Papua New Guinea - Third Agricultural Credit Project

Document of The World Bank FOR OFFICIAL USE ONLY MICROFICHE COPY Report No. 10608-PNG Type: (PPR) Report No. 10608 ALEGRE, I / X31755 / T9059/ OEDD1 PROJECT PERFORMANCE AUDIT REPORT PAPUA NEW GUINEA THIRD AGRICULTURAL CREDIT PROJECT (LOAN 2624-PNG) APRIL 28, 1992 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 no! otherwise be disclosed without World Bank authorization. CURRENCY EQUIVALENTS Name of Currency: Kina (K) Rate of Exchange: Year Average Appraisal (1985) US$1.00 - K 0.94 Intervening (1986-88) = K 0.91 Completion (1989) = K 0.86 ABBREVIATIONS AND ACRONYMS ABPNG - Agriculture Bank of Papua New Guinea ADB - Asian Development Bank BANK - World Bank Group DAL - Department of Agriculture and Livestock ERR - Economic Rate of Return FAO/CP - Food and Agriculture Organization/Cooperative Program FFB - Fresh Fruit Bunch FRR - Financial Rate of Return IASER - Institute of Applied Social and Economic Research IBRD - World Bank IER - Impact Evaluation Review MIS - Management Information System NBPOD - New Britain Palm Oil Development, Ltd. NPMA - National Plantation Management Agency OECF - Overseas Economic Cooperation Fund OED - Operations Evaluation Department OFL - Over Free Limit PCR - Project Completion Report PNG - Papua New Guinea PNGBC - Papua New Guinea Banking Corporation PNGDB - Papua New Guinea Development Bank PPAR - Project Performance Audit Report SAR - Staff Appraisal Report TA - Technical Assistance FISCAL YEAR Government of Papua New Guinea Agriculture Bank of Papua New Guinea January 1 - December 31 WEIGHTS AND MEASURES Metric System THE WORLD BANK FOR OFFICIAL USE ONLY Washington, D.C. 20433 U.S.A. Office of Directo-Ceneal Opetatins Evaluation April 28, 1992 AEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report on PAPUA NEW GUINEA Third Asricultural Credit Prolect (Loan 2624-PXG) Attached, for information, is a copy of a report entitled "Project Performance Audit Report on Papua New Guinea: Third Agricultural Credit Project (Loan 2624-PNG)" prepared by the Operations Evaluation Department. Attachment 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. FOR OFFICILL USE ONLY PROJFCT ERFORM IL AUDIT REPORT PAPUA NEW GUINEA THIRD AGRICULTURAL CREDIT PROJECT (LQAN 2624-PG) TALE OF CONTETS Pao no. Preface . . . . . .. . . .. . .. . .. . .. .. . ... . . . Basic Data Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . it Evaluation Summary . . . . . . . . . . . . . . . . . . . . . . . . . v I* BAKRON . . A. Preparation . . . . . . . . . . . . . . . . . . . . . . . . . 1 B. Objectives . . . . . . . . . . . . . . . . . . . . . . . . . 2 C. The Project . . . . . . . . . . . . . . . . . . . . . . . . . 3 II. IMPLEMENTATION EXPERIENCE ...... . . . . . . ...... 4 A. Project Costs and Dollar Disbursements . . . . . . . . . . . 4 B. On-Lending Component . . . . . . . . . . . . . . . . . . . . 5 C. Institutional Component. ................ 7 III. PROJECT OUTCOME . . . . . . . . . . . . . . . . . . . . . . . . 10 A. ABPNG's Financial Position . . . . . . . . . . . . . . . . . 10 B. Agricultural Impact . . . . . . . . . . . . . . . . . . . . . 11 The Commodity Price Decline . . . . . . . . . . . . . . . 11 Cocoa Farms . . . . . . . . . . . . . . . . . . . . . . . 11 Oil Palm Blocks . . . . . . . . . . . . . . . . . . . . . 12 The Socio-Economic Impact of Rural Credit Study . . . . . 17 Other Project Investments . . . . . . . . . . .. . . 18 A Follow-on Project . . . . . . . . .. . . . . . . . . . 18 IV. FIRDINGS AND ISSUES . . . . . . . . . . . . . . . . . . . . . . 19 A. A Good Project under Stress. ........... ..... 19 B. Role of Interest Rates . . . . . . . . . . . . . . . . . . . 21 C. Low Equity Contributions to High Input Investments . . . . . 22 D. Role of Technical Factore ................. 23 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. Paze No. TABLES 1. Allocation of Amounts Disbursed for Sub-Loans . . . . . . . . . 5 2. Number and Average Value of Subloans ....... ..... 6 FIGURES 1. Cocoa and Coffee Prices .................... 15 2. PalmOilandCopraPrices ................... 16 Hap IBRD 17314R PROJECT PERFORMANCE AUDIT REPORT PAPUA NEW GUINEA THIRD AGRICULTURAL CREDIT PROJECT (LOAN 2624-PNG) PREFACE This is a Project Performance Audit Report (PPAR) on the Papua New Guinea Third Agricultural Credit Project, involving an IBRD Loan in the amount of US$18.8 million. The objectives were to increase exports, reduce imports, improve rural incomes and strengthen the Agricultural Bank of Papua New Guinea (ABPNG) as a development financing institution. The Loan was approved on September 19, 1985. It was closed after an extension of six months, fully disbursed, on December 31, 1989. The final disbursement was made on January 16, 1990. The PPAR is based on the Project Completica Report (PCR) prepared by the Asia Regional Office and submitted to the Board on April 30, 1991, the Borrower's Socio-Economic Impact Study and PCR submitted to the Bank in 1989, the Staff Appraisal Report (SAR), the President's Report, the Loan documents, a report of the Executive Directors' meeting at which the project was considered, a study of project files, and discussion with Bank staff. An OED mission visited Papua New Guinea in November 1991 and discussed the effectiveness of the Bank's assistance and project execution with ABPNG and other relevant agencies. ABPNG's kind cooperation and valuable assistance in the preparation of this report are gratefully acknowledged. The PCR provides a good account and assessment of the project experience, and the performance of the Bank and the project executing authority. The PPAR brings the PCR up to date and elaborates on particular aspects of the overall lending period, especially on the impact of a decline in commodity prices on farmer incomes and the quality of ABPNG's portfolio. Following standard OED procedures, copies of the draft were sent to the Government on January 31, 1992. No comments were received. PROJECT PERZORMANCE AUDIT REPORT PAPUA NEW GUINEA THIRD AGRICULTURAL CREDIT PROJECT (LOAN 2624-PNG) BASIC DATA SHEET Key Project Data Appraisal Actual Estimate Performance 2 Project Costs (US$m) 31.1 30.5 97 Loan Amount (US$M) Disbursed 18.8 18.8 100 Area Planted (ha) Cocoa 4,580 4,826 105 Oil Palm 4,800 3,900 81 Beneficiaries Cocoa 2,030 1,323 65 Oil Palm 1,200 956 80 Poultry and Piggery 21 70 333 Transport n.a. 537 Economic Rate of Return Project below 10 Cocoa and Oil Palm Models (Z) 22-38 below 10 Financial Rate of Ret:rn Cocoa and Oil Palm (2) 26-45 near 0 Cumulative Actual and Estimated Disbursements Y 86 FY 87 L.4 I ..9 I..2 Appraisal Estimate (US$m) 2.40 6.20 9.00 1.20 Appraisal Cumulative 2.40 8.60 17.60 18.80 Actual - 3.60 7.30 5.80 2.10 Actual Cumulative - 3.60 10.90 16.70 18.80 Actual as Z Appraisal 0 42 62 89 100 Date of Last Disbursement: January 16, 1990 - iv - Proiect Timetable Item Date Planned Date Actual Identification - May 1982 Preparation - November 1982 Appraisal - May 1983 Post-Appraisa) - February/March 1985 Negotiations - July 9-12, 1985 BoarL. spproval - September 19, 1985 Loan Signature - May 2;, 1986 Effectiveness August 23, 1986 August 29, 1986 Closing June 30, 1989 December 31, 1989 Project Completion December 31, 1988 December 31, 1989 Staff Inputs Stage of Project Cycle I lanned Revised Final Through appraisal n.a. n.a. 76.1 Appraisal through Board approval n.a. n.a. 8.0 Board approval through effectiveness n.a. n.a. n.a. Supervision n.a. 30.0 26.0 Total 110.1 Mission Data Stage of Project Cycle Vonth/ No. of Days 14 SpecializatiT7 Performance -ating Year Persons Field ' Represented" Status Appraisal May 83 4 21 EC,EN,AG,FA Post Appraisal Feb/Mar 85 2 11 EClFA Supervision 1 1 May 87 1 17 AG I Supervision 2 4 Oct/Nov 87 1 23 AG 1 Supervision 3 Jun 88 1 12 EC 1 Supervision 4 Jun/Jul 89 2 10 EC,FA 2 Completion May 90 2 22 EC,AG - VI PCR mission estimates based on Bank*s project files. Specializations EC * Economist, EN - Engineer (Irrigation); AG - Agriculturist; FA - Financial Analyst. £t Performance Rating: 1 - Problem-free or minor problems; 2 a Moderate problems; 3 - Major problems. Development Impact has however been rated "2" in all supervision reports. Supervision mission also discussed a possible Fourth Credit Project. -v - PROJECT PERFORMANCE AUDIT REPORT PAPUA NEW GUINEA THIRD AGRICULTURAL CREDIT PROJECT (LOAN 2624-PNG) EVALUATION SUMMARY Introduction weighted avarage of prices fell by about 602 in the period 1985-1990 1. The Third Project would carry-on (para 3.3 and Figure. 1 and 2). There providing term credit for commercial were many positive features in this crops, livestock and some agricultural project experience, which are vel services. The emphasis vae on tree documented in the Project Completion crops for export, although Bank support Report (PCR) and drew in that document for coffee, which is Papua New Guinea's a satisfactory rating for t%e project (PNG) main agricultural export and had as a whole. But the earnings of the featured in the Second Project, was estates and farmers who borrowed under almost entirely withheld due to inter- the project, and the flow of repay- national quota restrictions. The Third ments to ABPNG9 continue to be de- Project was appraised in 1983. Pro- pressed by the unfavorable price cessing was suspended pending a reorga- trend. The audit mission followed the nization of the executing agency - what PCR mission by I8 months and could became the Agriculture Bank of Papua better assess the effects of the poor New Guinea (ABPNG) -and better recep- price regime. The problem for ABPNG tivity by Government for market-orient- has been worsened by Government's ed interest rates. A post-appraisal inability to allocate the budget sup- mission in early 1985 led eventually to port it had indicated to the Bank in Loan effectiveness in August, 1986, 1985 would be forthcoming to compen- without significant movement on inter- sate for an unwillingness to raise eat rates. The Loan (the previous two interest rates. The audit finds that projects had been Credit financed) was the incentives and resources of both for US$18.8 million, against total the bank and the farmers have been costs estimated at US$31.3. The pro- depleted to the point where the pro- ject covered the four-year lending ductivity of the assets created by the period 1985-8. loans, in particular the cocoa estates and cocoa and oil palm small holdings 2. Project lending coincided with the that were rehabilitated (and to a period of decline of international crop lesser extent expanded) with project commodity prices. For coffee, cocoa finance, may be compromised. The and palm oil, the three primary export satisfactory rating is no longer tena- commodities which are also ABPNG' a main ble (para 4.1). agricultural lending lines,i the -vi- Obiect2'ves (OFL) category". Approvals, sub-loan disbursement. and Bank Loan disburse- 3. The project had mixed objectivest metits all accelerated in 1987, and by to increase agric;ltural exports, the end of 1988 the onlendinS category reduce agricultural imports, raise of the Loan had been exhausted. Thir- rural incomes, strengthen ABPNG, and ty OIL subloans, all but two for cocoa "hel establish an appropriate policy estates, absorbed almost half of the framework for development of a finan- funds. Including small cocoa loans, cially sound rural credit market". this crop took 511 of project funds. ABPNG was established in 1985 at a time Oil palm smallholders took 141, trucks when the financial autonomy and compe- and other transport 17Z, and poultry tence of its predecessor bank - under at 3Z comprised the rest of the check the pressure of increasing arrears, list of important categories. Perfor- political influence on credit manage- mance against institutional targets ment, and the exodus of Australian was adequatet training, zonsultant staff - were beginning to erode. At assignmentep provision of vehicles for the same time, the previously expatri- field officers, etc. However the ate-owned cocoa estates on the outer attempt to improve the "policy frame- islands were being neglected, and, work" for ABPNG was unsuccessful. also, the trees planted during the Interest rates, though positive in the first stage of the successful small- sense of remaining above the modest holder oil palm settlements on New level of inflation, were much too low Britain had grown too tall too harvest. to help cover in any significant wty Both tree crops were candidates for the high costs of loan administration rehabilitation. In 1985, prices for (a consequence in part of the unusual- palm oil had already slipped, but not ly dispersed population) and rising much. Prospects for it and cocoa were delinquency. The Bank attributed both adequate to justify new loans. The these factors in part to politics, and project's production targets also APNG agreed, but neither were effec- included other new and old crops, pigs tive in persuading Govermment to alter and poultry, farm and marketing vehi- course. cles and earth moving equipment. ABPNG would continue to invest in coffee with Results its own resources and some loan funds from the Asian Development Bank. 5. The extent and quality of tree crop plantings were above (cocoa) or Implementation Experience close to (oil palm) appraisal targets. The PCR mission re-estimated economic 4. Project onlending got off to a rates of return for the two crops and slow start, despite ABPNG's experience found these within the range 8-181 and in previous projects. The principal acceptable. The 8% referred to small- demand was from the large cocoa es- holder oil palm, but that too would tates, where expatriate-led managing exceed a 121 opportunity cost of capi- agencies took the lead in preparing tal if the cost of upgrading the set- farm plans on behalf of the owners tlers houses were excluded. The audit (usually groups of local PNG citizens). has adjusted those calculations to ABPNG committed the project funds reflect changes, since the PCR was rather quickly; the lag was in arrang- prepared, in its price estimates and, ing for subsequent Bank approval of all more importantly, evidence that the large loans in the "over free limit PCR yield estimates also can no longer - vii - be maintained (paras 3.5, 3.13). income have caused it to steadily Estate and smallholder maintenance reduce its volume of annual lending. standards appear to be slipping under That is especially true for agrizul- the disincentives of the low prices. ture# but it reflects a trend across The same is true for coffee. The all its coercial and industrial situation may get worse, since owners lendiag lines as well (only 49Z of of the cocoa and coffee plantations ABPNG1s outstanding portfolio at the have lost confidence in the managing end of 1990 was in agriculture). In agencies - blaming them partly for the response to the coodity price de- reduced not earnings - -ad are seeking cines in 1989 the bank suspended to remove them. It vou.aLd appear that interest income from its plantation all the PCR models would now fall below loans. Neverthelesep all 52 of 10%, unless heroic assumptions are made ABPNG's largo cocoa loans (including about a reversal in commodity price the 28 under the project) fell into trends or Government's exchange rate arrears. Most of these accounts have policy. That refers to the economic had to be rescheduled, and a steadily rate of return. because of the ex- growing number have been shifted to change rate policy, which is overvalued the doubtful and bad debt ledgers. when measured for the agricultural ABPNG's collection rate fell to only economy, financial rates of return for 42Z in 1990. Thet veveals the deepen- cocoa and oil palm farmers are closer ing crisis. All the large cocoa loans to zero (paras 3.7, 3.8, 3.13). That to Madang province may have to be is where the impact of the tree crop written off. And almost all of the loans is most dramatic: on small farm- smallholder accounts in the project's era who own a settlement block or a oil palm zone are now in arrears. piece of a plantation and have not Exacerbating the situation, the polit- received any significant income from ical turmoil on Bouganwille has al- them for the last five years nor see ready forced ABPNG to write-off all much better prospects for the next ten. loans to that island. Bouganville was responsible for 50% of PNG's cocoa exports. The reported improvements in 6. The rest of the lending lines show ABPNG's institutional capacity attrib- a mixed performance. The popular two- utable to the project, measured ton trucks, convertible for haulage or against what would have been -he case busing, were financed by ABPNG under in the absence of Bank support, has to hire-purchase arrangements. Despite be viewed against this backdrop of wearing dourn of the fleet on the rough increasing financial stress, due to roads repayments have been relatively factors largely beyond ABPNG1c con- good. The poultry loans were good and trol. bad: repaymenti have held up although net earnings after debt repayment by Sustainabilitv these new commercial farmers have also been low. The small-scale irrigated 8. As suggested above, the sustain- vegetable scumes near Port Moresby ability of the rehabilitation work on have mostly been abandoned and the the cocoa and oil palm farms is uncer- pumps repossessed. tain. For cocoa, there is also con- cern that the hybrid variety that has 7. With respect to the institutional been the basis for the recovery may viability of ABPNG, the mounting losses have been inappropriatep and offers of expected loan reflow. and interest lower long term yields than projected. -viii- That is a seperate issue than the Fldings and Lesues decline in mairtenance standardc. For the four hectare oil palm blocks, half 10. This is a project that made satis- of that area is in the "third phase" factory progress during the early on- section planted furthest from the road lending period toward physical and and there is evidence that many set- institutional targets, but has been tiers are not fully exploiting thew. caught by unfavorable international These settlers are the pioneers who commrdity price trends and macroeco- came in the late-1960s. They are old nomic policies and faces losses now ii, now, and the majority nuither get nor the quality of both the new farm as- expect much support from their sons. sets and the loan portfolio. The root A turnaround in prices paid by the palm causes of the stress were beyond pro- oil mill would generate new enthusiasm, ject control. Since 1985 prices for although sccio-economic research in the its major products bave tumbled. settlements suggests that even then the Also, PNG'a foreign trade sector is commitment of the sons to continue this dominated by mineral evports - not heritage is suspect (para 3.11). agricultural commodities - and that Yields are reported to be declining positioning is going to be strength- throughout the smallholder oil palm ened as the mineral boom accelerates. (and cocoa and coffee) sector. Exchange rate policy is focused on -ontrolling that development, not on 9. With respect to the sustainability rescuing the tree crop sector from of ABPNG, the PCR indicates that one of adverse price movements. Neverthe- the project's major institutional less, the comparative advantage that targets - holding constant the level of these tree crops provide PNG cannot be public subsidy supporting ABPNG's wasted. The project was a wholly portfolio - was missed completely. inadequate instrument to protect those ABPNG depends much more on outside help interests. In this larger context, an now than at appraisal. The Bank otherwise good project becomes unsat- dropped a fourth loan from its pipe- isfactory. The Bank is now reformu- line. This was partly offset by a lating its approach to these crops in Japanese (OECF) credit for US$18 mil- the hopes of protecting the progress lion in 1988 which largely copied the in rehabilitating and expanding the Third Project design, though without export base that has been achieved in insisting on raising interest rates. the past decades (pares 4.39 4.4). The Asian Development Bank is providing Tat is the correct response for the funds for further rehabilitation of the Bank. cocoa farms on New Britain. Neither of these external borrowings, however, 11. The Bank's preoccupation with the compensate for the deteriorating quali- interest rate issue - during the run ty of the overall agricultural portfo- up to the Third Project, during a mid- lio, including those loans still in an term socio-economic evaluation of active status (i.e not yet reclassified borrower attitudes and, most conspicu- as doubtful or bad). The audit be- ouslyl in the PCR - seems irrelevant lieves the main explanation for that to the storm that was gathering. Under trend to poorer quality is the commodi- no circumstance could rates of inter- ty price decline, and the exchange rate est applied to smallholder settlers policy which let the agricultural and tribal group-owners of transferred sector absorb the full impact of that estate-9 be expected to cover either decline.* the cost of administering and collect- -ix - ing debt on this emerging island econo- 12. A close reading of the project my, with its scattered and often iso- history shows also the importance of lated branch banks, or the provisions technical agricultural factors behind for bad debt as they mounted under the the success or failure of a farm cred- impact of the commodity price decline. it project. These have been obscured In retrospect, it is clear that higher by the pricing problem. Examples interest rates would simply have led to referred to in the main text (para high:,r arrears. The socio-economic 4.12) are: (a) the rapid growth of the evaluation demonstrated that the bor- oil palm trees, bringing forward all rowers would have borrowed even at deadlines for replanting credits; (b) higher rates. The PCR highlights that the importance of the pollinating finding, and goes on to attribute many weevil - brought from Africa to New of the project weaknesses to the low Britain in 1982 - to assuring the rates that the Bank reluctantly had to survival of the smallholder oil palm accept. It goes too far. In fact that schemes; (c) the unusually heavy rains discussion is the weak spot of an of 1990 and 1991 that encouraged the otherwise informative PCR. The macro- spread of black pod disease in the economic backdrop to the project is not cocoa fields and reduced yields by illuminated well In that document. The half, on the heals of the price de- audit had the perspective of another 18 cine and dealing a mortal financial months of stress on the borrowers and blow to some estates; and (d) the the bank. It shows that the interest possibility that the popular cocoa rate was in this case not as important hybrid that spearheaded the recovery as other pricing issues and did not of the cocoa sector may have a genetic deserve the prominence it was assigned flaw. (pars 4.6 - 4.8). PROJECT PERFORMANCE AUDIT REPORT PAPUA NEW GUINEA THIRD AGRICULTURAL CREDIT PROJECT (Loan 2624-PNG) I. BACKGROUND A. Preparation 1.1 Preparation of this project was started by the Papua New Guinea Development Bank (PNGDB) in 1982, before its predecessor project, the Second Agricultural Credit Project, became effective. The intention was to ensure coverage of the next three-year time slice of PNGDB's rapidly expanding agricultural lending operation, beginning in 1984. The emphasis was again to be on tree crops, although the Bank indicated that its world-wide restrictions on lending for coffee and cocoa might apply. Ultimately coffee was excluded, except for a marginal amount for rehabilitation. These were term loans - ABPNG does not provide seasonal or other short term finance except as part of a line of permanent working capital. 1.2 The proposal was appraised in 1983, but processing was suspended later that year when Government declared its intention to substantially restructure the institutional apparatus for agricultural loans. An initial reorganization plan drew criticism from the Bank. In the same period the Bank's concern also increased over PNGDB's operational viability, prompting a post appraisal mission in late 1983 and formulation of an "Action Program". Continued disbursement under the Second Project was tied to that program. At the top of the agenda for the Bank was interest rate reform. In February 1985 the reorganization was carried out: PNGDB was replaced by the Agriculture Bank of Papua New Guinea (ABPNG), with the old staff and branch network largely intact but with an obligation to increase over time the proportion of the value of agricultural loans in its portfolio to 80% to reflect its new mandate. A second post appraisal mission was carried out in the same month, and the project moved quickly to Board approval in September 1985. 1.3 The subject of interest rates had dominated discussion with Government before and during the period when processing was suspended. The Bank argued in vain to get Government to authorize an increase of on-lending rates from the prevailing 8.5% to at least 11%. The Bank had a few supporters in the Cabinet but the majority felt that the Government was obliged by its campaign promises to provide special assistance to the farmers to encourage them to shift to commercial agriculture and that in any case an agreement on higher rates should be deferred until there was evidence that the crop surpluses generated by the loans were sufficient to improve family incomes after servicing the debt. Government did agree to review the general rate structure annually, to restrict to farmers below a specified poverty line the special subsidies modifying the general rate structure, to hold overall subsidies to ABPNG below a certain ceiling, to investigate alternative methods to subsidiz the farmers, and to use -2- project funds to finance a study of farmer sensitivity to the level of interest. Government also accepted a slight u,;Yard adjustment of the rates. The Agriculture and Rural Development Department in the Bank continued to argue that a tougher posture on eliminating subsidies was appropriate. Nevertheless, the Bank accepted these weaker provisions. It was clear that Government would not move furthert and the alternative was to give up what had until then been a clearly successful line of credits. It should be noted that the agreed rates - virtually unchanged since the 1982 proposal - were positive via a vie the relatively modest level of inflation and were expected to remain that way. What the Bank wanted was that ABPNG should be released altogether from dependence on an annual budget provision. This implied that farmer rates should be at least as high as to cover operational costs and perhaps a part of the bad debt, as well as the concessionary cost of the capital provided by the project. 1.4 The Bank was also concerned about the deterioration in its own projections of international prices for tree crop exports. Coffee viability was especially suspect, given Bank analysis of the growth in world export capacity. Government argued that PNG had few alternative enterprises where a comparative advantage had been so clearly demonstrated: coffee was the country's leading agricultural export, it occupied the greatest number of emerging commercial farmers, and it was the most important crop in ABPNG's agriculture portfolio and in disbursements under the Second Project. ABPNG would continue to lend for coffee with or without Bank involvement. A new loan from the Asian Development Bank (ADB) would help ABPNG to cover a part of that investment (the Third Papua New Guinea Development Bank Project; signed January 1983 and effective, also after a delay, in July 1985). The Bank approved small allocations for rehabilitation and processing, but otherwise abandoned the coffee program. Cocoa remained eligible - the Bank felt the economic profile was satisfactory. Twenty percent of project costs in the Staff Appraisal Report (SAR) were allocated to cocoa. As shown below, cocoa eventually took 51% of the on-lending funds. B. Objectives 1.5 The project had a mixture of three objectives, none of which clearly dominated. The production objective is presented first in the SAR, in particular rehabilitating and developing the tree crop sector. This was a carry-over from the Second Project, which had initially been conceived as a tree crop rather than a credit project. Tree crops would be the leading edge of a progriam to expand exports through the increased production of established and new exportable commodities. The old cocoa estates were a special target: already in the 1970. they were approaching senility while replacement rates were minimal. Poultry, piggeries and vegetable crops were to be encouraged too, to reduce the import bill and save foreign exchange. The second objective was to raise the incomes of PNG farm families. There was an implicit assumption that this would follow progress towards the production targets, though that assumption was later to be challenged as the results of a socio-economic survey were revealed (para 3.7). The third objective was the strengthening of ABPNG as the premier public lending agency for the rural sector. For this there would be a group of project activities targeted on improvements in ABPNG's operational capacity and procedures. Also, the project was designed to help establish an appropriate policy framework for development of a financially sound rural credit market in the country. The weaker formulation of the interest rate conditionality was part of that program. C. The Project 1.6 Project costs were estimated at $31.3 million, of which $23.7 million (76%) were allocated to on-lending and the rest to incremental operational expenses ($5.2 million) and vehicles, training, technical assistance (TA) and special studies ($2.4 million), all for ABPNG. That bank had already reorganized its field structure, creating four regional offices each with branches and "representative" offices. Under the Action Program, ABPNG had agreed to a reduction in the number of these subsidiary offices, and a greater than proportional reduction in total staff. A Deputy Managing Director would be contracted, and several crop, training, management information (MIS), and other TA consultants would also be recruited internationally. The loanable funds would finance a three-year slice (1985-88) of ABPNG's planned agricultural lending. Indicative subsector lending targets in the SAR allocated 43Z to agricultural transport and equipment (mostly trucks and earth-moving units), 33Z to cocoa and oil palm, 11% to poultry and piggery, and 141 to new crops, miscellaneous agriculture and agro-processing (a collection of categories which included coffee, rubber and lumber). The prominence of transport and equipment had not benn a feature of the Second Project. 1.7 For the cocoa program, the SAR distinguished three categories of loans. First was the replanting of existing estates, which were described as "100 ha" plantations for cost estimating purposes but in reality were usually larger and sometimes much larger than that model. These estates were mostly on the outer islands, with a concentration in East New Britain. Estates still owned by expatriates were not eligible for ABPNG loans. Of the estates that had been transferred to nationals during the previous two decades, a few were owned outright by PNG individuals. The great majority, however, were owned by clans or other groups of nationals, usually residents of the area. Invariably even in the groups a few members would dominate, including businessmen, politicians etc. These decision makers, and the individual owners, were later to be described as the "notables" in the socio-economic survey report, though in terms of numbers they were insignificant on the long list of nationals identified as owners. The primary public bank, the Papua New Guinea Banking Corporation (PNGBC), and a few of the foreign banks operating in PNG, were also making loans for rehabilitation, expansion and even establishment of cocoa estates during the 1980s, including the expatriate-owned estates. Government had several other incentive programs to encourage the recovery of the cocoa industry, although the ABPNG activities were to offer the major source of funding during the project period. 1.8 The second cocoa category was described as 20 ha plantations, almost always new farms to be carved out of natural forest. That area was a minimum, and was related to a critical mass of harvested bean thought necessary to justify the use of a management agency. ABPNG would insist that all groups owning these small plantations as well as the larger estates must employ management agencies as a condition of the loan. The small plantations were also almost always group- -4- owned, though individuals were eligible. The 20 ha model had become popular during the Second Project, and was especially so in the coffee zone on the mainland. The Third Project would accelerate adoption of the concept by the cocoa industry. The third cocoa category in project design was the smallholder blocks in the Warangani settlement scheme of East New Britain. These were presented as a 6 ha model though there was a range around that figure. They were individually owned and managed. The SAR projected that 16 estates, 104 of the 20 ha plantations (ABPNG calls both categories "plantations" and lists them together), and 150 settlement blocks would benefit from the project. 1.9 For oil palm, the project aimed specifically at 1,200 settlers at the Hoskins scheme in West New Britain. These were the families on the scheme that pioneered oil palm smallholder activity in PNG in the mid 1960s. The four hectares of trees of the first and second plantings had grown too tall, and the principal objective of this project component was to poison, cut down and replace them. II. IMPLEMENTATION EXPERIENCE A. Project Costs and Dollar Disbursements 2.1 Actual project costs were $30.5 million, 97% of SAR estimates. The on- lending component took $26.9 million, or 88% of the total. That was higher than the 762 share allocated at appraisal and reflects mostly the transfer of project funds from "incremental operating expenses" - which ABPNG elected to cover from its own budget - to onlending. There was a savings also in the training component, where an expatriate consultant and overseas courses were dropped in favor of a purely local program. Distribution of the on-lending component among the various enterprises is shown in table 1. A subgroup of 30 loans labelled Over Free Limit (OFL) - requiring prior Bank approval - is separately identified in the "actual" columns of the table. It focuses attention on the importance of borrowings by the cocoa estates (all OFL) in the final accounts. ABPNG approved 29 such estates and disbursed against 28. This refers to project funds. Altogether ABPNG has 52 conos, estate accounts, some financed before the project and some during the projact using other sources (for example, the ADB Third Project). ABPNG did not restrict the project estate rehabilitation finance to 100 ha, and the Bank during supervision did not argue that it should. 2.2 The table shows that 512 of the value of the sub-loan total went to cocoa: 42% for the estates and 9% for the 20 ha plantations and smallholder blocks. The SAR had allocated 21% to all three types. Agricultural transport - mostly trucks - was the second largest category at 172 (SAR 19%), followed by oil palm at 14% (SAR 11Z). The next largest categories in the actuals were well behind at 42 and 3%. During 1987 supervision missions reported that the cocoa estate component might 0bsorb as much as 692 of the value of the subloans, based on commitments already made by ABPNG. The subsequent weakening of the cocoa -5- price and retrenchment of the loan requests cut back and extended disbursements against these applications beyond the project period. Table l: Allocation of Amounts Disbursed for Subloans (US$ 9000)- Free Limit Over Free Total imit _ _I Cocoa 2,332 10,202 12,534 51 Oil Palm 3,474 3,474 14 Poultry & Piggery 517 333 850 3 (of whichs Poultry) (502) (333) (035) (3) New Crope & Kisc.Agr. 1,612 666 2,278 10 (of which: Rubber) (194) (666) (860) (4) ( Gardening) (189) (189) (1) ( Fishing) (373) (373) (2) Agro-Processing 477 477 2 (of which: Cocoa) (344) (344) (1) Agro-Transport 4,183 4,183 17 Agro-Equipment - 619 3 Total 13,213 11,201 24,414-1 100 i All Ma figures converted at US$1 - K .9. ki Total conforms to ABPNG table dated November 1, 1989 for Free Limit disbursements. The OFL figures are separately derived from "Approved* Tables. The grand total also conforms to PCR Tables 4A and 4B. It does not conform to PCR Table 5, which shove US$ 26.9 million. 2.3 The Bank Loan of $18.8 million was fully disbursed. Of that, 91% was disbursed against the sub-loans, slightly higher than the SAR estimate. Most of the difference was transferred over from the unrealized foreign costs of the training com,?onent. B. On-Lending Component 2.4 Approximately 2,401 sub-lou i were disbursed under the free limit, and another 30 over it. The largest nuuber of loans were made to the oil palm settlers (954), followed by transport (537), fishing equipment (271), and cocoa processing (167). The latter is not included in the "cocoa" subloans discussed elsewhere. Table 2 aumarizes these statistics, showing also the average size of the subloans for both free limit and OFL. 2.5 This is a project which can be described in four parts: (1) the cocoa loans, especially to the estates which took 42% of all the funde, (2) the Hoskins oil palm small holder sub-project, (3) the shorter term (three year repayment) truck and bulldozer loans, and (4) the large variety of all other loans, which accounted for about 302 of the number and 152 of the value of the total. The average value of the estate loans (K328,000 - approximately $364,000) dwarfs the other categories. The audit mission was primarily concerned with the cocoa and -6- oil palm loans, which together absorbed almost two-thirds of all onlending funds. It is important to note that had the mission been concerned with the other two parts it would have been observing 1,300 accounts in different settings, with other emerging small and medium size commercial farm, farm/transport, and farm/processing enterprises, and may have made other findings. Table 2: Number and Averaie Value of Subloans (Diabursements, K9000) Average Share of Total Number Value Disbursements FL OFL Total FL OFL FL OL Cocoa 116 28 144 18.1 327.9 9 42 Oil Palm 987 - 987 3.2 - 14 - Poultry & Piggery 70 1/ 71 6.6 300.0 2 1 New Crops & Misc. 458 Jk1 459 3.2 599.0 8 2 Agro-Processing 195 - 195 2.2 - 2 - Agro-Transport 537 - 537 7.0 - 17 - Agro-Equipment 38 - 38 19.9 - 3 - Total 2,401 30 2,431 55 45 *1 For the then-ABPNG-owned Nuigini Table Birds Company. k1 For the ABPNG owned Sogeri Rubber Development Corporation. 2.6 Repayments of most of the project loans have only started in the last two years (e.g. cocoa rehabilitation had a four-year grace period), and it is too early to report meaningful trends for that subset of ABPNG loans. The three-year loans for agricultural transport and equipment are exceptions. These were concentrated in the first two years of project lending, ABPNG retained ownership of the assets until repayments were complete, there was no grace period, and about 60Z of the loans to both groups had been fully recovered even before the project was completed at the end of 1989. The quality of the rest of the project portfolio, in terms of the expectancy of repayment, must be judged by the bank's collection experience with earlier lending and by its treatment of arrears in the early phase of project debt recovery. In both respects, the prospects are poor. 2.7 The ABPNG MIS system does not provide a collection ratio showing collections as a percentage of collectables (now and old dues). Rather it shows collections as a percentage of the total portfolio and of the portfolio that has entered the repayment period. However, ADB acquired the data required to calculate the first collection ratio for the Project Performance Audit Report issued in 1991 on its Third Project. It gives 462 and 422 for 1989 and 1990 respectively. That is for ABPNG's overall lending, including non-agriculture which comprised 512 of the ABPNG portfolio in 1990. The collection ratios for the agriculture sector are not separately identified in ADB's PPAR - the ADB project provided funds for all sectors. But there is a strong presumption that the agriculture performance is worse than the total. These figures hide other - 7 - indicators of deteriorating collection performance, especially as they reflect conditions under the Bank's project. The great majority of the estate loans - project and others - have had to be rescheduled (the 28 OFL cocoa accounts, which Bank supervision had examined individually before approval, are in no better shape than the rest). Also, under the impact of the declining coffee and cocoa prices, a decision was taken in early 1989 to treat any and all accounts in ABPNG's Category 10,, which comprises coffee and cocoa plantations over 20 ha (i.e. both the estates and the 20 ha "plantations"), as doubtful (Category 16/). Interest payments were suspended, interest accruals in the accounts were stopped, and ABPNG set aside an enormous sum of K43 million in that year's profit and lose statement as provision for bad debt. In the early 1980s the Bank had argued that the annual provision of about K1 million was inadequate to reflect the potential for bad debt. In 1982 the ABPNG set aside K5 million to catch up - a sum that satisfied supervision under the Second Project. The 1989 provision was in response to a threat of unprecedented dimension. 2.8 In 1990 part of this threat was realized as, with civil turmoil in Bouganville, ABPNG closed its offices there and shifted to the "bad debt" category (19/) all loans, large and small, on the island. This included four of the project estate loans and the majority of the project 20 ha plantation loans. ABPNG staff told the audit mission that the two estate and most of the 20 ha cocoa loans in Madang Province - the only important cocoa producing province on the mainland - may also have to be written off. The deepening delinquency problem is not confined to cocoa. Almost all of the settler oil palm loans were rescheduled in 1989 - this included remaining debt from an earlier program that planted the "third phase" (i.e. the third of the 2 ha subdivisions comprising the 6 ha block), plus the initial payments for the project replantings). The audit mission reviewed a sample of the smallholder accounts in ABPNG's branch in Kimbe (the actual site of the so-called Hoskins scheme), and found that in 90% of the cases the borrowers were falling behind each month in repayments even of the rescheduled debt. This is not a willful act of theirs (though it otherwise might have been). The repayments are deducted automatically by the Small Holder Affairs Division at the Mosa palm oil factory at Kimbe. Partly because the company!' and ABPNG policy is to allow usually at least a small payment to the farmer, the price deduction formulas do not permit Mosa to repay to ABPNG all that is due. Given the bleak projections of future palm oil prices, ABPNG is not hopeful that the payment trend can be reversed. The audit mission did not visit the coffee zones, but these were investigated by the socio-economic survey researchers in late 1988 and again by the ADB audit mission in 1990 and the delinquency situation there seems also to be unmanageable. C. Institutional Component 2.9 The problem of rising arrears and the worsening financial situation of ABPNG is discussed again in later sections. The PCR attributes the "grim" status It New Britain Palm Oil Development Ltd. (NBIPOD), which operates a nucleus estate and the mill and accepts (depends upont) smallholder fruit from both the settlers and villagers. NBPOD is 502 owned by the Harrisons and Croafield Group. This audit does not discuss the progress of the company or the efficiency of its mills. - 8 - to several factors - the decline in commodity prices of course, along with rising operational costs during the project period, the low level of interest rates, and the bank's continuing inability to access other resources through deposit collection. The cost phenomenon is a reflection of what amounts to a reversal of the SAR forecasts. To achieve better control over the portfolio, and the arrears situation, and to compensate for a perceived deterioration in the quality of pukli* extension services (after provincialization) and of the management agencies, ABPNG staff was increased rather than decreased from the appraisal level. The relevant figures are, for total staff, 334 actual in 1984, the SAR target of 317, and 424 as a December 1990 actual. ADB reports that ABPNG's personnel costs rose from K2.7 million in 1985 to K5.2 million in 1990. The size of the expatriate cohort also increased at first - as part of the drive for improved quality. But this trend was short-lived and the complement of fully- paid professionals is down from the 1984 total of 21 to 9 in 1991 (plus a few Peace Corps Volunteers). ABPNG cannot pay the high salaries of the former Australian contingent, who have almost entirely been replaced by contract staff from other continents. The SAR attempt to force a reduction in the number of branches failed, as Government reacted to the announced closings of 12 offices in 1986 with an instruction to reopen them. A few of the least active offices have subsequently been closed, but one must anticipate that political factors will continue to intervene with any effort to withdraw the services of the country's only official agricultural bank from any of the even isolated regions. There are now 10 branches and 17 representative offices in operation (none on Bouganville). 2.10 Project implementation was attended by the types of delays that are common to the Bank's rural and credit projects. The supervision files and the PCR mention the delays in audit, the poor quality of the earliest Statements of Expenditure, the inadequacy of the project appraisals submitted to the Bank for approval of the OL applications, delays in Government's conversion to Kina of Bank reimbursements for the ABPNG account, and, perhaps most important, the failure of Government to honor its commitment to provide counterpart resources matching the Bank's share. This audit report does not expand on these matters, some of which are addressed in the PCR. 2.11 One of the prominent issues of implementation in both the Second and Third Projects was the variable performance of the management agencies. The variability factor, prompting the PCR and ADB to remark on the poor quality of many of the agents supervising individual farms, is more apparent in the coffee zone, where the number of agencies proliferated to handle the rapidly expanding list of 20 ha plantations. In response, ABPNG decided to set up in 1979 its own public company, the National Plantation Management Agency (NPMA) to assume these responsibilities in areas and for farms where the private sector was doing poorly. For various reasons that company did poorly too, and was liquidated in 1987. Some of its staff and farms were taken over by a new private agency. At the time of the audit there were three private management agencies operating on the cocoa plantations in New Britain, and a fourth in Madang. 2.12 Despite complaints from group and individual owners that the agencies were retaining too much of the net profit - after debt repayment - for -9- themselves, leaving little or nothing for the owners, the problem seems rather to be more a function of the depressud market. The agency fees are set by ABPNG. The PCR argues that the fees ought to be converted to a system that rewards performance measured by overall earnings, and not a flat fee proportional to the size of the cocoa area, as it is now. But in the present circumstance the agencies could hardly agree to that adjustment. The nationals talk also about Australian managers taking home high salaries, but by now almost all of the agencies' plantation managers, and most of the area supervisors overseeing groups of plantations, are nationals too. There is little doubt that agency performance in cocoa is variable, but whether this reflects the quality of the agents more than the physical attributes of the individual farms is uncertain. The audit's conclusion is that the ethical posture of the managing agencies is appropriate for the job at hand. What is showing up now are differences in the quality of work as these companies react under stress. 2.13 To dispense with the agencies, especially on the group cocoa plantations, could court disaster. That is exactly the direction most of the groups intend to take them. The PCR anticipates this shift as inevitable at the termination of the repayment period, when the ABPNG condition is removed. But already a few of the individual and group owners within the program have dropped, or petitioned ABPNG to allow them to drop, their agencies. ABPNG is expecting the movement to spread. The branch in Rabaul, which handles the East New Britain cocoa accounts, has stepped in to support the first of those freed-up group farms with its own staff, with good results thus far. But the branch is over-burdened by that work and cannot absorb many more. The prospects are improved by the fact that the groups tend to retain their agency's farm manager, at a renegotiated wage, while getting rid of the agency contract and the overheads and image that goes with it. But that trend may itself be reversed, as the individual members of the group opt to break up the property into family blocks, shift to informal management, and exploit the harvests while the picking is good without putting resources back into preserving the trees. 2.14 Implementation of the TA components went reasonably well. It was mentioned above that the training program was not internationalized. But ABPNG designed a local training operation that has been assessed as quite good. For transport for ABPNG, 51 care (SAR-46) were bought through competitive bidding, spurred on by Bank agreement to dispense with conventional international bidding procedures. The consultant program was carried out, with the exception of the training advisor, and the agronomy crop manuals (Practical Guidelines) and manpower survey are assessed as acceptable. The HIS advisor appears to have performed well, although the PCR says much more must be done to provide ABPNG with an adequate monitoring system. The socio-economic study report ran into problems that are discussed below (para 3.15). 2.15 The impression given by these TA activities, fitting in as they did with general improvements in ABPNG's procedures, is that the project was generally successful in achieving the near-term targets implied for the 1989 project completion date. A supervision mission in 1987 had reviewed the cocoa applications against the then, already lower, cocoa prices and found the returns below SAR expectations but nonetheless acceptable. The Bank's supervision - 10 - ranking of the project remained at either 1 - no significant problems - or 2 - moderate problems that could and probably would be overcome - throughout the four year implementation period and its four supervision missions. The PCR mission followed in May 1990, and, despite the commodity price collapse, still found that the project could be considered satisfactory. III. PROJECT OUTCOME A. ABPNG's Financial Position 3.1 The losses in income due to the 1989 suspension of interest payments and the deteriorating repayment profile have severely reduced ABPNG's resources to maintain, let alone expand, its agriculture portfolio. The rate of annual approvals for agriculture grew from an average of K1 million in the period 1980- 1984 to K19 million in 1986 - the peak. It fell to K8 million in 1990. Since the bank does better from its industrial loans, it prefers now to shift the overall portfolio in that direction. The 1985 agreement that established ABPNG called for a move over time' to an 80% position for agriculture. That was later revised to 75%. The 1990 Annual Report shows the present share at 49%. ABPNG has petitioned the Government to reduce the formal target to 70%. It does not accept criticism that it is abandoning its mandate in agriculture. Rather it claims that it needs to strengthen the other sector portfolio, to provide the income to cross-subsidize the agriculture portfolio and keep it and the bank itself afloat. 3.2 The PCR and the ADB PPAR on the two Third Projects provide substantial detail on the financial accounts and ratios of ABPNG. What ADB calls "substantial" annual operating losses have been the pattern since 1988. In the late 1970s and early 1980s the report card on PNGDB had been much better, and even during most of the rest of the 1980s the ADB tables show that ABPNG's financial accounts and operating performance have been reasonable, especially considering the unfavorable impact of the exodus of expatriate managerial and field staff and the growth of the "risky" agricultural portfolio. In 1985 the SAR reported a 1984 collection ratio of 782, which can be compared with 422 in 1990 (para 2.7). Some observers report their impression of "slackness" in ABPNG collection efforts, but it is hard for the audit to evaluate this factor given the low farm incomes and obvious difficulty in repayment. The audit did perceive elements of less than satisfactory performance in the agency's operations - the information management area is also unimpressive - but these may be signs of a young institution learning to get along without heavy expatriate operational support. - 11 - B. Agricultural Impact The Commodity Price Decline 3.3 The Bank's price series for cocoa, oil palm, coffee and coprall are shown in the next two pages. They are included here rather than as an annex to ensure that the reader gras.s the magnitude of this unfavorable trend. Other exporting countries have weathered the storm because of different ("farmer friendly") exchange rate and wage policies. PNG feels it can afford neither (para 4.2). Of special relevance is the interval from 1985, the appraisal year, to 1990. Collectively, the weighted real value per ton exported for the four crops has fallen by 60% in that interval. Cocoa has shown signs of a partial recovery in 1991, but it is clear nevertheless in all four cases that a shift of kind rather than degree has taken place. The Bank had been concerned during appraisal that export prices might not hold up (para 1.4). But it had not anticipated a shift of this severity. The Bank does not foresee substantial recovery in the next decade. Cocoa Farms 3.4 The PCR reports that 4,826 ha were replanted under the project, 5% over the SAR target of 4,580. Most of that was on the OFL estates. The targets for number of loans for the 20 ha plantations and the settler blocks were both substantially undershot. Planting standards are described as reasonable in the PCR, especially in the first years before the impact of the price decline had been appreciated, though the PCR does note some deficiency and predicts a slight reduction in overall yields below SAR expectations. The impact of prices on management standards and yields is concentrated rather on the maintenance period. Low prices have meant cutbacks in otherwise routine services$ those where full treatment is not considered essential to the harvest of a basic crop: weeding, pruning, fertilization, disease control, etc. 3.5 The SAR used yield estimates of 1.60, 1.75, and 1.25 tons of dry bean per hectare as the averages for the estates, 20 ha plantations, and settler blocks, respectively. The PCR revised two of those figures downward somewhat: 1.60, 1.50, and 1.10. However yield estimates provided by a consultant for the Bank's present PNG tree-crop review (para 4.3) gives figures much below either of those earlier sets of estimates. For example, the consultant reports, for the estates, 1.3 tons/ha for the best 8 of 32 sampled estates in 1989. That was the last good year, before the cocoa zone was devastated by the extremely heavy rains and cloudy weather of the last two years. Only 2 of the 32 reached 1.1 tons/ha in 1990. The excessive rain and sunless days have encouraged the spread of black pod disease everywhere, contributing to as much as a 50% reduction in the harvests. The consultant claims that the better estates would have managed to survive the price decline - in the sense of being ablA to pay off most of the II Copra and copra oil are derived from coconut and were once the preeminent export crops of PNG. Tog,ather they ranked fourth in 1990 in export earnings. However they occupy a minor position in ABPNG's portfolio. - 12 - dues - were it not for these rains. Partial recovery from their effects can be e-pected. But given the simultaneous reduction in field maintenance expenses and their inevitable impact on future harvests, yield forecasts in the PCR are no longer tenable. 3.6 The consultant suggests a figure of 1.1 tons/ha as appropriate for average estate yields in the future. At that level the economic rate of return (ERR) re-estimated in the PCR for the "100 ha" cocoa model would fall from 151 reported there to 9%. Should prices unexpectedly recover sooner and further than the Bank's latest projections - strengthening the estate earnings and permitting a resumption of normal practices - the figure would rise. However, a countervailing factor is the growing restlessness of the group and individual estate owners, who may throw out the agencies before such a recovery occure. That outcome is now a real possibility. It is the most worrisome vision of the whole audit exam. 3.7 The suspicion that this could occur is based not so much on the ERR figures but in the recalculations of financial rates of return (FRR). The PCR re-estimates these as 7% for the 100 ha model and 5Z for the 20 ha model. In retrospect, both figures must be considered over-optimistic. The Bank's revised price projections for cocoa are slightly better for 'the next five years, and slightly worse for the following five years, than the forecasts that were used by the PCR mission. So the price factor does not suggest a significant change in the PCR's FRR projection. The damage is done by the downgrading of the yield estimates. The audit mission met cocoa estate clan owners who have never received any dividends from their property, and do not see any sign of doing so in the fucure. That sorry state is said to be common to all. That concern is shared by the banks. ABPNG, PNGBC and the private commercial banks financing cocoa estates have discontinued rolling over working capital loans to support continued maintenance of these estates. 3.8 It should be pointed out that the new ERR and FRR estimates refer to models of functioning farms. Analysis of all pr-4ect loans would include the farms on Bouganville and elsewhere that have gone back to bush or are receiving minimal service. From that wider perspective the project rates of return would be driven closei to zero. Oil Palm Blocks 3.9 The Hoskins story has similar features. The replanting program vent well, most of it completed before the end of the second project year (1987). It was executed not by ABPNG (which has a hands-on role in executing the cocoa projects), but by the Department of Agriculture and Livestock (DAL). This is a remnant of the erstwhile Department of Primary Industry, which at the time of appraisal was responsible for extension services for all projects in the country but subsequently was broken into pieces as these extension services were provincialized. DAL, however, was made responsible for the oil palm settlements, and these remained under the residual national extension authority. (Notes cocoa extension was provincialized). DAL has maintained a strong presence in Kimbe throughout. It supervises the planting and replanting stages, sometimes - 13 - insisting that the settler contributes his labor, otherwise using contractors. The project ultimately did not finance replanting of both the first and second phase two-hectare subdivisions as previewed by the SAR. InstAad, only the first was replanted. The second was allowed to remain standing, beyond the harvester's reach, ready for poisoning and removal at the settlers discretion.!' 3.10 Here again the SAR projections are breaking down under the impact of the price decline. This audit report has not yet mentioaed the role of the price stabilization funds which were established for export crops in the early 1980s to transfer farmer earnings from good to bad years. The record of monthly payments by the Mosa mill to each farmer thus reflects the not available after deducting the mill's charges, and whatever repayments will be made to ABPNG, plus or minus a "bounty" or "levy". Surpluses in the better years of the earlier 1980s were exhausted for most schemes and most crops in 1989. Government since has injected public money into the stabilization funds to preserve some level of subsidy. These appropriations are considered public loans to the industry. They may never be repaid. Nevertheless, despite the bounty, settler take-home payments have declined substantially since the mid 1980s, and are far below the payments made when the Hoskins scheme reached maturity in the mid-1970s. This has impacted, as with the cocoa farms, on the oil palm settler's commitment to manage his mini-plantation. The Small Holder Affairs Division at the Mosa mill told the audit mission, and this was repeated by ABPNG staff, that maintenance and harvesting standards are both falling. The biggest threat is to the third phase planting, at the back of the block. Although the distances from there to roadside are not long, the majority of settlers - or so it is alleged - are selectively cutting down only the largest bunches for haulage in wheelbarrows to the road, leaving the majority of the bunches to fall and rot. Maintenance standards are down too, fertilization for example. 3.11 The impact of settler indifference has been reported on earlier occasions. The Hoskins smallholder oil palm program (not the NBPOD nucleus estate, with which this audit does not deal - see footnote -t in para 2.8)) is said to have been threatened by extinction in the 1970s because of the failure of the families to properly pollinate the plants by hand (as the trees kept growing at their extraordinary rate). What saved the program was the identification and import from West Africa of a suitable weevil that has, happily, taken over responsibility for pollination. A less tractable problem at Kimbe, and this is also reported from PNG's other oil palm settlement schemes, is the resistance of the sons to assume a responsible role in the management of their father's block. They help, but not much, and not enough to make sure the flow of bunches to the roadside is maintained ("waving to their father as he goes to work"). The price decline of course had an important impact on the sons' motivation. But the phenomenon has deeper roots in the several tribal societies from which these families came. Sons, for example, sometimes are unwilling to 11 The third phase two-bectare subdivision (of the 6 hectare block), most of which were planted in the early 1980s, when added to the part newly replanted, completes the 4 hectare total oil palm mini- plantation which under the present strategy the program f-a obliged to provide to each settler. What he does with the other 2 hectares - the second phase planting with its tall, useless, trees - is up to him, although the concept is to encourage diversification. - 14 - exert themselves until the inheritance of the block is defined, and until their position via a vie their brothers is clarified. Meanwhile the father waits as long as he can to name his successor, because his authority would disappear immediately thereafter. These issues are documented in several anthropological studies of the area and of the homelands, most recently by a remarkable three- volume 1991 study of Kimbe commissioned by the ADB. The audit will not explore further the social organization except to say that earlier predictions of farmer behavior are now looking overly optimistic, as the original farmers grow old and their immediate male relatives look around for less demanding jobs or hang around the village stores. One has to try hefting the long harvesting poles, with knives extended, to appreciate the difficulty of this job as the trees grow up and the fresh fruit bunches of oil palm fruit rise further out of reach. 3.12 There are grounds for optimism in this bleak landscape, if the payments can be increased. This will happen automatically as the crop delivered to the mill continues to grow. The expansion of production is a fact. It reflects the results of replanting of the nucleus and settler plantations, and the harvest coming on stream from the newer village plantings. Together these increasing deliveries will bring the Mosa mill up to capacity throughput and improved economies of scale. Nevertheless the overall scenario is discouraging, especially compared to the enthusiasm of the 1970s. That enthusiaa. is best reflected in the up-beat assessments in ORD's PPAR and Impact Evaluation Report (IER) on the first Hoskins project, issued in 1976 and 1980 respectively (OED Reports No.1400 and 3070). That earlier optimism has faded. The mood of the settlers has changed. Inter-tribe fights, which were common in the early years, have stopped. But settler frustration has begun to erupt in riots against Mosa and DAL, both of which were attacked by mobs of settlers earlier in 1991. 3.13 The PCR recalculated the economic rate of return for the 4 ha oil palm model at 8Z. All statistics on smallholder yields from Hoskins show declining yields for the settlers. This is not true for the villager program, which provides 2 hectares to interested residents of the areas on communal ("customary") land that had not been "alienated" for leasing to foreigners or for Government programs. But the village program is new, much smaller than the settler program in extent, and the planted area is close to the roads. For the settlers, the PCR used yield estimates of 17 ton/ha ffb (fresh fruit bunch), down slightly from the SAR 18 tons/ha. The Bank's present consultants refer to a better benchmark figure of 14 tons/ha, but admit that is optimistic and later quote an actual average of 11.3 tons/ha for all smallholders in PNG. The audit mission did not bother to recalculate the PCR's ERR for the 4 ha oil palm plantation: it would have to bb lower than 82. The PCR also shows an FRR of 62. Since yields have dropped, and the Bank's price forecasts available to the PCR mission in 1990 have not changed significantly, we can assume that an updated FR would be close to zero. The SAR estimate was 45%. 3.14 The story for coffee smallholders is said to be similar, with yields and take-home earnings reportedly in retreat everywhere. - 15 - FIGURE 1 Cocoa Prices (0/kg, 1985 constant) 40 00.......... 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2006 Coffee Prices (0/kg, 1985 constant) 800 700 100 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2006 Deflated by Manufacturing Unit Value (MUV Index ........ Deflated by US GNP deflator Sources World Bank, International Economic Department - 16 - FIGURE 2 Palm Oil Prices ($/ton, 1985 constant) 140k 12Dl ............./ 600 190 1955 1980 1986 1970 1975 1980 1985 1990 1995 2000 2005 Copra Prices ($/ton, 1985 constant) 140 1000 400> 200- 950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Defiated by Manufacturing Unit Value (UV Index Deflated by US GNP deflator Source: World Bank, International Economic Department - 17 - The Socio-Economic Impact of Rural Credit Study 3.15 The study was commissioned by ABPNG in 1988, and carried out by the PNG Institute of Applied Social and Economic Research (IASER). The contract was the culmination of an extended interaction between the Bank and the Borrower over execution of a monitoring program and group of evaluation reviews agreed at negotiations. As a separate exercise, the Government had also agreed to carry out the study of the farmer's sensitivity to changes in interest rates - part of the Bank's effort to persuade Government that farmers would accept higher rates. Ultimately the interest rate study was incorporated in the IASER contract. The monitoring component was dropped. Because a sample drawn exclusively from Third Project borrowers would have been too fresh to provide evidence of impact, the Bank agreed to let IASER sample intensively in three areas, one from the present project and two associated with older Bank projects. One area was the East New Britain cocoa zone. This sample would include all three categories of ABPNG cocoa borrowers under the Third Project. A second was the coffee zone, which included Second Project (but not Third Project) beneficiaries. The third was oil palm settlers from the Popendetta scheme on the mainland, younger than the Hoskins scheme but with the settlers in the midst of the payback period. Draft reports on each zonal study were discussed in late 1988 at IASER. The final report was presented to ABPNG in January 1988. It received special attention in the PCR, particularly in Part II prepared by Government. Who prepared and released this commentary in Part II is uncertain, because ABPNG decided neither to approve nor release the IASER Report. IASER has never received any further feedback, and the audit mission showed the favorable remarks in the PCR to the IASER authors for the first time. 3.16 We must assume that a decision to withhold the report vae taken at the board level in ABPNG and was related to the potential embarrassment of the principal finding, and the one the report elevates to prominence: that, largely due to the price decline, but also to the recognizable self-interests of the management agencies, whatever profits remained after debt repayment were reported by IASER to have been consumed by two participants who had nothing to do with the primary targeted beneficiaries identified in the SAR. First, the agencies themselves. Second, the migrant workers employed on the large and small plantations. The owners - the target group - received practically nothing. One of the reasons was a fact highlighted in the report but not yet mentioned above - that the group owners had no intention or interest in working on their own plantations. Some worked on other estates. But they saw ownership as a sign of prestige and not as a source of wage income. Even as prices fell, and the prospects disappeared of receiving dividends in the near future, the owners were not willing to work there. In the cocoa zone, in fact, the management agencies had to establish patrols to keep the owners off the plantations, because the trespassers' main intention was to steal "the-Ir" beans to sell to the roadside buyers. The report deals with other interesting social issues, that will not be discussed here. 3.17 With respect to the interest rate issue, IASER was able to demonstrate that group owners and settlers were generally insensitive to the rate of interest. This was particularly true on the large cocoa and coffee estates, - 18 - where the group owners had not the foggiest notion of the rate and left it and other financial business to the agents. The latter were concerned with the rates, of course. But that was in the context of the financial analyses they had to present to ABPNG. There they saw that the expected financial returns, based on the same price forecasts the Bank was using in the early project period, showed returns that could easily cover the ABPNG charge of 9.52.Y1 The Bank seized on the interest rate findings in the IASER Report: they are highlighted in the PCR as evidence that the Government's position on interest subsidies was mistaken. Other Proiect Investments 3.18 As mentioned above, the audit mission did not investigate the results of other parts of the lending program. It collected anecdotal evidence on a few loans to poultry, irrigated vegetable and truck enterprises in the vicinity of Port Moresby. The results are mixed. The two poultry farmers visited cannot expect to see significant profit to their enterprise for many more years, given the remaining debt, and this is so after four years have past without appreciable return. The branch agent said that was a common condition for poultry participants in the Port Moresby area. The poultry loans are not in arrears, however, primarily because repayments are deducted by the commercial broiler processor that signs contracts with the farmers. Half of the irrigation pumps in the small vegetable schemes in the branch's area have had to be repossessed. Finally, most of the truck loans were fully paid (the hire-purchase contract almost guaranteed that), although it was alleged that a sizeable number of the trucks have been immobilized on the rough roads, were packed up in the borrowers' yards, or, if the breakdown had occurred before the debt was liquidated, may be sitting in ABPNG yards, waiting for buyers. A grand judgement based on this quick exposure in the environs of Port Moresby would raise doubts about the economic and financial impact of the other parts of the Third Project lending program, and the quality of these sections of the bank's portfolio. But it is too tiny a sample of the 1,300 accounts (para 2.5) to extrapolate. The socio- economic report, the PCR, and the ADB's PPAR do not shed any light on these other investments. A Follow-on Proiect 3.19 A fourth project was discussed with Government during the first supervision mission in 1987. The Bank advised that neither cocoa nor coffee could be included for Bank finance, and that that restriction would apply to rehabilitation and/or completion of projects initiated on farms included in the Third Project. Discussions of the follow-on continued, but for various reasons the interest on both sides waned. The crop restriction was one. Also, the Japanese aid agency - the Overseas Economic Cooperation Fund - had entered the arena, and in 1988 signed a US$18.0 mill-on credit agreement to support ABPNG VJ The principal officer of one of the agencies told the audit mission that this complacency was not unlaliteds that he supposes now, and supposed then, when most of the applications were being prepared in 1986 and 1987, that an interest rate above 152 would render some if not most of these cocoa estate rehabilitation projects nonviable. - 19 - across all of the components covered by the Third Project. This included cocoa and coffee, and in many cases picked up the same estates. The Japanese position on farmer interest rates and subsidies was diametrically opposite the Bank's. Although the agreement did not call for lower onlending rates, there was no pressure on ABPNG or Government to raise them. The Bank, meanwhile, had been disappointed with the lack of any serious response by Government to the provisions of the Third Project dealing with interest rate review and the reduction of subsidies. That, plus ABPNG's failure to reduce costs and other evidence of political interference with the objectives agreed in the 1985 Action Program, finally persuaded the Bank to drop the proposed project. A comment in a Bank letter dated July 1990 to FAO/CP, discussing the draft PCR, captures the Bank attitude toward Government and ABPNG intransigences "the fact that the Bank is not preparing a follow-up project to Agricultural Credit III indicates our firm resolution to seek critical changes in the policy and institutional framework of ABPNG, which should be allowed to operate as an independent, financially viable, and self-sustaining financial intermediary". With the deepening financial predicament, ABPNG is becoming more rather than less dependent on Government support. Also, with the decline in prices, the absorptive capacity of the agricultural sector for additional lending has itself declined sharply, so that the need for further support from the Bank at this time is not as clear as it was during supervision in 1987. 3.20 That is likely to remain the case until there is a change in the terms of trade for agricultural products vis a vie other sectors of PNG's economy. The macroeconomic setting is the focus of the Bank's ongoing review of the tree crop sector, as discussed below. IV. FINDINGS AND ISSUES A. A Good Proiect under Stress 4.1 This is the first in a series of PCRs, PPARs and IER - dealing with PNGDB and ABPNG lending to the tree crop sector - to break a string of good ratings. As mentioned above, the Hoskins oil palm settlement program was reviewed by OED in an audit dated 1976, and revisited during an impact study in 1980. OED issued a cluster audit in 1987 that included the Second Credit Project, with its finance for coffee, cocoa and oil palm (the First Credit Project had dealt exclusively with livestock), and the Popondetta Smallholder Oil Palm Development Project. The findings were invariably favorable, especially with respect to the quality of the plantings, the increase in exports and the earnings of the farmers. The role of expatriates in project management was an important element of success, along with the dedication of the farmers and fertility of the land. At Hoskins, the 1976 PPAR elevated the SAR ER to 19%, and the IER four years later raised it again to 25Z. The Second Project audit gave an ERR estimate of 39% for the typical oil palm sub-loan (some of which had been disbursed at Hoskins), substantially above the SAR projection of 291. That PPAR also showed high re-estimated ERRa for coffee and cocoa subloans. The ERR - 20 - calculations were all accompanied by high values for recalculated FRRs. The PCR prepared in 1990 for the project under audit was influenced by the commodity price decline, and gave ERR and FRRs substantially below the SAR estimates. However, the recalculated EUR were still above opportunity costs, though the FRR were generally below 10%. The PCR concluded that the Third Project was also satisfactory. With a year and a half of additional hindsight, this audit is obliged to change the sign to unsatisfactory. 4.2 The important question is what does this reversal imply - not so much about the appropriateness of the original investment decision but about the Bank's role in future finance of farm credit and tree crops in PNG. The audit concludes that the cocoa and oil palm lending programs proposed for this project were sensible.- They were ambushed by an unprecedented collapse of international commodity prices from which early recovery is not anticipated. The appropriate response for the tree crop sector would have been an adjustment of the exchange rate in recognition of the decline in the comparative value of the country's overall trading position. The burden of the price decline would have been spread, real wages would have fallen, and financial returns to tree crop farmers would have been maintained at levels that would have kept the farmers' commitment to preserving their assets. However, government policy was turned in another direction. 4.3 This is a classic case of the "Dutch Disease", disabling commercial agriculture as economic policy is oriented around mineral export. The overvalued exchange rate and extremely high wage rate (by Asian standards) are part of that scenario. Nevertheless the country's tree crop sector must be protected from the threat of irreversible physical degeneration, signs of which are now evident. The Bank is responding to this threat by focussing on the tree crops as a group within an overall sector study, with the intentien of formulating a recovery program that is consistent with the objectiv6s of the macroeconomic scenario. Appropriate devaluation, strengthening of the price support programs, and some form of debt relief, especially for the plantations, are among the alternatives. ! 4.4 Another important observation is that this "credit" project faced an unanticipated hostile economic environment over which it had no control. The tools at the disposal of an agricultural financial institution cannot deal with a commodity price decline of this magnitude. Further support by the Bank for the tree crop sector must shift now from a line of credit projects to a line of subsector projects that have leverage over the larger policy environment, presumably including though not prioritizing a credit component. It made sense in 1980 and again in 1983 to appraise projects aimed in part at tree crops as P Recenx sector work by the Bank includes Domestic Resource Cost analysis. It shove that plantation cocoa is at a comparative disadvantage - via a via smallholder cocoa and all oil palm models. Sigh wages are part of the problem. Thue some restructuring of costs in the plantation cocoa sector is necessary even if coHEDdity prices were to Improve. See net Footnote. Il The white cover draft report is dated March 19, 19921 Pana New Guinea e Revitalising A&ricltare - Issues and Optiona. - 21 - Second and Third Credit Projects. It would not make sense under present conditions to prepare further support for tree crops as a fourth credit project. B. Role of Interest Rates 4.5 Government's position during appraisal was that PNG's subsistence farmers needed special support to prompt them to take up modern agricultural methods. The subsidy provided through a low interest rate was seen to be essential. The Bank argued that higher rates had advantages for the lending institution and resource allocation in general, and would not deter small farmer investment in viable enterprise. The issue dominated the protracted appraisal period. Government agreed to raise the basic rate from 8.5% to 9.5Z, and, within the first year of effectiveness, raised the rate again and added a "bank charge" as well. In 1988, under the pressure of the collapse of commodity prices, the trend was reversed and the rate was lowered to 8%. The Bank believes Government has not moved significantly throughout the period from what the Bank perceives to be an ideological platform. The Bank highlighted the finding of the IASER study - that the participating farmers were both ignorant of and indifferent to the interest rate. But that finding had no impact on Government's attitude. 4.6 The role of the interest rate in influencing behavior of the borrowers was exaggerated by the Bank as well as by Government, in opposite directions.e Whereas the Government claimed subsidies were needed to stimulate investment, the Bank argued that subsidies would invite credit diversion, substitution and other forms of misallocation. But these undesirable effects seem actually to have been smaller than the Bank had anticipated. Diversion of funds for purposes other than those agreed between ABPNG and the farmers was negligible. At Hoskins, given the procedures for financing the poisoning, removing and replanting of the oil palms, diversion would have been impossible. In the cocoa zone also, diversion of credit aimed at the estates and 20 ha plantations would have had to have been engineered by the management agencies, not the farmers. Given the agency fee formula and ABPNG's inspection routine, that maneuver is implausible. Diversion of credit intended to finance the trucks is equally implausible, given the hire purchase arrangements. At least that is the conviction of ABPNG, and Bank supervision did not argue differently. Diversion of the trucks themselves for non-agricultural uses was likely, but the multiple use of these vehicles was a fact of life and did not invite the Bank's objection. 4.7 There is general agreement that another undesirable effect of low interest rates - the substitution of project funds for resources already available from the farmers or from their other banks - did occur. This issue comes up in the socio-economic survey, where the participation of "notables" is described. It concludes that some of these leading citizens had access to other funds, but switched instead to cheaper project finance. However the IASER report does not pretend that this element dominated the project portfolio: the evidence V The ADB has also argued strongly for higher interest rates, in a report *Agricultural Credit and Rural Savings in Papua New Guineas issued in three volumes in 1991. The report was prepared by a Sri Lankan consultant firm in collaboration with LASER. Whether ADB headquarters subscribes to the forceful argument for removing the subsidy is unclear. The Japanese Government apparently does not. - 22 - is rather anecdotal and illustrates more the point that the target population cannot always be reached. Given the nature of the average member of the group owners, the Hoskins settlers and the small scale entrepreneurs who bought the trucks, the incidence of substitution must have been low. 4.8 If there vas no significant misallocation effect attributable to the subsidy, then the Bank's case for higher rates mainly depended on the damage done by subsidies to the earnings and institutional viability of the bank itself - ABPNG. That case is mu2h stronger: the loss in interest income contributed to ABPNG's mounting financial distress. However, in this case the effects of the subsidy were swamped by the effects of the commodity price decline. The poor position of ABPNG at present is not closely related to the low interest rate agreed at appraisal. Rather, it is a function of the mounting losses on the borrowing farms and the suspension of all interest charges and increasing arrears that were their consequence. Higher interest rates would only have meant higher write-off a and arrears. The Bank could not have anticipated this turn of events. But the continued hot pursuit by the Bank of the interest rate issue in the face of the commodity price decline seems, in retrospect, inappropriate. Since misallocation was not a factor, the Bank's insistence could be justified only in terms of institutional damage. Yet, in that respect, the interest rate was the less important of the pricing problems. The subsidy was given a prominence it did not deserve in Bank-country dialogue. The main threat to the project was coming from a different direction. C. Low Equity Contributions to High Input Investments 4.9 For the tree crop loans, group owners and individual farmers and settlers entered the project with a low cash equity input, often zero. This contributed to the negligible earnings available to the owners after debt service. One conclusion is that the cash equity requirement was too low, that few enterprises can sustain that level of debt, and that the formula was an invitation for farmer dissatisfaction. The audit is persuaded, however, that when posed at this simple level the conclusion is uninspired. In fact higher equity was in many cases unavailable: the group owners of the plantations would not have been able to make that contribution. The expectation that the returns would be good enough to reward the owners as well as the bank was simply another casualty of the extraordinary lose in commodity value. Apparently the cocoa rehabilitation loans placed by PNGBC during the 1980s required a higher equity share and on the whole have fared no better than the ABPNG loans. 4.10 But there are two other issues hiding in this discussion which are being taken seriously. Both are brought up in the IASER report, the ADB PPAR and the more recent consultant reports submitted to the Bank. One is that the group ownership models, including the estates and the 20 ha plantations, have revealed themselves under the stress of the price decline as being inappropriate forms of enterprise for PNG's farmers. That is the basis for IASER's conclusion that the project's objectives for production and incomes were fundamentally inconsistent. Target groups operating through management agencies may achieve production increases, but not the associated income. Since they do not work on the plantation, they do not get wages. They could rent the property back to the - 23 - agents, and hope to collect a steady rental charge, but the agencies are unlikely to accept that arrangement and the risks it implies. So the group members are left with the risk and, in the present circumstance, no dividends. That is not acceptable either. Both the agency and group concepts are now under attack. 4.11 The second issue is that the Second and Third Projects' farm investment models may have been mis-specified for the target populations. These models are now described as high-input high-output programs, which are especially vulnerable to irregular management and major commodity price swings. The consultants' concern is mainly directed at the coffee zone, and in particular the 20 ha coffee plantation. It is thus somewhat outside the orbit of this audit, although the same factors are influential also in the cocoa zone. The turnaround in thinking about the group model has been abrupt. In OED's 1987 cluster PPAR which included the Second Project, the 20 ha coffee plantation was highly praised. The argument now would be that the individual owner who can make a substantial equity contribution may be correct in adopting the high-input approach. But the smaller farmer would be better advised to manage on his own a mini-planation of scattered trees within the context of his traditional farm plan, rather than join a group and pay for an agent. These are important and interrelated issues that must be addressed in future planning for the PNG tree crop sector. D. Role of Technical Factors 4.12 The input-output issue just mentioned is one of a number of important technical subjects which helped determine the course of the project. This is one of those types of credit projects which are heavily influenced by the successes and failures of the technologies underpinning the investmentst technical factors which require the attention of project authorities (and Bank supervision). Among the others are: (a) the explosive growth of the oil palm trees on New Britain, which reduced their economic life from 25 years (SAR) to 18 years (actual) and forced the project authorities and settlers to replant - with credit - much earlier than expected; (b) the discovery in West Africa of a weevil suited for pollinating the Hoskins oil palm. Starting in 1982, the weevils quickly replaced the farmers, who were finding it difficult to keep up with that job; (c) the rains of 1990 and 1991, triple the norm in many parts of the cocoa zone and inviting the rapid spread of the black pod disease. As mentioned above, it is believed that some of the estates would have otherwise survived the price decline; and (d) the fact that the cocoa hybrid, which has been the basis for the rehabilitation program during the 1980s, may carry a genetic flaw, revealed during the recent rains, which will substantially shorten its period of peak yields and follow them with rapid decline. 1BRD 17314R 1444 7- 3 PAPUA NEW GUINEA M U S toen9°u NEW THIRD AGR ICULTURAL CREDIT PROJECT PROPOSEO PROJECT PAEVIOUS PROJECT KvengCOFFEE Vammo RUsBen Aitope -1.PAS N NE IR ANDPOIL,TRY NEW IRELAND AUOR A S ONATIONAL I FUTUE MA.0 OADS INATONAL) WES Wewak Mo. -NA PoAs IP-V S E P / K MOUNTAINOUS AAEAS r-.- --..J---- -- ------ - PAINPOVINCIAL BOUNDARIES ~ M A DAN G ---- INTERNATIONAL BOUNDARIES E AS5T SE PIK WEST NEW BRITAIN 0 0 e0 120 ff KVOMETES --BoL MAE a 0 0 to eo o 4 Madang Hoskins Sohano z Kibt 1--BOUGAINVILL El z oNk - 0° /..~~~~ NEW BRITAIN j OUANIL it - nW- Ki.t. < ung. 5 UTHERN .ILN HIGHLANDS *A ST NEW BRITAIN GHANDS MOROBEI z- WESTERN NORTHERN K K,,.win. o Eena . / Pp-ndetto ( BeK D- OroBo .N9. Port Ne C DENTRECASTEAUXN -o b -IStANDS . Mi L N E dyA y - 1GUL F PAPUA NEW GUINEA CENTRAL 1 12- ONDIAtU NEW 'mo. /N-/AN-A-A- -..-------------------------- ---- Sc A v 150 A53Y 19 __ MAY 1985