconut lending, and with the informality that characterized the relationship between Australian offi- cers in PNGDB and the 15 private expatriate ranching enterprises, the insti- tutional objective was not reached (PCR, paras. 8.03 and 8.05). The PCR says that the Bank's attempt to broaden the scope of PNGDB lending helped that organization to be "probably more valuable to IDA than to PNG" (PCR, para. 9.01). OED feels that conclusion is correct because by IDA becoming more familiar with PNGDB's appraisal and loan supervision procedures, this enhanced knowledge facilitated the appraisal of the smallholder project (Credit 348-PNG). Expatriate Record Keeping 20. The expatriate character of the project's livestock lending pro- gram, which tied Australian credit officers to Australian ranchers in the years before independence, reduces the transferability of this experience to other Bank projects. The solid technical achievements described by the PCR on the ranches must be associated in part with that expatriate owner- ship. What is remarkable is the poor quality of reporting on this ranch performance, a source of continuing concern to supervision teams (PCR, paras. 6.01, 7.08). OED is used to reading in PCRs of deficiencies in records, of on-farm changes, made and transmitted by Bank-supported credit agencies. But it was not anticipating that performance reporting would also be neglected in credit projects involving well-trained expatriates. The informal nature of the banker-rancher relationship explains part of the shortfall. The offi- cers knew what was happening on the ranches, though they did not bother to record it. -6- Technology Transfer 21. The announcement of early independence had severe repercussions on the project's coconut component. Expatriate estate owners proved unwil- ling to invest in new plantings when confronted by pressure from the land- hungry indigenous population -- the project site was in an area of relatively high population density -- and uncertain about the new government's land policy. The risks were multiplied by the long gestation period of tradi- tional coconut plantings and the declining long term trend of copra prices. Under existing circumstances, the expatriate's attitude is understandable. 22. OED would like to comment on one weakness of the appraisal design, however, which might have unnecessarily reduced the possibilities of lending for coconut development. We refer to the type of technology promoted by the project, in particular the exclusive attention given to traditional coco- nut varieties. As an alternative, the introduction of early maturing hybrid varieties would have reduced considerably the gestation period. We do not suggest that a shift to hybrids would have salvaged the coconut project since the estate owners willingness to accept new varieties on a large scale has never been tested, but the absence of hybrids in project design raises a question about constraints on the Bank's capability to promote technology transfer. 23. Traditional tall coconut varieties require six to eight years between planting and obtaining first yields, with full production not reached until fifteen years after planting. This would make any coconut project a marginal venture from the economic as well as financial point of view unless there occurred a significant rise in the price of copra. The Papua New Guinea project was no exception, with an economic rate of return esti- mated at only 9% during appraisal. 24. The extremely long gestation period of the tall coconut varieties had challenged researchers for many years. A major breakthrough in developing hybrid varieties was accomplished by the French Institut de Recherches pour les Huiles et Oleagineux (IRHO), resulting in palm trees which yielded after three to four years and reached full production in seven to nine years. The Bank knew of this development at least as early as 1965, when the new varieties were mentioned by the Permanent Mission to West Africa in a report identifying agricultural projects in the Ivory Coast. The Bank consequently assisted in the preparation of a coconut project mostly based on hybrids, and a project was appraised.and financed in 1967 (Loan 613-IVC). While the Bank is and has always been interested in technology transfer from research institutes to farmers, from nucleus estates to outgrowers and from project farmers to non-participating smallholders, the flow of technological infor- mation between country or regional offices within the Bank seems to have - 7 - been left more to individual staff initiative than to formal procedures.1/ 25. The view of project staff that farmers would not have been prepared to risk planting considerable areas with hybrid coconuts imported from West Africa seems more to underline OED's case than to refute it. While focus- sing on IRHO research results in the Ivory Coast, staff never explored the organization's accomplishments in adapting these varieties at their Pacific Region stations at Rangiroa and Saraoutou. Project staff also commented on hybrid planting material being in short supply in 1968. It is true that hybrid nuts were not available in quantities required during the late sixties, but the supply situation improved during the early seventies. To take advan- tage of this foreseeable development, project design should have allowed for at least partial planting of hybrids, an approach followed by the Ivory Coast project mentioned above. 26. It is interesting to note that due to the Bank staff's language composition, flow of information from the anglophone hemisphere to the francophone is much stronger than vice versa. That pattern may have weak- ened the coconut subproject. As late as 1973 the Bank turned down a request from Papua New Guinea's administration for financing hybrid coconut research (Supervision Report, August 1973) without even referring the Borrower to exist- ing coconut research at the French Pacific research stations on New Hebrides and Tahiti. Government should have been briefed on the extent of research already available, to avoid duplication of effort, and to identify the areas still open to adaptive research. Instead of rejecting the request on the grounds that results of hybrid research were already becoming available in the region, the Bank turned it down because of the high proportion of local costs and because research activities would have been extended beyond the credit's closing date. 27. What lessons could the Bank learn from this experience? One is that Bank contacts need to be intensified with agricultural research insti- tutes, especially in the francophone hemisphere. Closer cooperation has to be sought with research institutes during the time of project preparation and appraisal. Finally the flow of technical information within the Bank, to keep its technical staff constantly briefed on the fast advancing agri- cultural technology, needs to be strengthened. 1/ The then Division Chief of Projects,.who was a member of the 1971 super- vision mission, recommended hybrid coconut development to the PNG Director of Agriculture. As a result of this discussion, a PNG offi- cial visited the IRHO station in the Ivory Coast in 1972.  PAPUA NEW GUINEA (SECOND) AGRICULTURAL DEVELOPMENT PROJECT (Credit 175-PNG, Part) Project Completion Report TABLE OF CONTENTS Page No. 1. INTRODUCTION . . . . . . . ... . . . . . . . . . . . A 1 2. BACKGROUND. . ............. . . . . . . A 2 3. PROJECT FORMULATION ............. ... . . A 3 4. THE COCONUT COMPONENT . . . . . . . . . . . A 4 5. THE CATTLE COMPONENT . . . . . . . . . . A 6 The Subsector . . . . . . . . . .......... . .. A 6 Implementation . . . ...... . . . . . . . . . . . . . . A 8 Cattle for Smallholders .. . . . . . . . A 17 6. PROJECT COST . . . . . . . . . . . . . . . . . . . . . . . . A 19 7. ECONOMIC IMPACT .......... . . . . . . . . . . . . A 24 Macro Impact................ . . . . . . . . A 24 Microeconomic Analysis....... . . . . . . . . . . A 26 8. INSTITUTIONAL PERFORMANCE AND DEVELOPMENT . . . . . . . . . A 28 9. THE BANK'S PERFORMANCE . . . . . ........... . . A 30 This report is based on the findings of a mission which visited Papua New Guinea in January 1977. The mission comprised B. Berman (IDA) and M. Walshe (consultant): N. Owens assisted in the completion of this report, and L. Mitchell assisted in editing. -2- ANNEX TABLES Table 1 - Cattle Purchases Table 2 - Herd Development Summary Table 3 - Pasture Summary Table 4 - Schedule of Exchange Rates Table 5 - Project Costs Table 6 - Domestic Fresh Beef Consumption and Supply PROJECT MAP - IBRD No. 12828 PAPUA NEW GUINEA (SECOND)AGRICULTURAL DEVELOPMENT PROJECT (Credit 175-PNG, Part) Project Completion Report 1. INTRODUCTION 1.01 In August 1968 the Bank and the Government of the Commonwealth of Australia (GOA) began negotiations for a proposed IDA credit of US$6.0 million for oil palm, coconut and livestock development in the then GOA administered Territory of Papua New Guinea (PNG). The meetings were adjourned in September due to a delay in IDA replenishment, a temporary setback that the Bank had warned might occur. The Bank then rejected a GOA request for retroactive financing from the IDA credit when ultimately signed. The GOA in turn rejected a Bank offer to switch to IBRD loan terms, but accepted an alternative proposal to proceed immediately with a smaller credit for the oil palm component, and Credit 137-PNG for US$1.5 million was signed on January 21, 1969. Following the IDA replenishment in late 1969, a second credit, 175-PNG for US$5.0 million was negotiated and signed on January 30, 1970. This Credit included the beef and coconut components dropped from the First Agricultural Development Project, and expanded the oil palm development. 1.02 The second credit was fully disbursed by June 1976, two years later than the appraisal estimate. The oilpalm component was completed on schedule, and in conjunction with Credit 137-PNG is the subject of a Project Performance Audit Report./l This PCR will therefore discuss only the coconut and cattle components in detail, but basic project cost data will include the oil palm component for completeness. 1.03 A mission, consisting of Messrs. Berman (IDA) and Walshe (consultant) visited PNG in January 1977 to prepare this completion report with the coopera- tion of the Government of Papua New Guinea (GOPNG), through its Department of Primary Industries (DPI),/2 Livestock Section, and the Papua New Guinea Development Bank (PNGDB). /1 Report 1400 - Project Performance Audit Report: PNG-New Britain Smallholder Development Project, December 28, 1976. /2 Formerly the Department of Agriculture, Stock and Fisheries (DASF). - A 2 - 2. BACKGROUND 2.01 A Bank economic survey mission in 1963 /1 recommended the general policies for agricultural development in PNG; these policies were adopted in large measure, and progress towards achieving them was reviewed by a Bank economic mission in 1967./2 The economic development plan for PNG for 1968/69- 1972/73, followed in many respects the various recommendations of the Bank missions, taking into account current developments. The principal objective was to expand production with the maximum participation of nationals.13 2.02 In the agricultural sector, both the 1963 and 1967 Bank economic missions recommended further encouragement of coconut, cocoa, rubber and cattle production through a strengthening of extension services and the pro- vision of credit. The Bank's 1969 Economic Report /4 endorsed the development plan to diversify and increase the production of crops for export and for import substitution, such as oil palm, pyrethrum, and tea, and the increased participation by nationals in agricultural production. 2.03 To support this strategy the 1967 economic mission proposed a project for Bank/IDA financing which included crop diversification into smallholder oil palm, expansion of export-oriented production by nationals of coconuts, coffee and rubber, additional investment in the expatriate /5 sector with the rehabilitation of coconut and rubber estates, and the increase and expansion of cattle ranches. 2.04 It was recognized that the expansion of participation by nationals in these subsectors would be slow, and that where an expatriate sector was already established (e.g. beef, coconuts, rubber), increased productivity should initially be fostered through this group while building up the agricul- tural services needed to expand and promote participation by nationals. /l The 1963 mission's report was published: "The Economic Development of the Territory of Papua and New Guinea," Johns Hopkins University Press, Baltimore, 1964. /2 Report AS-1296, August 25, 1967. /3 In IDA/Bank terminology, "smallholder" and "national" are interchangable in the context of this report. /4 Report No. EAP-8a, September 3, 1969. /5 "Expatriate" was the term used in the appraisal report to indicate the "non-nationals" sector. For the purpose of this report "expatriate" will indicate, rather, a foreign source for the initial capital invest- ment on cattle ranches and coconut estates. -A 3- 2.05 Credits 137-PNG and 175-PNG were formulated to correspond with these objectives and the realities of the sector at that time. Therefore they supported participation by nationals in export-oriented oil palm develop- ment, while supporting expatriate participation in the beef and coconut subsectors. 2.06 To support further agricultural development in PNG, the Bank has financed two follow-up projects, the first /1 to promote smallholder livestock and the second /2 to expand smallholder participation in oil palm development. 2.07 It must be recognized in reviewing this Credit that the project, and indeed the underlying development strategy embodied in the 1968-1973 development plan, was evolved at at time when independence for PNG was a certainty, but appeared to be 15 to 25 years away. It was only with the change of government in Australia in 1970 that a definite program for self-government leading to independence was evolved. This project spans the period from PNG as an admin- istered territory to independence; consequently some of the judgments and deci- sions made during appraisal were no longer appropriate during implementation. 3. PROJECT FORMULATION 3.01 The conceptual basis for this project was developed by the 1963 economic survey mission, and placed into an "agricultural credit project" context in the 1967 Economic Report. This project was IDA's first involve- ment in PNG (viewing Credits 137-PNG and 175-PNG as financing for one activity). Only one Bank loan had been made in 1968 for telecommunications./3 3.02 The project as finally presented to the Board consisted of four subprojects: (a) a five-year tranche of a ten-year program for the planting and replanting of 10,000 acres /' of existing coconut estates; (b) the development and stocking of 150,000 acres of private beef cattle ranches and the improvement of government services to the beef cattle industry; (c) the settlement of 980 smallholders to grow oil palm over an area of about 8,000 acres on the island of New Britain adjacent to thelarea covered by the previous IDA project; and (d) the construction of a wharf at Kimbe to serve the oil palm subproject. /1 Credit 348-PNG signed April 1, 1973. /2 Loan 1333,-PNG signed November 19, 1976. /3 Loan 546-PNG signed June 28, 1968. /4 Since appraisal PNG has converted to the metric system; the PCR will follow this system in reporting results, but use the appraisal measures when reporting appraisal estimates etc. - A 4 - 3.03 The then Department of Agriculture, Stock and Fisheries (DASF) was jointly responsible with PNGDB for the implementation of the coconut, beef cattle and oil palm subprojects. The construction of the Kimbe wharf was supervised by GOA's Department of Public Works. Approximately $2.5 million of the proceeds of the credit was to be onlent to PNGDB, for a 25-year term with interest at 1-1/4%. PNGDB in turn would make loans to coconut planters, oil palm smallholders and beef cattle ranchers for periods of up to twenty years. In the case of cattle and coconut subprojects, PNGDB was responsible for loan processing based on a technical examination and recommendation of DASF field staff. Administration of oil palm loans was the responsibility of DASF staff assigned to the New Britain Oil Palm Project, with PNGDB acting only as the loan accountant/debt collector. 3.04 In the case of cattle and coconut development the borrowers were expected to self-finance 30% and 19% respectively of the cost. As the Credit was to finance only a five-year tranche of the coconut development and initial investments in new cattle ranches, it was expected that PNGDB would have to make additional funds available to project beneficiaries beyond the closing date of the project. 4. THE COCONUT COMPONENT 4.01 No record could be found of enquiries made between 1969 and 1973 for loans for replanting or new planting of coconuts. The appraisal report does indicate that by May 1969 PNGDB had received enquiries totaling US$146,000; however, both they and the completion mission could not identify exactly what this amount referred to, since current records only show loan enquiries at that time for tractors, plantation purchase and copra processing - none of which was eligible for financing under the Credit./l In fact no loans were made under this subproject and the funds were subsequently reallocated. 4.02 Opinions differ in PNG as to why no loans were taken up and include: (a) The Lack of Promotion of These Loans. Early Bank supervision missions indicated that "further advertising would be undertaken," but this never materialized, possibly due to the lack of a coordinat- ing officer for the coconut subsector. (b) Copra/Cocoa Prices. Between early 1971 and mid-1973 prices for copra were extremely low and the future for the plantations seemed bleak without rehabilitation using hybrid material. At that time cocoa prices were good, and planters took advantage of the favorable market by double-cropping coconuts with cocoa, thus achieving a far better return than could be achieved by coconut rehabilitation. /I The appraisal team leader reported to the PCR mission that they had been shown actual applications outstanding at the time. DPI reports that a 1967 survey indicated that "some 25 growers were interested in borrowing for replanting of 20,000 acres, and new plantings of 8,000 acres." - A 5 - (c) Local unrest in the Rabaul area, the focus of this component, caused by disputes over expatriate land ownership. (d) Independence. The timetable of independence was announced in 1970 and contributed to an unwillingness by expatriate owners to commit themselves to long-term investments. (e) Rate of Return. The long period needed for available coconut varieties to reach maturity gave a rate of return of about 9% - not a very attractive investment for an expatriate uncertain of his land ownership position and aware that independence would be granted before his replantings reached full development. Further- more, the borrower's equity contribution and interest payments cumulatively would reach 48% fl of the capital investment before the first operating profits would be realized in the eleventh year. It is the PCR mission's opinion that the political changes were the most significant factors in the lack of investment in this subsector. 4.03 The contrasting reactions of expatriate estate owners in the coconut subsector and the cattle ranch owners faced with the same political realities of independence were that while one group chose not to invest, the other expanded. The timing of the announcement for independence and the geographic location of the two groups were the two important factors affecting their courses of action, and both are interrelated. Most of the coconut estates are located in East New Britain; the land riots there in 1970 illustrate the political problems resulting from extensive expatriate holdings in a land pressure area. The politically active population there ensured that the concerns of the local people were reflected in GOPNG decisions as independence approached. The September 1975 supervision mission reported on the implementation of the Land Acquisition Act (passed in January 1975): "...so far fewer than 30 plantations (mainly copra and cocoa) have been acquired... All acquisitions except one have been by mutual agreement between vendor and Government. Priority for acquisition has gone to areas where customary owners are short of land for food and cash crop production, e.g. Gazelle in East New Britain. No private cattle.ranches have been nationalized and DASF does not anticipate any to be taken over in the foreseeable future." 4.04 This apparent difference in status of the cattle ranches is a result of the Markham Valley not being a land pressure area and the local population not being so politically active in pressing for expropriation. This does not imply that GOPNG has responded to unreasonable political pressure in the Gazelle Peninsula, but rather reflects its concern with the problems of land pressure. /1 In the appraisal farm model for new plantings of 400 acres, this cumula- tive deficit reached A$25,400 (Annex 1, Table 2). - A 6 - 4.05 The appraisal mission's assessment was that both estate and ranch owners would be willing to invest: their judgment was mistaken for the coconut subsector, but the role of exogenous factors must be borne in mind when evaluating their assessment in retrospect. The 1969 appraisal mission members met with coconut plantation owners in Rabaul and were assured of the need for such loan financing and that the loans would be taken up. They felt confident on both counts, though they reported to the PCR mission that political uncertainty with regard to private investment strategy had been discussed at these meetings. At a time when independence was considered a distant certainty, the appraisal mission concluded that planters were being unnecessarily cautious but had been persuaded by the mission of the value of loan financing for replanting and new planting. Either the announcement on independence reinforced the planters' uncertainty about the political climate in spite of the mission's positive stand, or the planters as a group recognized that the mission's proposal was good for the industry and did not wish to discourage IDA by their individual reluctance to take the investment risk. 4.06 It is difficult to judge the impact of the nonimplementation of this project on the coconut subsector. If replanting had gone ahead under this Credit, it might today be considered a constraint to the introduction of new varieties in the large plantations, the traditional leaders of the industry. In 1974, DPI began the development of a nursery for new hybrid species in preparation for a future coconut replanting program. 5. THE CATTLE COMPONENT The Subsector 5.01 The opinion of the Bank's 1963 economic survey mission was that the beef cattle industry in PNG was very promising and it recommended a program of development aiming at a tenfold increase in the cattle population in 10 years. However, by 1967 actual development was equivalent to only a threefold increase over a 10-year period, from 27,300 head in mid-1964 to 42,800 in mid-1967. To achieve the target growth rate, breeders from Australia should have been imported at a rate of 2,000-3,000 per year; however, in the three year period ending mid-1967 only 2,500 had been imported. At the time of appraisal the total herd was estimated at about 50,000 head, 80% owned by expatriates. 5.02 The 1969 Economic Report indicated possible causes for the slow growth in the national herd: (a) difficulty in obtaining imports of suitable breeders: Brahman crossbreeds from Queensland were considered to give the best performance in terms of weight gain and adaptability to PNG conditions; -A7 - (b) difficulties in obtaining suitable land to expand existing cattle ranches and establish new ones, since the traditional owners became increasingly reluctant to release their land;/1 (c) the lack of credit, particularly for the purchase of breeders. 5.03 The cost of importing breeding cattle into PNG was estimated to add between 60% and 80% to the purchase price in Australia. To encourage importa- tion GOA paid a freight and disease control subsidy, and conducted disease inspection and control at the port of origin. The Credit Agreement required GOA to review the subsidy with IDA before the end of 1972, with a view to terminating it by the end of 1973, or as soon as the development of the cattle industry warranted it. The subsidy was terminated in 1973. 5.04 At the time of appraisal PNG's development plan was for the total herd to reach about 300,000 by 1980, with 60% owned by nationals. Slaughter rates were expected to be between 30,000 and 40,000 head p.a. The appraisal report considered these targets optimistic but did not quantify its expectations. The 1969 Economic Report indicated that an interim goal was to reach a total herd of 138,000 by 1972/73, and that the program would require the importation of 10,600 breeders in the five-year period. 5.05 From 1969 to 1973 only about 2,200 j2 head were imported. However, as the following table shows, the national herd's growth rate, while being slower than expected, did increase significantly through a strategy of maximum retention of breeding stock. National Herd Growth % annual avg. growth 1969 1972/73 1975/76 1980/81 1964-1976 1969 Projection - 138,000 192,000 300,000 21.5 Actual 50,000 121,600 152,700 (est.) - 17.5 % Owned by expatriates 80 63 60 - f1 GOA was by this time also becoming increasingly reluctant to acquire land for subsequent lease to expatriates even though it had the legal right to do so under the Land Ordinance 1962-66. IDA's concern over this issue is reflected in the Credit Agreement's Section 4.04 obligating GOA to acquire ownership of 50,000 acres (20,250 ha) suitable for leasing to ranchers. While this type of clause is standard in credit agreements, it assumes significance in this project beyond its originally intended function (para. 5.13). /2 Last available figures quoted in "The Livestock Industry of Papua New Guinea," Malynicz, Herington & Eccles, (DPI, Konedobu, January 1977). -A8 - DPI's current national herd projections show that the target of 300,000 head could be reached by 1980/81, but this would require conservation of breeding stock, through a reduction in slaughtering for domestic consumption, and an increase in carcass beef imports. 5.06 The reasons for the slower than expected growth rate in the national herd are easily discernible, as compared to appraisal estimates or assumptions. These are: (a) slower than expected increase in consumption of fresh beef which, in part, explains a lower than expected level of investment in private ranches; (b) slower than expected rate of development of the smallholder sector with concomitant slaughtering of surplus ranch-produced breeders. In addition, during project implementation, the constraint on breeder cattle supply identified by the 1969 Economic mission (para. 5.02) was not relieved. In Australia during the period 1968 to early 1975, cattle prices were high and relatively few breeding stock were being placed on the market as herds were being built up in response to general optimism on the future of cattle ranching. 5.07 The causes for the two factors listed above are difficult to identify. The PCR will discuss only (a), as it directly relates to the performance of this component; (b) is currently the subject of -intensive investigation by DPI and PNGDB in relation to Credit 348-PNG, and will not be dealt with in this report. Implementation 5.08 To assist PNG in reaching these goals the subproject was designed to make credit available for the purchase of 4,400 head of imported breeding stock, ranch improvements including the opening up of 20,250 ha of new ranch land and - A 9 - the full stocking of about 40,500 ha of existing ranches.I The subproject was also to support importation of 200 /2 breeding heifers for two government ranches, on-farm improvements on those ranches, and the building of four farmer training schools. The latter component was in anticipation of the development of a smallholder livestock sector, to be supplied with breeders from the ranches assisted under this Credit. 5.09 Implementation Schedule./I While the appraisal used a basic 10,000 acre model, making it appear as if 15 ranches should have been funded in the two fiscal years 1969/70 and 1970/71, with a four-year development period on each ranch, comparison should be on the basis of the total ranch area and cattle imports. Using number of cattle imported as the measure of the speed of implementation, it is clear from Table 1 below that implementation was far slower and more variable over time than the appraisal team had expected. The size of each loan was smaller than expected and the 1974 supervision report indicates that PNGDB was having difficulty identifying "suitable borrowers," so as to draw down the full allocation under this category (Cat. 2). Government ranch development, however, proceeded according to the appraisal schedule, both for cattle imports and physical development. This was primarily because of delays in the processing of this Credit (see para. 1.01); GOA had begun ranch development prior to IDA agreeing to a second credit, and IDA funds were therefore used for limited retroactive financing and for the last phase of this development. /I Paragraphs 3.08 and 3.09 of the appraisal report main text indicate that "existing ranches will be interested in further development of some 100,000 acres ... it is forecast that ... 50,000, 30,000 and 20,000 acres would be developed successively in the first three years..." Yet in Annex 2, Table 4, the model of a 10,000-acre existing ranch shows pasture improvement on a total of only 4,300 acres over 4 years, implying a total implementation schedule of 43,000 acres; Annex 2, Table 1 indicates investments in pasture improvement for just 38,000 acres over five years, with yet another implementation schedule. Therefore the extent of pasture development and the implementation schedule intended by the appraisal mission is unclear. For cattle on private ranches the herd model indicates that each existing ranch would import 350 breeders in year 1, with new ranches at 400 breeders (Annex 2, Tables 3, 5). Yet model cost estimates for existing ranches indicate all stock purchased in the preproject year. In the total cost table (Annex 2, Table 1), two years of expenditure are shown, equivalent to the import of only 4,417 breeders, against the estimate (main text, para. 3.09) of 5,500 over three years. For the purpose of the PCR the implementation schedule implied by the annex cost tables will be used. /2 The appraisal report's main text and Annex 2 suggest 700 head; however, the cost tables indicate expenditure for 200 head in 1969. The first supervision mission reported that government ranches had "completed their cattle imports at 200 head, as indicated at appraisal." - A 10 - Table 1: CATTLE COMPONENT: IMPLEMENTATION SCHEDULE 1969/70 1970/71 1971/72 1972/73 1973/74 1974/75 Total A: Private ranches Number of loans approved /a 11 2 3/b 2/c - - 18 Cattle imported (head) - appraisal estimate /d 2,180 2,235 - - - - 4,415 - actual /e 550 83 930 603 1 293 2,460 B: Government Ranches Cattle imported (head) - appraisal estimate 200/d - - 200 - actual 200 - - 200 /a No explicit appraisal estimate. /b Includes one loan which subsequently lapsed. /c Includes loan renewals. /d See text for discussion of appraisal report inconsistencies. /e DPI's final report did not distinguish between imports and local purchase. From other sources total imports during the period 1969/70 to 1973/74 amounted to 2,236 head (including imports for Government ranches). - A 11 - 5.10 Private Ranch Development. Table 2 summarizes the key data for ranch development for the private sector. The main points that emerge from this table /1 are: (a) the project embraced 53% of the existing grazing area projected at appraisal and opened up 65% of the expected new ranch area; (b) when the appraisal estimates are adjusted for the reduced size, the herd development just exceeds appraisal estimates. These results are impressive since the number of breeding animals purchased was 40% lower than the adjusted model projections, although the opening breeding stock was 40% higher; and (c) when adjusted for ranch area, appraisal estimates of pasture improvement are greater than actual investments in pastures. However, the actual mix of fully improved and partially improved pastures shifted in favor of fully improved, accounting for the same herd size being carried on apparently less improved pasture. 5.11 Although the technical coefficients used in the appraisal model (Annex 2, Table 2) are high, they have been reached or exceeded on most ranches. For example, the weaning rate projected was 71%, but the PCR mission's estimate based on examination of ranch records suggests that an average of about 75% is being achieved for the project. The high technical coefficients reflect a very high level of ranch management and emphasize the importance of good management. In contrast the weaning rate in the smallholder sector is generally considered to be about 55% and the difference is mainly attributable to differences in management. 5.12 The outstanding question in both sectoral development and project performance is the lower rate of development of cattle ranches than expected -at appraisal, or reflected in national herd projections developed at that time. The loan-was introduced to the various Graziers Associations with a high level of promotion. The appraisal team had held a meeting with the graziers in Lae (reported to be the largest meeting ever held), and willing- ness to take up loans was clearly evident. The fact that 11 loans were made in 1969/70 also indicates the immediate acceptance of loan financing for herd expansion. However, it now appears that the first group comprised essentially all the eligible borrowers. While there were about 40 properties running cattle at that time, many as a sideline (under coconuts), DPI officials believe that "the remainder (of ranches) were simply unable to expand (herds) due to land limitations." The opening herd size, and level of breeders on existing ranches participating in the project, was higher than appraisal estimates Al A detailed variance analysis'has been performed on physical parameters, herd development, and .pasture improvement, and placed in the project file. The main conclusions are reflected here. - A 12 - Table 2: CATTLE COMPONENT: KEY DATA FOR PRIVATE RANCH DEVELOPMENT Before project (1969/70) After project (1975/76) Appraisal Actual as % Appraisal Actual as % estimate Actual appraisal estimate a Actual appraisal A: Ranch Development Existing ranch area (ha) lb 40,500 25,880 64 40,500 25,943 64 Existing grazing land (ha) /b 40,500 21,659 53 40,500 23,575 58 New ranches' area (ha) - - - 20,200 13,300 65 Improved pasture (ha) /c /hd e - - - 9,850 6,507 66 B: Herd Development Breeding cows /c 3,136 4,360 139 8,637 10,308 119 Total cattle C 8,954 9,584 107 22,313 22,196 Breeding stock purchased /c - - - 3,951 Ij 2,460 J 62 /a Based on Development Models Year 4, Appraisal Report, Annex 2. LL: Unclear in appraisal report if area is net or gross. /c Appraisal estimate adjusted for total size of ranches (64% of appraisal estimate for existing ranches, and 65% for new ranches). /d No pasture improvement expected on new ranches, but 451 ha was improved. fe Stated as incremental improvement. f See footnote /1, page 9 on appraisal report inconsistencies. A 91% imported, 9% local bulls. - A 13 - (Table 2). From DPI's comments it must be assumed that other ranches were at an even higher stocking rate, and could.only sustain a herd increase through pasture improvement, /1 or expansion. 5.13 As mentioned in para 5.02, leasing land for expatriate ranch develop- ment was becoming difficult, and while GOA complied with Credit Condition 4.04,/2 some of the land so secured was not taken up by new ranch development. There were two reasons for this (a) PNGDB/DASF could not find suitable entre- preneurs, (in both the financial and technical sense); and (b) when some suitable entrepreneurs indicated an interest in leases on two large blocks acquired by GOA for the project, these leases were not awarded by GOPNG, as they were unwilling, just prior to independence, to make additional land available to expatriates. Therefore, while GOA had strictly complied with the letter of credit conditions 4.04, the spirit of the agreement was violated by GOPNG. Given the changed political circumstances, it is not surprising that GOPNG was unwilling to honor a covenant entered into by a previous government./3 5.14 Pasture Improvement. The pasture areas embrace a broad range of environmental conditions. Rainfall ranges from about 3,200 mm in the wet forest areas to as low as 1,200 mm in the Imperata /4 dominant grasslands. In the wet forest areas improved pastures (tropical grasses and legumes) have to be established. Trees and scrub are cut and burned. Grass and legume seeds are scattered on the ashes without fertilizer. After about four years the pasture is cultivated with discs and resown when all the burned logs have virtually decomposed. Weed control is a constant problem and slashing yearly or biyearly is the most economic and preferred control method. If weed control is practiced, improved pastures can be highly productive with a safe carrying capacity up to about 1 to 1-1/2 animal units per ha. The cost of, clearing and establishing pasture on wet forest areas has increased about fourfold since 1970, from about K 60 to K 240 per ha. However, when estab- lished they can be maintained cheaply because fertilizer is not used and weed control (slashing) costs about K 2.5 per ha. A large proportion of the pasture areas under the project are natural Imperata dominant grasslands with rainfall ranging from about 1,300 to 2,000 mm p.a. (e.g. Markham Valley). In fact, Imperata dominant grasslands are found from sea level to nearly 2,000 m, with rainfall varying from 1,300 to 4,000 mm. /1 The constraints on pasture improvement are discussed in para. 5.15. /2 See footnote L1, page 7. /3 The PCR mission, after discussion with members of various supervision missions, has concluded that' IDA staff were aware of this apparent violation, but judged it inappropriate to raise it as a supervision issue, or even report on it in writing. /4 Imperata cylindrica - a coarse grass, low in nutritional value, that dominates after traditional slash and burn cultivation. - A 14 - 5.15 The strategy for utilizing Imperata pastures is not clear cut. Although it has been shown that large increases in animal production per ha are possible following improvement, this does not automatically ensure profitability, which is governed by many interrelated factors. It must be emphasized that unimproved pastures can sustain a high level of animal production. For example, a carrying capacity of about 1 animal unit to 1-1/2 to 2 ha can be sustained producing a steer of about 450 kg liveweight at 3 to 3-1/2 years. Therefore, the first priority in development should be to stock unimproved pasture to the safe limit of its carrying capacity. (This strategy was recognized in the appraisal report in the development model for new pasture areas where no provision is made for improved pasture.) When this is achieved, pasture improvement can be considered as the next step. Full improvement involving ploughing and sowing costs about K 180-200 per ha (US$230-256 per ha). If successful, animal production (liveweight) per ha can be at least doubled, but output and profitability will depend on good animal and pasture management coupled with appropriate increases in stocking rates. If these conditions are absent a strategy of development based on pasture improvement is subject to considerable financial and technical risk. Those ranches where this pasture development strategy was followed proved to be profitable if the owners were good ranch and stock managers. Those cases where pasture improvement was followed by poor management proved to be unprofitable and plagued with pasture maintenance problems, especially weed control. Partial improvement, based on scattering legume seed after burning, is an alternative where risks are small and investment is much less (about K 10-15 per ha). In this project the most profitable ranches were those with large areas and numbers of cattle but with some emphasis on semi-improved pastures (para. 7.10). 5.16 Table 3 shows the detail of the pasture development. While it is clear that the farmers did not follow the assumed strategy of very limited full pasture improvement in combination with extensive semi-improved pastures, they also limited the total area under any type of pasture improvement. Most of the pasture improved under this project was associated with exceptionally good management and was therefore successful. The area placed under both types of improved pasture amounted to about 66% of the adjusted appraisal estimate. The shortfall in total area reflects to some extent ranchers' reservations about the economics of improvement which are extremely difficult to quantify. - A 15 - Table 3: CATTLE COMPONENT: PRIVATE RANCH PASTURE IMPROVEMENT of which: Total Grazing Fully Partially area land improved improved Unimproved BEFORE PROJECT: EXISTING RANCHES Appraisal estimate: - ha 40,500 40,500/a - 4,000 36,500/a - as % total area 100% 100% - 10% 90% Actual: - ha 25,880 21,659 297 1,553 19,809 - as % total area 100% 84% 1% 6% 77% AFTER PROJECT: EXISTING RANCHES Appraisal estimate:/b - ha 40,500 40,500/a 800 14,170 25,530/a - as % total area 100% 100% 2% 35% 63% Actual: - ha 25,943 23,820 2,091 5,815 15,914 - as % total area 100% 91% 8% 22% 61% AFTER PROJECT: NEW RANCHES Appraisal Estimate: - ha 20,500 20,500/a - - 20,500 - as % total area 100% 100% - - 100% Actual: - ha 13,300 9,886 437 14 9,435 - as % total area 100% 74% 3% - 71% /a From appraisal report it is unclear if model for 10,000 acres in net or gross. A See footnote 1, page 9 for appraisal report inconsistencies. - A 16 - 5.17 Government Cattle Breeding Ranches. The project also supported completion of development on two government cattle breeding ranches begun while financing for the project was delayed (see para 1.01). This sub- project was implemented satisfactorily in accordance with the broad objec- tives of the appraisal report. Under the subproject the breeding herd on the government ranches increased by 965 cattle /I and the total number of cattle increased by 1,686 (Table 4). The two government ranches now comprise about 10,500 ha carrying a combined herd of about 4,600 head. Pasture improve- ment was carried out on 6,060 ha. Other developments included fencing (at least 35 km), stockyards and roads, but records were not available on quanti- ties. The appraisal report did not specify on-farm developments for government ranches but included only overall cost estimates for both on-farm development and support services. The appraisal report envisaged the importation of 200 breeding cattle for government ranches/2 which was achieved in FY69/70. /1 Urimo began to run buffalo in 1971/72 - these numbers are excluded from the analysis, but totaled 670 in 1975/76 with 170 breeders. /2 See footnote /2 page 9, on appraisal report inconsistencies. Table 4: CATTLE COMPONENT: KEY DATA FOR GOVERNMENT RANCHES Increase/decrease Before project (1969) Completion (1975/76) over project life Actual Appraisal as % of Actual Actual estimate Actual appraisal Breeding cows 1,514 2,479 - 965 - Total herd /a 2,917 4,603 - 1,686 - Imports - 200 200 200 100% rurnoff (animal) 763 1,560 - 797 - rotal grazing area 11,741 10,526 - (1,200)/b - Improved and semi- improved pasture 3,239 9,299 6,000 6,060 101% /a Excluding Buffalo on Urimo. /b About 1,200 ha excised from Baiyer River. - A 17 - 5.18 The project achieved its objectives of completing the development of two soundly based cattle breeding ranches. Baiyer River (which was visited by the PCR mission) is an impressive example of a practical and sensibly developed ranch; current sales of breeder cattle are reported to cover its operating budget. 5.19 The recording of information and its retrieval on government ranches left much to be desired, since it was not possible to find out what developments had taken place or to get physical or financial records. The main difficulty is the system of public service recording and accounting used. This problem should be considered when future projects are being formulated. 5.20 Farmer Training. Under the project provision was made for training potential cattle smallholders in cattle management. A classroom and equipment, together with dormitories, kitchen and a toilet block, were provided to each of three government ranches /1 in accordance with the appraisal report (including Baiyer River and Urimo on which on-farm development was also financed). Accommodation for 48 resident students was provided at Erap and Urimo and for 24 at Baiyer River. Instructors and support staff are employed at each location, and the training program is under the direction and guidance of an experienced expatriate Director. The training courses, aimed at smallholders receiving cattle under the Smallholder Livestock Project (Credit 348-PNG), provide rudimentary training in ranch management and practical animal husbandry. Particular attention is given to such matters as fencing, catching and handling animals, and pasture establishment and management. In the early stages the duration of the residential courses was three months but this was considered too long and was reduced to two months over the past two years. Furthermore 48 students are now considered too many for the courses and the number is presently limited to 24. Since 1975 the training staff have also provided two-week refresher courses at village level; these are proving both popular and successful and will play a more important role in the future. DPI is also considering conducting some initial training at the village level, in place of residential courses at the government ranches. 5.21 Under the project 2,189 farmers, as well as 276 rural development and livestock officers, have received residential training at three training centers. The Government runs one additional training center /2 not included in the project. The Director of Training estimated that about 50% of the trainee farmers were actively working with cattle and made use of the information covered in the courses. In summary, it is considered that this aspect of the project is successful and fulfills appraisal objectives. Cattle for Smallholders 5.22 An important objective of the project was that private ranches and government ranches financed under the project would supply breeding stock and steers to smallholders. It was recognized during appraisal that the limited supply of cattle was a major constraint to smallholder development /1 Baiyer River, Urimo and Erap (Map 12828). /2 Bisianumu - A 18 - and therefore ranch development was a prerequisite for any smallholder develop- ment. Thus, in 1973, Credit 348-PNG was signed to finance about 870 small- holders, with about 15 breeding cows per smallholder, the cattle being supplied by the private and government breeding ranches financed under Credit 175-PNG. This objective has been completely met and in fact private and government ranches have been able to supply more than adequate stock for the smallholder project. Government records, which are incomplete and under- state the position somewhat, show that private ranches supplied about 7,000 head to the smallholder sector from 1973 to end of 1976 and government ranches financed under the project supplied 5,311 head, which together amount to about 60% of the projected number provided under the smallholder project. - A 19 - 6. PROJECT COST 6.01 Project costs were estimated at appraisal on the basis of specified physical development. In attempting to establish the total cost of the project, the PCR found that: (a) in the private cattle ranches, no record was available of the assumed 30% equity contribution, or the physical development it represented; (b) for the government ranches 200% of appraisal costs have been reported, but its relation to actual physical development is unclear. Unit costs of development were not available; and (c) to exhaust the credit, IDA allowed GOPNG to charge road construc- tion costs not originally in the appraisal estimates. Consequently a cost figure associated with the physical development under review in this PCR is not available. Appropriateness of appraisal cost estimates, and the impact of cost increases on project viability cannot be examined. 6.02 For the smallholder oil palm component, unit cost changes have been discussed in the PCR and PPAR previously mentioned. For the cattle component PNGDB has supplied IDA wth a table of standard costs in 1970, and 1976; these are compared, where available, with appraisal unit cost assumptions in Table 6. 6.03 For the private ranch sector, total costs, as reported, are 55% of appraisal estimate, not significantly different from the scaledown in the size of this component. 6.04 Credit funds were reallocated between categories (Table 7), and the loan was fully disbursed by June 1976. (See Table 8 for appraisal estimate of fund disbursements and the actual disbursement schedule.) Initial delays in disbursements, as noted in the 1970 supervision report, resulted from GOA's accounting practices rather than delays in project implementation./l /1 Fully discussed in the PPAR, p. 19. - A 20 - Table 5: CATTLE COMPONENT: PROJECT COSTS (A$'000) /a Actual as Appraisal Actual /b % appraisal Private Ranches /c Ranch development 810 345 43 Cattle purchases 508 385 75 Subtotal 1,318 730 55 Government Ranches /d Ranch development 184 393 213 Cattle purchases 24 32 133 Subtotal 208 425 204 Government services /d 531 956 180 Total cost 2,057 2,111 103 /a Expressed in Australian $, because of numerous changes in US$ to A$ exchange rate during 1969-75. lb See full cost detail, Annex Table 5. /c Excluding self-financed operating costs. /d Actual costs through March 1975. - A 21 - Table 6: CATTLE COMPONENT: UNIT COST COMPARISON (Kina) Appraisal Actual/b estimate (1969)/a 1970 1976 Fencing (per km) - boundary 342 250 468 - subdivisional 248 241 250 Pasture (per ha) - fully improved 12 50 85 - partially improved/c 3 12 12 Cattle imports: - private ranches: breeders 110 135 126/f bulls 400/e 777/e 589/f - government ranches: breeders 200 131 bulls 400/e 690/e - Cattle sales - heifers (for smallholders) 100 95-110 125-159 - steers 137 115-135 145-165 - cull heifers 70 65- 90 85-115 - bulls 400 250-350/d 350-450/d Freight - ex-Australia 70 45 90 /a Annex 2, Table 2, converted to cost per metric unit where appropriate. /b Source: PNGDB. /c Burning and seeding only. /d Not specified for cull or breeding. /e Appraisal assumed local bulls. /f Last available cost data, June 1973. - A 22 - Table 7: ALLOCATIONS OF PROCEEDS OF CREDIT BY CATEGORY (US$'000) Category Appraisal /a Reallocation /b Actual /c Coconuts 1. Long-term PNGDB loans to coconut planters 300 - - Livestock 2. Long-term PNGDB loans to ranchers 1,075 840 910.6 3. On-farm investment, 590 670 666.1 including livestock & educational facilities and freight and disease control, in Borrower's ranches Oil Palm 4. Long-term PNGDB loans to smallholders 940 1,280 1,365.8 Borrower's Services to Smallholders: 5. Land subdivision and crop extraction roads 640 605 1,049.7 6. Agricultural extension services 360 315 307.9 7. Kimbe wharf 680 700 699.9 8. Unallocated 415 590 - Total 5,000 5,000 5,000 /d /a Schedule 1, Credit Agreement. /b June 1974. /c June 16, 1976. /d Exchange rate adjustment of US$518,880 not included. - A 23 - Table 8: SCHEDULE OF DISBURSEMENTS OF CREDIT /a (US$'000) Fiscal Appraisal Revised Revised Actual Actual as % year estimate /b estimate (3/74) estimate (3/75) disbursements appraisal or current estimate 1969/70 1,600 - - 1970/71 3,300 1,362 41 1971/72 4,190 2,120 51 1972/73 4,690 3,348 71 1973/74 5,000 3,790 3,666 97 1974/75 - 4,700 4,500 4,270 95 1975/76 - 5,000 5,000 5,000 /c 100 /a Excluding exchange rate adjustment of US$518,880.50. /b Not shown in appraisal reports; first record as "appraisal estimate" in the April 1975 supervision report. /c Closing date of June 16, 1976. - A 24 - 7. ECONOMIC IMPACT Macro Impact 7.01 Beef Consumption and Imports. The cattle component was justified on the basis of import substitution. While the appraisal report contained no projections for future carcass imports, it assumed that the turnoff from the ranches would substitute for imports of both beef carcasses and breeding stock. 7.02 The first record of beef consumption projections appear in the appraisal report for Credit 348-PNG. DPI has analysed these projections and actual development for the PCR mission (Annex Table 6). Consumption over the period 1969/70 to 1975/76, while growing at an annual rate faster than projected, (13.5% actual vs. 9.5% projected) began in 1969/70 from a base only 75% of that expected. The level reached by 1975/76 is therefore only 92% of the projection. However, the degree of substitution of domestic production for imports was marginally lower in 1975/76 than projected (67% projected, 61% actual). The impact of this project on supply is difficult to quantify: turnoff data from the ranches are not available. The impact of this project (and Credit 348-PNG) can be seen from the following summary results below, reflecting an increasing proportion of local beef in annual consumption. Annual average 1969/70 1975/76 % growth Tons % Tons % (1969/70 - 1975/76) Domestic slaughter 1,193 41 3,790 61 21.7 Imported carcass 1,690 59 2,399 39 6.5 Total fresh consumption 2,883 100 6,189 100 13.5 Over the period 1969/70 to 1975/76 the increase in annual domestic slaughtering is equivalent to 13,000 head, of which about 60-70% can be attributed to both projects. By comparison the increase in annual imports is equivalent to 3,000 head. 7.03 Prices. The New Guinea Graziers Association and the retail traders have practiced an informal price stabilization system. This did result, when beef prices in Australia were low, in short-term marketing problems for the graziers; the balance between import and local slaughter figures cannot, therefore, be interpreted as strictly reflecting choices based on prevailing market prices, although the extent to which importers responded to low Australian cattle prices in 1974-76 is difficult to measure. 7.04 A DPI official describes the price support situation as follows: "There was not at any time a formal Government move to support beef prices at a certain level. The New Guinea Graziers Association (NGGA) (strongly representative of all major ranch producers) had, however, always recognized the need to take a responsible approach to beef prices in negotiations with retailers. Once again, there was no formal mechanism for this - individual ranchers simply agreed that they would follow NGGA "policy" in price negotiations with the particular retail organization with whom they dealt. In arriving at the "agreed price" NGGA usually had prior discussions with major retail interests and with DASF. - A 25 - The attitude of the NGGA generally was to maintain reasonable stability in beef prices irrespective of world (and particularly Australian) market fluctuations. Hence, beef prices have shown a steady, and relatively small increase over the years. DASF was in general agreement with this philosophy and in fact saw a need to hold prices down in relation to prices of other goods and services. This is also reflected in prices for breeding cattle. Pricing "policy" for beef was therefore the result of close communi- cations between producer, retailer and Government, led primarily by the NGGA, and without the pressure of any formal measures by Government. As a result, PNG prices were well below those of equivalent classes of stock in Australia during the earlier boom there, but are now somewhat higher. Bull and cull cow beef prices have, however, been held back to a greater degree than prime steer prices. This policy occasionally resulted in very short term marketing problems for PNG producers, but all were satis- factorily resolved, usually at the ruling PNG beef price at the time. DASF played rather more of a directive role in price policy over breeders. Sale prices for Government-owned livestock were subject to (and still are) control by the Government's Supply and Tenders Board. DASF recommends these prices and did so, with IDA's agreement on the basis of ranch budgets and smallholder budgets. Once again the aim was to try and hold prices down and at a reasonably stable level, but bearing in mind the relatively high cost of imported breeders at the time./1 Private ranches were more or less bound to sell their breeding stock at similar prices. In fact Government stock purchasing officers were given a range of prices to work on which were the same as the approved selling prices of Government stock." 7.05 Supply to Smallholders. This project was also designed to supply cattle for smallholder development. Current DPI records indicate, for the period July 1973 to December 1976, that for the smallholder project, of a total of 20,603 head purchased for the smallholder project, ranches funded under this project supplied 6,990 head to DPI, and Government ranches funded under this project supplied 5,311 head (assuming that all recorded turnoff was for smallholder development). On this basis, project ranches contributed about 60% of the cattle for smallholder development. 7.06 Additionality vs. Substitution. The PCR mission could not find any meaningful measure of the extent to which the private ranches would have developed without IDA financing. Ranch owners, DPI and PNGDB staff had varying opinions, from "we couldn't have survived without it," through "major contributors to the development of the Markham Valley" to "ranchers were L_ Referring to the 1973-75 period. - A 26 - encouraged to utilize PNGDB loans rather than commercial bank loans, even though these could have been available." The PCR mission has concluded that with ranches under varying financial and economic circumstances, all these statements are valid: the PCR mission's guess is that about 40-50% of the development would have occurred without IDA financing, through commercial bank lending. 7.07 Transfer of Technology/Demonstration Effect. Aside from the obvious case of the funding of training for smallholders under the project, a number of the ranchers have been active in encouraging their staff to become cattle smallholders. In one instance a ranchowner has gone into a joint enterprise with the tribe from whom he has leased the land, and, aside from legal share- holder participation, the tribal leaders play an active part in herd management. Quantitative measures of number of employees or former employees of ranchers now active as smallholders was not available. However, the obvious success of cattle ranching must have acted as an inducement for nationals to enroll in the training schools and apply for smallholder cattle loans. Microeconomic Analysis 7.08 Financial and Economic Rates of Return./1 The necessary data on investment and operating costs, and sales returns is missing or very incomplete on most ranches. The PCR mission was able to collect data on three ranches that, with some additional estimation, allows for financial analysis. All three ranches are special cases, as indicated below,.and cannot be viewed as representing the total project. As data for all ranches is not-available, however, the economic rate of return was judged by the PCR mission to fall in the range of 17-25%, as compared with the appraisal estimate of 11% for all expenditures on this component (including Government ranches). This judgment was reached after analysis of data for "Ranch A," and adjusting sales income upwards by 15%; this adjustment parallels the appraisal assumption of import- substitution benefit; the use of this assumption in the PCR analysis is neces- sary due to the absence of current and historic cif/fob prices. 7.09 Ranch A. In 1969 grazing land with 20% partially improved pastures was being used to hold cattle for trading. A series of PNGDB/IDA loans brought the improved pasture land to 70% of total grazing, and established a permanent herd with imported stock. However, about 1975, the herd size was reduced in an attempt to lower the capital value for ease of sale. Current PNGDB/DPI herd and cost projections indicate that the financial rate of return is in the range of 12-15%./2 7.10 Ranch B. A large ranch, almost fully stocked at the time of appraisal, with no improved pasture, the strategy here was to use PNGDB/IDA loan funds to partially improve pastures, thereby increasing carrying capacity. Cattle imports amounted to only 5% of the opening herd, yet within the loan period fA Details of this analysis are in the project file. /2 Appraisal estimates of financial ROR are 11.1% (existing ranches) and 9.1% (new ranches). - A 27 - herd size increased by 75%. This strategy resulted in a 90% financial rate of return, reflection the large level of sunk cost, and high returns to minimum-investment in partial pasture improvements. 7.11 Ranch C. A new ranch financed through this project. To date it has had no significant sales; pastures are essentially unimproved, and the ranch is now reaching full carrying capacity. Current herd projections indicate a financial rate of return of about 12%./1 7.12 Government Ranches. Given the absence of any cost data, but observing that the herds are well managed and very productive, the PCR mission estimates the economic rate of return to Government ranch investment would be in the 20-25% range./2 This estimate is dependent on the assumption that Government development costs were the same as those of the private sector, and that the observed efficient management was similar to that on both Ranches A and B. /l Appraisal estimates of financial ROR are 11.1% (existing ranches) and 9.1% (new ranches). /2 No separate appraisal estimate was made. - A 28 - 8. INSTITUTIONAL PERFORMANCE AND DEVELOPMENT 8.01 The coconut and cattle components were to be administered directly by PNGDB, with DASF assisting the borrower to prepare technical submissions to PNGDB, and assisting PNGDB with technical supervision. Responsibility for the oil palm component was assigned to DASF, with mill management by the New Britain Palm Oil Development (NBPOD). PNGDB acted as loan accountant and debt collector from its local branch in Kimbe. This section will examine the performance of the PNGDB and DASF in the cattle subproject only. 8.02 Table 9 shows that PNGDB approved a total of 36 subloans for 17 ranches from 1968 to 1974. Initially these loans formed a significant propor- tion of lending to the cattle subsector, but rapidly dropped off in both numbers and as a proportion of lending, as the emphasis on smallholder loans began to significantly increase PNGDB's cattle portfolio. 8.03 In terms of institution building, this loan was of marginal import- ance, and its administration fell into normal channels PNGDB had already established with DASF. As planned at appraisal, DASF assigned specific staff to assist with subloan appraisal and supervision. There is some evidence, however, that the division of responsibilty was not clear and that DASF and PNGDB spent some time on institutional disagreements. This conflict was not resolved until the personalities involved had ended their appointments. 8.04 Because of the size of each borrower and the nature of the enterprise, officers of PNGDB and DASF were personally well acquainted with the progress of each subloan. This has had a negative effect in terms of reporting: on reading the files on each subloan, the PCR mission was struck by the fact that inspection reports were written with the assumption that the reader was very well informed on the past history and present conditions on the ranch. Especially conspicuous by its absence is any comparison of actual progress with past estimates. It is therefore impossible to gauge whether PNGDB and DASF officers found the borrowers' physical progress meeting either the projections of the original submission or any adjusted targets. 8.05 At a time when PNG was moving towards independence, this project, because of the size of each subloan, served to reinforce the traditional 'old boy' relationships between DASF and PNGDB rather than acting as a catalyst for institutional change. With the advent of independence and the repatriation of most of the very experienced staff, the 'old-boy' network has been replaced. In the newly restaffed institutions, much of the informal contact has been lost; however, in the now necessary systematization of reporting, and especially the technical and service relationships between DPI and PNGDB, little adjustibent to the new operating conditions is yet evident. While this has had little, if any, impact on this Credit, its effects have been felt in the smallholder project (Cr-348 PNG). GOPNG is fully aware of this situation, and has launched a systems study financed under that credit. 8.06 The development of farmer training schools through this Credit did benefit DPI's livestock extension unit. The facilities are used for training Table 9: PNG DEVELOPMENT BANK - 1IA ATTLE LOAN VOLUME APPROVALS 1967/68 1968/69 1969/70 1970/71 1971/72 1972/73 1973/74 1974/75 1975/76 No. Amt. No. Amt. No. Amt. No. Amt. No. Amt. No. Amt. No. Amt. No. Amt. No. Amt. (K'000) (K'000) (K'000) (K'000) (K'000) (K'000) (K'000) (K'000) (K'000) Total approvals 498 2,250 707 5,103 I 1,414 4,898 I 1,999 4,152 I 2,395 8,029 1,499 5,781 1,654 8,547 2,403 10,190 1,853 10,890 Total agriculture 412 908 I 505 1,896 I 983 2,294 1,398 2,069 [ 1,695 3,155 988 3,160 996 1,861 1,278 1,980 923 2,831 II I I II of which - Cattle 14 153 I 75 622 | 179 605 | 349 623 552 1,391 468 1,333 512 1,216 431 1,205 312 876 II I III - Cr. 175L 1 66 I 9 183 | 6 183 | 6 236 | 4 189 3 179 5 323 1 36 - - II I I I Cr. 175as1 I I I (by amunt) of : ------------------------ -_- _ _ ___ _-_-____-_-_____-- I 1 I II I I - Total approvals 2.9 | 7.6 3.8 I 5.7 | 2.4 3.1 3.8 0.4 - Agriculture 7.3 20.4 | 8.0 | 11.4 | 6.0 5.7 17.3 1.8 - Cattle 43.4j 62.4 30.3I 37.9 13.6 13.4 26.6 3.0 I IIII II /a These data are inconsistent with those presented in Table 1, also presented by PNC.DlB to PCR mission. Source: PNGDB. - A 30 - field officers: to date 30% of DPI's training activities have been for in- service training./l Nine hundred and seventy-six livestock assistants/rural development assistants (LA/RDA) have attended 46 courses, of which 10 courses for 296 LA/RDAs were held in the facilities funded under this Credit. 8.07 The project (and the Smallholder Oil Palm Project) highlighted weakness in DPI's internal reporting and accounting, and especially the need within government departments for project accounts. For this project, the lack of such accounts has, for example, made it difficult to adequately measure all expenditures on the government ranches. The lessons learned have been applied in the Popondetta Project (Loan 1333-PNG). 8.08 The outstanding questions in terms of institutional performance that are highlighted /2 are: (a) the relationship between PNGDB and DASF/DPI; and (b) the ability of PNGDB to promote borrowing. The first question is currently being examined through the systems study, funded under Credit 348. PNGDB is also discussing the desirability of expanding its own agricultural staff; this goes beyond the scope or expertise of the current study, and will have an important impact on all future PNGDB activities in agriculture. It must also be viewed in the light of desirable promotional activities that PNGDB (and currently DPI) should undertake through field level staff, to encourage smallholder agricultural productivity. In the past PNGDB has relied on informal contacts for loan promotion, or waited for the borrower to initiate an enquiry. DPI field level staff have often not considered it part of their extension task to initiate loan requests. A clear policy direc- tive is now imperative, before any further expansion in agricultural lending is considered. 8.09 Covenant Compliance. The Credit Agreement contained only two covenants specific to the cattle component. Section 4.04, on land acquisiton has been discussed in paragraph 5.13. Section 4.05, on the review of the freight and disease control subsidy, was complied with on time. Section 4.13 (b), a general covenant on auditing, appears not to have been enforced by IDA for the cattle component. 9. THE BANK'S PERFORMANCE 9.01 Identification to Negotiations. PPAR 1400 discusses the delays in project preparation and appraisal due to: (a) attempts by IDA to create an "omnibus credit project"; and (b) the lack of IDA funds. While the latter was unavoidable, IDA's move to broaden the scope of its lending was in /l An activity not expected at appraisal. /2 Current supervision of Credit 348-PNG also points to these questions. - A 31 - retrospect, probably more valuable to IDA than to PNG. Table 9 indicates that IDA's contribution to PNGDB's loan portfolio was relatively small. The expan- sion of the loan to cover agricultural subsectors in which PNGDB was already active allowed IDA to become familiar with DASF/PNGDB field appraisal and loan supervision capabilities. This experience facilitated the appraisal of the smallholder project (Credit 348-PNG). 9.02 No major issues were raised during appraisal. For negotiations, the Loan Committee requested IDA staff to explore the possibilities of changing the proposed terms of onlending from GOA to PNGDB. The Loan Committee indicated two alternatives: (a) GOA use the loan as an equity investment in PNGDB; or (b) PNGDB repay GOA as project beneficiaries repaid their loans. It is difficult to discern from the files the Loan Committee's reasons for this request. In the first instance the appraisal analysis indicated that PNGDB's capital structure was quite adequate to service the proposed debt; and secondly, if alternative (b) were to be followed, GOA would in effect be the beneficiary of an IDA credit for the period between bene- ficiary repayments and IDA's repayments. GOA's original plan to onlend part of this Credit to PNGDB at the same repayment terms as the original Credit had the benefit of retaining IDA funds in PNGDB after repayment by the original beneficiaries. Clearly this would be more beneficial to agricultural development in PNG than alternative (b). IDA's negotiation team finally agreed to again recommend the GOA proposal to the Loan Committee, and the recommendation was accepted. 9.03 Supervision. As this Credit was supervised in conjunction with Credits 137 and.348, with most missions also performing project appraisal and identification work, it makes an analysis of man-weeks of field super- vision impossible. Table 10 shows the schedule of missions and their duties. - A 32 - Table 10: SUPERVISION SCHEDULE Visit to PNG Date of report Supervision duties I June 1970 July 1970 Cr. 137/175 II May-June 1971 July 1971 Cr. 137/175 III May-June 1972 July 1972 Cr. 137/175 IV May 1973 Aug. 1973 Cr. 137/175/348 V March-April 1974 May 1974 Cr. 137/175/348 + appraisal /a VI Feb.-March 1975 April 1975 Cr. 137/175/348 + appraisal, reconnaissance /b VII Aug.-Sept. 1975 November 1975 Cr. 175/348 + identification/c VIII July 1976 August 1976 Cr. 175/348 + identification/c PCR Jan.-Feb. 1977 July 1977 Preparation/c /a Appraisal of Bialla/Popondetta Project. /b PCR 137 + Appraisal update (Popondetta) + Highlands Reconnaissance. /c Southern Highlands Project Identification/Preparation. 9.04 IDA's supervision of the oil palm subproject is discussed in the PPAR for Credits 137 and 175. IDA's supervision of the other subprojects will be discussed here. 9.05 The technical strength of DPI and PNGDB allowed IDA to supervise this project at annual intervals./l The issues raised in the supervision reports and letters to the Government indicate concern centered on (a) the accounting and reporting procedures; (b) possible additional subprojects designed to utilize surplus funds, and the subsequent reallocation of funds; and (c) loan adminis- tration. 9.06 Reports and Accounting. The reporting format was established during the June 1970 supervision mission. Because of interdepartmental coordination problems and the lack of accounting staff in DASF the first quarterly report was prepared as of the end of December 1970, only for the cattle subproject./2 /1 The exception is 1975 when Credit 348 was experiencing problems, and the proposals for the future Southern Highlands Project needed to be examined. /2 IDA's files show no record of the quarterly reports being received again until September 1971. The supervision mission in June 1971 noted that lack of accounting data made it impossible to complete the quarterly reports. - A 33 - The first report on the oil palm component was prepared as of the end of December 1971. Reporting was confined to the activities of PNGDB, DASF and NBPOD, and no measure of the progress or performance of the beneficiaries (cattle ranches) was prepared on a systematic basis. Only an occasional narrative report for PNGDB's cattle consultant (Dr. K.V.C. Kesteven) was received. By contrast, production output data for smallholder oil palm beneficiaries were readily available from NBPOD's -records. 9.07 An examination of PNGDB's files on each subborrower indicates that the loan supervisers did collect financial data to update their assessment of the borrowers debt service capacity, but no attempt was made by DPI or PNGDB, on a ranch by ranch basis, to systematically measure growth in herds, turnoff of cattle for smallholders, or supply to the beef markets. Except for the supervision report of September 1975, there are no other supervision documents assessing borrower, project and sectoral performance./l Credit 348 was at that time encountering problems, and the beneficiaries under Credit 175 were left with heifers that had to be fattened rather than sold as breeders. The supervision mission analyzed the supply of breeder heifers from Credit 175 in relation to the problems of Credit 348. 9.08 By contrast, stock records for the government ranches were available to the supervision missions, and herd development was charted in supervision reports, although it was not reported in the quarterly returns to IDA. 9.09 Reporting on costs included only those items reimbursable by IDA, and a total cost picture was therefore never constructed, even though the appraisal report indicated a 30% cost level to be borne by the ranch owners. Similar limited cost reporting was presented for government ranches and the training schools. 9.10 Reporting on the cattle component was designed to supply IDA with general performance benchmarks. These reports have been of little use in preparing the PCR, or allowing the mission to guage the pace or problems of implementation. By contrast the reporting on the oil palm component delivered by NBPOD did assist post-execution evaluation. Possibly today's increasing awareness by IDA staff on the necessity for ongoing project monitoring and evaluation, and the requirements for PCR's, makes it difficult to judge the quality of these reports in their context. 9.11 Additional Subprojects. As early as the 1970 supervision mission it was clear that the full credit could not be spent, and from then until 1976, DPI and IDA expended a considerable effort to identify and evaluate additional subprojects. Besides,the coconut subproject, cattle expenditures were lower than estimated at appraisal, and unit cost savings in oil palm further increased the surplus. Projects staff, and the then Area staff, encouraged this search, except for a period in 1971/72 when Area decided /1 The PCR mission was informed by a member of Supervision Mission IV, that this was addressed in the August 1973 Supervision Report; a copy of this report was not available to the PCR mission, as neither the authors nor the Bank's files have one. - A 34 - that with the shortage of IDA funds, the unspent proceeds should be canceled. GOA continued to offer alternative subprojects, and by 1973/74 IDA was indica- ting its willingness to be flexible in its consideration of alternatives. Finally GOPNG in 1975 claimed for earlier expenditures on oil palm road con- struction which effectively utilized the remaining funds. The volume of correspondence and supervision attention given to this question was perhaps warranted, as the surplus at one time appeared to be reaching about 40% of the total loan amount. 9.12 While it was clear that IDA intervention could not counteract the lack of progress in the coconut subproject, it is perhaps surprising that the supervision missions accepted the relatively low performance in the cattle subproject so unquestioningly. Although it is not reflected in supervision reports, perhaps IDA's appreciation of GOA's increasing sensitivity to the land tenure questions on the eve of independence accounts for this. However, this is no substantive discussion of appraisal targets vs. actual results,/1 and supervision reports concentrate on surplus fund reallocation and the management of existing loans, rather than a recognition of only 50% lending achievement by PNGDB. Consequently, there is no recognition of the slowdown in sectoral growth: this is compounded in the later years by the smallholder problems, which superficially indicate an oversupply of breeding stock. 9.13 Loan Administration. The promotion and administration of loans by DPI/PNGDB was identified to GOPNG as a problem by the September 1975 supervision mission as a result of IDA's experience under the smallholder project, Credit 348. Cr. 175 supervision missions reported on these problems, because the proposed systems study was to have been financed from this Credit, though subsequently it was decided that Credit 348 would finance this. A full discus- sion on this topic cannot yet be conducted as the various studies are still under way, and should be reviewed rather in the more appropriate context of the Borrower and Bank's performance under Credit 348. /1 The 1973 Supervision is reported to have addressed this question - see footnote /1, page 33. ANNEX Table 1 PAPUA NEW CUINEA AGRICULTURE DEVELOPMENT PROJECT II (Credit 175-PNG) PROJECT COMPLETION REPORT Cattle Subproject: Cattle Purchases Imports Local Total Heifers Bullb Heifers Bulls Heifers Bulls Private Ranch Ampan - - - - - Hammersley 97 6 - - 97 6 Vunakambi 38 2 10 10 48 12 Wanaru - - - - - - Doa Mants 32 - 35 2 67 2 Awilunga 97 6 - - 97 6 Takoa 150 - - - 150 - Jackson 350 11 - - 350 11 Jimi River 950 36 100 3 1,050 39 Kanosia 301 12 - - 301 12 Dumpu 200 10 - 12 200 22 Gili Gili 90 7 - - 90 7 Munum 125 24 - - 125 24 Murray 100 7 - - 100 7 N. Kindino 78 2 15 15 93 17 H1agita 100 4 - - 100 4 Rigo 100 - 100 - 200 - Total 2 127 260 42 3 169 Government Stations Baiyer R. 190 } 10 - - - - Urimo PAPUA NEW GUINEA AGRICULTURE DEVELOPMENT PROJECT II (Credit 175-PNG) PROJECT COMPLETION REPORT Cattle Subproject: Herd Development Summary Before loan (1969/70) Current (1975/76) Changes Breeders Total Breeders Total Breeders Total herd herd herd Private Ranch Ampan 253 491 170 396 -83 -95 Hammersley 125 241 229 434 +104 +193 Vunakambi 10 29 39 102 +29 +73 Wanaru 164 245 2 7 -162 -238 Doa Mants 3 26 88 184 +85 +158 Awilunga 226 397 531 996 +305 +549 Takoa 883 1,865 1,105 2,333 +222 +468 Jackson 273 420 356 994 +83 +574 Jimi River - - 1,171 2,878 +1,171 +2,878 Kanosia - - 304 542 +304 +542 Dumpu 1,500 4,018 3,590 7,083 +2,090 +3,065 Gili Gili 65 152 220 408 +155 +256 Munum 604 1,268 1,800 4,212 +1,204 +2,944 Murray 107 213 300 614 +193 +401 N. Kindino 87 119 32 81 -55 -38 Hagita 37 49 140 309 +103 +260 Rigo 23 51 231 623 +208 +572 Total 4,360 9,584 10,308 22,196 5,956 12,562 Government Stations Baiyer R. 1,514 2,917 2,250 4,699 +736 +1,782 Urimo /a 632 (32) 1,451 (41) 776 (169) 2,200 (672) +144 (137) +749 (631) Grand total /a 6,506 (32) 13,952 (41) 13,334 (169) 29,095 (672) 6,836 (137) 15,043 (631) /a ( ) = buffalo which are included in total. PAPUA NEW GUINEA AGRICULTURE DEVELOPMENT PROJECT II (Credit 175-PNG) PROJECT COMPLETION REPORT Cattle Subproject: Pasture Summary (ha) Before project After project Increases Date of Total Grazing Unim- Partially Fully Total Grazing Unim- Partially Fully Improved Fully loan land land proved improved improved land land proved improved improved pasture improved Private Ranch May 71 Ampan 596 571 104 467 - 596 568 89 379 100 12 100 Dec 71 Hammersley 329 329 256 73 - 592 411 123 188 100 215 100 Jul 73 Vunakambi 513 114 114 - - 513 114 74 40 - 40 - Dec 69 Wanaru 508 320 223 - 97 508 326 27 200 99 202 2 Jan 70 Doa Mants 265 254 254 - - 265 254 42 212 - 212 - Oct 70 Awilunga 394 354 250 104 - 394 354 111 243 - 139 - Jun 71 Takoa 6,644 6,515 6,011 504 - 6,644 6,515 5,930 - 585 81 585 Dec 69 Jackson 3,434 3,434 3,434 - - 3,434 3,434 2,447 904 83 987 83 Aug 71 Jimi River - - - - - 8,190 8,186 8,125 - 61 61 61 Jun 74 Kanosia - - - - - 5,110 1,700 1,310 14 376 390 376 Dec 68 Dumpu 6,986 6,986 6,886 100 - 6,986 6,986 4,718 2,268 - 2,168 - Feb 69 Gili Gili 740 150 100 50 - 740 740 300 360 80 390 80 Mar 69 Munum 1,970 729 324 205 200 1,970 1,970 570 500 900 995 700 Feb 69 Murray 871 871 871 - - 871 871 771 40 60 100 60 May 69 N. Kindino - - - - - 245 245 - 245 - 245 - Sep 71 Hagita 2,025 640 590 50 - 2,025 640 490 100 50 100 50 Jul 69 Rigo 405 392 392 - - 405 392 222 136 34 170 34 Total 25,680 2 19,809 1,553 297 39.488 33,706 25,349 5,829 2528 6 2 Government Stations Baiyer R. 6,073 6,073 2,834 --- 3,239 --- 4,858 4,858 1,093 --- 3,765 --- ----- 526 -- e Unmo 5,668 5,668 5,668 - - 5,668 5,668 134 --- 5,534 -- 5,534 -- ANNEX Table 4 PAPUA NEW GUINEA AGRICULTURAL DEVELOPMENT PROJECT II (Credit 175-PNG) PROJECT COMPLETION REPORT Schedule of Exchange Rates /a US$ A$ A$ K Date of Change Through 12/23/72 1.000 0.840 1.191 1.000 12/23/72 - 03/13/73 1.000 0.784 1.275 1.000 03/13/73 - 09/09/73 1.000 0.704 1.420 1.000 09/09/73 - 09/30/74 1.000 0.675 1.488 1.000 09/30/74 - 06/30/75 1.000 0.752 1.330 1.000 06/30/75 - 07/26/76 1.000 0.787 1.000 1.000 /b 1.270 1.000 1.000 1.000 07/26/76 - 11/29/76 1.000 0.809 1.000 0.952 1.235 1.000 1.050 1.000 11/29/76 - 12/13/76 1.000 0.983 1.000 0.847 1.017 1.000 1.181 1.000 12/13/76 - 12/29/76 1.000 0.950 1.000 0.866 1.052 1.000 1.155 1.000 12/29/76 - 1.000 0.920 1.000 0.882 1.087 1.000 1.134 1.000 /a Source of information: Report No. 1150-PNG; Papua New Guinea: Economic Situation and Development Prospects (Green Cover); 1976 Data: PNG Ministry of Finance, IMF. /b Kina first issued in April 1975. ANNEX Table 5 Page 1 PAPUA NLW GUINEA AGRICULTURAL DEVELOPMENT PROJECT 11 (Credit 175-PNG) PROJECT COMPLETION REPORT Project Costs (A$'000) /a Reported cost Appraisal Reported as % appraisal estimate cost estimate A. CATTLE Private Ranches /b On-farm Developments Fencing 65 77 118 Yards 47 22 47 Water 60 16 27 Pasture 340 116 34 Equipment 110 87 79 Buildings 109 28 25 Cattle 508 385 75 Contingencies 79 - - Subtotal 1.318 730 55 Other Costs Operating costs financed by subborrowers 390 n.a. - Total private 1,708 n.a. - Government Ranches On-farn Developments /c Fencing 16 23 143 Yards 4 24 600 Water 10 9 90 Pasture 120 118 98 Equipment 20 27 135 Cattle 24 32 133 Contingencies 14 - - Other - 192 - Subtotal 208 425 204 /a Australian dollars used because of numerous exchange rate changes against US$ between 1969 and 1976. /b /c /d Source: Reported costs expressed in Kina (but expended when K 1 - A$1) as per: Withdrawal application # Dated /b 24 March 1976 /c 18 September 1974 21 May 1975 ANNEX Table 5 Page 2 Proiect Costs (continued) (A$'000) /a Reported cost Appraisal 'Reported as % appraisal estimate cost estimate Services Freight & disease control /d 263 233 89 Freight on Government cattle /d 16 239 1,493 Training facilities /c 115 228 198 Supervision /d 110 256 232 Contingency 27 - - Subtotal 531 956 180 Total Government 739 1,381 186 TOTAL CATTLE 2,447 n.a. - B. OIL PALM Smallholding Development b Settlement expenses 252 268 106 Housing 343 686 200 Field development 664 1,224 184 Contingency 51 274 537 Subtotal 1,310 187 Smallholder labor 434 n.a. - Total 1,744 n.a. - Government Services /b Subdivisions, roads & road maintenance 865 1,512 175 Extension services 485 390 80 Social services 191 355 185 Contingency 49 - - Total 1,590 142 PNGDB expenses 195 -/e - TOTAL OIL PALM 3,529 n.a. - C. KIMBE WHARF /b 890 863 97 TOTAL PROJECT COST 6 If n.a. /g -Lg /a Australian dollars used because of numerous exchange rate changes against US$ between 1969 and 1976. /b /c /d Source: Reported costs expressed in Kina (but expended when K 1 - A$I) as per: Withdrawal application # Dated 24 March 1976 /c 18 September 1974 /d 21 May 1975 e Included in smallholding development. Jf Excluding coconut subprojects JA Recorded cost - A$7,683,000, 112% of cost of funded subprojects. n.a. - not available ANNEX Table 6 PAPUA NEW GUINEA AGRICULTURE DEVELOPMENT PROJECT II (Credit 175-PNG) PROJECT COMPLETION REPORT Cattle Subproject: Domestic Fresh Beef Consumption and Supply (tons) Average Annual % 1969/70 1970/71 1971/72 1972/73 1973/74 1974/75 1975/76 Growth Rate 1969-75 Domestic Slaughter /a 1972 projections /b 1,980 2,271 2,657 3,115 3,576 4,034 4,524 14.1 Actual /c 1,193 1,459 1,750 2,317 3,094 3,302 3,790 21.7 - as % projection 60 64 65 74 86 81 83 Import - Fresh, chilled, frozen fa 1972 projections 1,875 2,298 2,301 2,257 2,234 2,222 2,173 2.6 Actual 1,690 2,022 2,116 1,900 1,304 1,912 2,399 6.5 -- -is % projection 90 88 92 84 58 86 110 Total 1972 projections 3,855 4,564 4,958 5,372 5,810 6,256 6,697 9.5 Actual 2,883 3,478 3,866 4,217 4,390 5,214 6,189 13.5 -- as % projection 75 76 78 79 75 83 92 /a As carcass weight. /b At 225 kg carcass weight per head, except 1975/76 at 200 kg per head. /c It is unclear why "1972 projections" have projected figures for 1969/70 to 1971/72. Source: DPI.  I RIlAN J AY A lIN DO0N ESIlA) n 1ý 1 1 1,y 7 - - - - - - - - - - - - cQ' J: 0 JzQ o - ,, > el - - - - - - - - - 74 KD - - - - - - - - - - - - - - - b -ts 0/0 700 - .. .. .. .. . .. .. .. . o Zz 3 2 ~