China - Guangxi Rural Credit Project

Document of The World Bank FOR OFF1CIAL USE ONLY ReportNo. 11125 PERFORMANCE AUDIT REPORT CHINA RURAL CREDIT PROJECT (CREDIT 1462-CHA) SEPTEMBER 22, 1992 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their ofricial duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS Name of Currency: Yuan (Renminbi) Rate of Exchange: Appraisal US$1.00 - Y 2.0 1985 US$1.00 - Y 2.9 1986 US$1.00 - Y 3.2 1987 US$1.00 - Y 3.7 1989 US$1.00 - Y 4.7 1990 US$1.00 = Y 5.2 Fiscal Year of Borrower: July 1 - June 30 Weights and Measures: Metric System 1 Meter (m) - 3.28 Feet (ft) 1 Kilometer (km) - 0.62 Mile 1 Hectare (ha) X 2.47 Acres ABBREVIATIONS ABC Agricultural Bank of China CAM Country Assistance Management CCE Brigade and Commune Enterprises EDI Economic Development Institute ERR Economic Rate of Return FAO Food and Agricultural Organization of the United Nations FRR Financial Rate of Return GDP Gross Domestic Product GZAR Guangxi Zhuang Autonomous Region IDA International Development Association IFAD International Fund for Agricultural Development IWC Incremental Working Capital MOF Ministry of Finance OMS Operational Manual Statement PAR Performance Audit Report PCR Project Completion Report PRE Policy, Research, and External Affairs RCC Rural Credit Cooperatives SAR Staff Appraisal Report  FoR omlcIL US ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation September 22, 1992 HEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on CHINA Rural Credit Proiect (Credit 1462-CHA) Attached, for information, is a copy of a report entitled "Performance Audit Report on China Rural Credit Project (Credit 1462-CHA)" prepared by the Operations Evaluation Department. Attachment This document has a restricted distribution and may be used by reipients only in the performance of the official duties. Its contents may not otherwise bo disclosed without World Bank authorization.  I  FOR OFFICIAL USE ONLY PERFORMANCE AUDIT REPORT CHINA RURAL CREDIT PROJECT (Cr. 1462-CHA) TABLE OF CON1 age No, PREFACE ............................................. i BASIC DATA SHEET .iii EVALUATION SUYMARY. v Bakground. 1 Project Identification and Preparation. 2 Project Objectives and Description. 4 Inpenntation Experience. 6 Introduction. 6 Project Size and Subloan Processing Problems. 7 Counterpart Funding and Incremental Working Capital. 7 Procurement .8 Training, the Role of EDI and Technical Assistance. 8 Supervision by the Bank. 9 Results .10 Findings and Lessons Learned .11 (a) Project Size .11 (b) Technology Transfer-a Missed Opportunity .12 (c) Supervision by IDA .13 (d) The Orange Market .14 (e) Filing .15 MAP: IBRD 17483R4 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.   - 1.- PERFORMANCE AUDRR RURAL CRED1T PROJECr (Cr. 1462-CHAA PRE3FACE This is the Performance Audit Report (PAR) of the Rural Credit Project for which Credit 1462-CHA in the amount of SDR 47.2 million (US$50 million equivalent) was approved on April 24, 1984. The objectives of the project were to increase agricultural and rural production and incomes by supporting the lending programs and institutional development of the Agricultural Bank of China (ABC) in Guangxi Zhuang Autonomous Region (GZAR). The credit was fully disbursed and closed on schedule, on June 30, 1989. The PAR is based on the Project Completion Report (PCR) prepared jointly by staff of the Asia Region of IDA (Preface, Evaluation Summary, Parts I and III) and the Borrower (Part II), which was issued on March 11, 1991, and on the Staff Appraisal Report, the President's Report, the credit documents, a review of IDA files and on discussions with Bank staff. An OED mission visited China in August/September 1990, discussed the effectiveness of Bank assistance with Borrower agencies and visited a sample of ABC branches and enterprises financed under the project. The kind cooperation and valuable assistance of Chinese staff in the preparation of this report is gratefully acknowledged. The PCR provides extensive details of the project's formulation, implementation and results. The PAR notes the project's substantial physical achievements and contribution to ABC's institutional strengthening. The report then elaborates on aspects relating to the expansion of the on-lending program as affected by devaluation, probable over-investment in citrus, the missed opportunity for technology transfer, the failure to allocate the foreign exchange risk to subborrowers able to bear it, and on the inadequacy of the Bank's contribution through supervision of the project. Following standard OED procedures, copies of the draft PAR were sent to the Borrower for comments. However, no comments were received.   - il$ - PERFORMANCE AUDIT REPORT CHINA RURAL CREDIT PROJECT (Cr. 1462-CHA) BASIC DATA SHEET PROJECT DATES Date Planned Dat- Actual Identification (PB) 01/05/83 Preparation 04/02/83 Appraisal Mission 08/29/83 Negotiations 03/12/84 Board Approval 04/24/84 04/24/84 Signing Date 06/01/84 06/01/84 Effective Date 11/01/84 11/01/84 Completion Date 06/30/89 06/30/88 Closing Date 06/30/89 06/30/89 CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (USS ' 000) Appraisal Actual as Z Actual as Z Ouarter Ending Estimate Actual of Estimste of Total 1984/1985 I 0.5 0 0 II 1.0 4.95 495 9 III 2.5 5.70 228 10 IV 4.0 9.57 239 17 1985/1986 I 6.0 12.72 212 23 II 8.0 15.67 196 28 III 11.0 18.56 169 33 IV 14.0 22.11 158 40 1986/1987 I 17.5 24.61 141 44 II 21.0 26.71 127 48 III 25.5 30.58 120 55 IV 30.0 34.75 116 62 1987/1988 I 34.5 37.47 109 67 II 39.0 43.24 111 77 III 43.0 46.06 107 83 IV 47.0 48.00 102 86 1988/1989 I 48.5 55.64 115 99 II 50.0 55.82 112 a/ 100 gl Th acua asmo disbunud incrsaad in kim of USS dw to devauation of US$ vs. SDR during imple.mantion.  - iv - MISSION DATA Stage of Month/ No. of Staff Days Performance Proiect C.clo Year erns in Fieldj Specialization Rating Reconnaissance 10/81 4 21 Econ/Eng/Fin. - Identification 4/82 4 21 Econ/Fin.Ag.Ec. - 7/82 1 9 Ag. Econ. Preparation 1/83 1 14 Agronomist 4/83 8 28 Ag/Econ/Econ/Fin. - Eng/Poultry Spec. Aquacult. Appraisal 9/83 6 28 Ag/Econ/Econ/Eng. - Aquac. Poult. Spec. Supervision 1 A/ 11/84 6 28 Ag/Econ/Econ/Aquac. 1 Poultry Spec. Supervision 2 12/85 1 7 Ag. Econ. 1 Supervision 3 9/86 1 14 Fin. 1 Supervision 4 12/87 2 7 Fin. Agron. 1 Supervision 5 b/ 10/88 1 n.a. Fin. 1 PCR Mission 3/90 2 10 Econ/Agroproc. 1 a/ Combined with identification of Rural Credit II. b/ In Beijing. Combined with identification of Rural Credit IV. STAFF INPUTS FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 Total Preappraisal 9.0 54.2 12.7 75.9 Appraia-l 50.7 50.7 Negotiations 5.1 5.1 Supervision 15.5 6.9 5.4 7.2 2.1 37.1 Total 9.0 54.2 68.5 15.5 6.9 5.4 7.2 2.1 168.8  - v - PERFORMANCE AUDIT REPORT CHINA RURAL CREDIT PROJECT (Cr. 1462-CHA) EVALUATION SUMMARY Introduction processing), and (b) strengthening of ABC through technical assistance (es- 1. About 80 percent of the people of pecially project appraisal methodolo- China depend on agriculture and Guangxi gy). The enterprises to be financed Zhuang Autonomous Region (GZAR) is one were orchards (mainly citrus), fisher- of the poorest provinces, with per ies, livestock and agro-processing. capita gross output a third lower than the national average. Introduction of the "production responsibility system" Implementation Experience in China has accelerated growth of agriculture and increased demand for 4. The project was completed well credit. ahead of schedule and far exceeded appraisal lending and physical tar- 2. The Agricultural Bank of China gets. This expansion resulted from (ABC) lends for production, processing devaluation such that on-lending funds and marketing, provides a range of more than doubled. Consequently the financial services and manages the processing capability of ABC was rural credit cooperative system. ABC/ stretched beyond safe limits and de- Guangxi has about 11,000 staff in about mand for counterpart funds could not 1,100 offices around the region. The be met. The rate of subloan process- Rural Credit Project was the fifth Bank ing was increased greatly to keep to Group assisted project in China and the the original commitment schedule. As first foreign funding of ABC. a result, processing was sometimes skimped and "some marginal quality Project Objectives subprojects" were approved. Also, follow-up by ABC staff to assist sub- 3. The major objectives of the pro- borrowers was unavoidably limited. ject were to (a) support a three-year lending program of the ABC (which was 5. There was also a shortage of part of a high-priority government dev- good quality planting material and of elopment effort to increase agricul- agricultural inputs. Shortage of tural and rural production and incomes fundings was a constraint throughout through the diversification and modern- implementation, resulting in inade- ization of agricultural production and quate working capital for subprojects,  - vi - which affected especially the large ahead of schedule, according to the area of perennial orchard crops. PCR. 6. Project implementation began 9. Other indicators of results before the credit was approved with given in the PCR are very impressive respect to important institutional and with estimated households benefitting training activities. The Bank's Eco- more than doubling over the SAR figure nomic Development Institute (EDI) to about 200,000 and creation of the played a valuable role in organizing equivalent of some 380,000 full-time training courses in China on rural jobs (a fourfold increase). credit, especially on appraisal meth- odology. Further training courses were 10. Some 1,240 subprojects were held for both ABC/Guangxi staff and ABC completed, but the project mix by staff from other provinces. ABC also scale and cost was quite different found Bank staff assistance very valu- from that expected at appraisal. All able in training of auditors and in sub-components were greatly expanded, setting up the project's evaluation and with investments in citrus more than monitoring system, with the help of doubling and agro-processing invest- foreign short-term consultants. ments tripling. About 28,000 ha of citrus were planted, three times the 7. Funds allocated for hiring for- appraisal plan of 9,000 ha, and 118 eign consultants for training assis- agro-processing units were financed tance were not used, partly because instead of 12. The citrus area fi- training received in EDI courses and nanced was a substantial expansion of Bank staff contributions during ap- 28 percent for GZAR as a whole and 5.6 praisal resolved some technical issues. percent of China's total citrus area Bank staff also assisted ABC to draft of 506,900 hectares in 1985. Taken a subproject appraisal manual, which is together, the tree crop planting pro- now widely used in ABC, and with ar- gram financed over 70,000 hectares, rangements for study tours. compared with 18,600 hectares origi- nally planned, an almost fourfold Results increase and a very large additional burden on ABC's processing and super- 8. The project made a substantial vision resources. Livestock financing contribution to development of agri- also expanded greatly. culture in GZAR, especially for tree crops, livestock and agro-processing. 11. Institutionally, ABC was greatly Financial rates of return on subloans strengthened with respect to appraisal are reported as very healthy at around methodology and trained staff, espe- 25 percent, similar to appraisal. Eco- cially in GZAR but also throughout the nomic rates of return range from 17 bank by the training of trainers and percent to 26 percent, with a "solid wide distribution and use of the ap- overall average of 20 percent," com- praisal manual. pared with 27 percent at appraisal. However, there may be some uncertainty Findings and Lessons Learned surrounding some rates of return given the falling prices for citrus, the main 12. Project Size. It was unfortu- investment. Repayments on subloans are nate that ABC and government staff,  - vii - with no apparent objection by IDA, orange planting and to place emphasis stuck to a project processing schedule on assisting post-harvest investments, and commitment period which was too was given by the Bank to ABC. Subse- short once the size of the project quent to the project under audit the expanded, driven as it was by devalued Bank Group has assisted a surprisingly Yuan generated by the IDA Credit. Had large number of projects where the the commitment period been extended, main component was establishment of presumably fewer marginal projects orange orchards. would have been financed and supplies of working capital and good quality 16. Although demand for citrus has planting material would have been less been expanding rapidly, clearly the constraining. The consequences of spectacular planting response in the approving marginal loans and of estab- citrus areas to the earlier high pric- lishing orchards with inferior planting es, and the Bank Group's widespread material will emerge in due course. support of new planting, have been instrumental in lowering prices and 13. Technology Transfer--a Missed Op- making oranges more generally avail- portunity. The major objective of the able. This is a great success story project was to support a lending pro- but the market situation needs to be gram by ABC which was part of a high kept under close review and the mea- priority government development effort sures adopted, inter alia, under the to increase agricultural and rural project to expand post-harvest facili- production and incomes through diversi- ties and to spread the peak harvest fication and modernization of agricul- period, need to be continued. The tural production and processing. The financial performance of orange sub- project's procurement arrangements loans, and therefore the loan repay- restricted most purchases to domestic ment rate could be threatened by the goods. Constraints to importation did fall in prices. exist at the time but, in view of the importance of enhancing the level of 17. Supervision by IDA. The PCR, in technology, the Bank should have made both the Bank's and ABC's sections, a greater effort to ensure that appro- records that the borrower expected priate, technically advanced equipment more assistance from the Bank during was available to subborrowers. This implementation. This was verified in equipment was not available and the comments to the audit mission. ABC in project therefore did not support particular, while complimentary about government's modernization strategy to the Bank staff's contributions, stated the extent intended. that there were too few supervision missions and that the technical exper- 15. The Orange Market. Through the tise of missions was not always appro- project, IDA has been prominent in priate. The audit confirmed that the support of the rapid expansion of the Bank's contribution to the project's Guangxi citrus industry, but there is implementation was less than the situ- now concern that falling prices may ation demanded. Current supervision lead to marketing and profitability guidelines require that a supervision problems. This problem was identified plan be agreed with the borrower be- during project implementation and ap- fore Board approval to minimize sur- propriate advice, to cease funding prises on both sides.  - viii - 18. Filing. References to file prob- ject implementation schedule, when lems occur several times in the PAR circumstances change, can have unde- text. In particular, there is a nota- sirable consequences, and that flexi- ble absence of some important incoming bility is preferable; that procurement messages on file, although other docu- arrangements can have a critical im- ments refer to such messages. Also, pact on technology transfer objectives the original "project proposal," sub- for better or worse; that perennial mitted by ABC and referred to in a Bank crop expansion can lead to marketing project brief, was not found in the and profitability problems when supply files. expands too quickly, and that Bank supervision inputs need to be better 19. Management attention to filing planned and agreed with borrowers.a/ deficiencies, of which this is but one example, would be worthwhile and possi- bly could make a small contribution to improving operational efficiency and a/ Comments on a draft of this report project results. received from the Bank's regional office and the Legal Department 20. This project experience suggests have been added to the main text that adhering to a predetermined pro- or footnoted.  PERFORMANCE AUDIT REPORT CHINA RURAL CREDIT PROJECT (Cr. 1462-CIA Background 1. The agricultural sector provided in the early 1980s about 35 percent of China's GDP and supported about 800 million people out of a total of about one billion. Production systems are intensive since only about a tenth of the land area is arable, with China meeting its basic food requirements for 22 percent of the world's population from less than 8 percent of the world's arable land. People's communes, state farms and privately managed plots accounted for 87, 5 and 8 percent respectively of arable land and 80, 4 and 16 percent respectively of the gross value of agriculture production (SAR, para. 1.02). 2. The historical sector growth rate of 3.2 percent from 1955-79 increased to over 5 percent after 1979. Reforms since 1976 have decentralized production decisions, increased farm gate prices substantially, increased family plot size, expanded rural markets and placed more emphasis on loan financing through ABC rather than grant finance. Introduction of the "production responsibility system," which provides incentives by allowing farm labor to benefit directly from increased output, led to greater demand for credit. 3. Guangxi Zhuang Autonomous Region (GZAR), the project area, is in southeast China (see map) in the tropical or sub-tropical zone on the border with Vietnam. To the east Guangdong Province is the hinterland of Hong-Kong and Guangzhou (Canton). GZAR had a population of 37 million in 1982, with about a third from the Zhuang minority group, and was the eighth poorest of the 29 provinces and autonomous regions. Per capita gross output is a third less than the national average. Ninety percent of the population is rural. GZAR is tropical and subtropical (with annual rainfall of about 1,560 mm.), with 60 percent of cultivated area irrigated. Rice production has increased rapidly since 1970 (4-6.5 percent p.a.). As a proportion of national production, GZAR produced 18 percent of the sugarcane, 8 percent of the jute, about 3 percent of both grain and tobacco and 4 percent of the fruits (mostly citrus). Fruit production in GZAR grew by 50 percent from 1980-82. In the early 1980s, GZAR's labor force of 12.9 million was organized into 268,300 production teams, 13,768 production brigades and 1,038 communes, with 541 state farms run mostly by county bureaux. Physical infrastructure is relatively well developed, including 1,710 km of railway line and river passage to Hong-Kong. 4. The Agricultural Bank of China (ABC), one of eight national-level financial institutions, was restored in 1979 by the State Council as the dominant institution for the mobilization of rural savings and the financing of the agricultural sector. ABC is charged with contributing to the effective execution of government economic policy in the rural areas, including not only agricul- tural production lending, but also lending for commercial and industrial activities which support agricultural development. In addition to its main task of mobilizing rural saving and extending credit, ABC supplies a wide range of banking services to rural areas, provides leadership to rural credit cooperatives (RCC) and financial advice to communes, brigades and production teams (SAR, paras 2.01 and 2.02).  - 2 - 5. In the early 1980s, ABC had 27 branch offices in the provincial capitals and autonomous regions, 2,499 subbranches, 86 business offices in major urban centers, 24,803 offices at communes and 403 offices at state farms. In addition, 59,000 RCCs were effectively part of the ABC network. ABC staff proper totalled 285,000, plus 250,000 RCC staff, of which less than 5 percent were graduates (SAR, paras 2.02 and 2.05). In 1982, ABC's balance sheet showed the equivalent of US$26.5 billion in deposits, US$38.6 billion in loans (an increase of 26 percent on 1981), and capital grants from government totalling US$11 billion. Income was US$2.4 billion, expenditure US$1.8 billion, giving an "Accumulated net gain" of US$577 million (SAR, Annex 1). About 80 percent of ABC's portfolio was short-term loans and 42 percent was loans to individuals. Information on overdues was not available as loans were "generally rolled over and kept nominally current" (SAR, para. 2.17). 6. ABC in Guangxi (ABC/Guangxi) had a regional office in Nanning, 194 municipal, prefectural or county offices, 974 commune offices and 25 offices on state farms. In addition, 3,087 RCC credit stations provided services at commune and brigade levels. There were 11,390 staff in ABC proper and a further 11,360 staff in the RCCs controlled by ABC/Guangxi. Sixty percent of the ABC and RCC staff had at least 20 years banking experience, but only 8 percent of ABC staff were university graduates. Deposits in 1982 totalled US$446 million equivalent, equity from ABC Beijing was US$318 million and loans outstanding amounted to US$840.5 million, of which a fifth were medium- and long-term loans. ABC/Guangxi reported a profit of US$18.5 million equivalent in 1982 (SAR, paras 3.04-3.05). Project Identification and Preparation 7. This fifth Bank-assisted project for the agricultural sector in China was scheduled in the Bank's lending program from the early days of Bank operations in China (the 1981 Country Program Paper included proposed rural credit projects in FY84 and FY87). As part of the identification process, the Bank sent a rural credit mission in late 1981 which produced "Rural Finance: A Sector Study" (Report No. 3864-CHA, of August 31, 1982). The mission noted that "Extension of the production responsibility system is likely to add significantly to the demand for credit and banking services at the household level" (para. 25). It also reported that ABC's interest rate structure, staff training and appraisal process ("the weakest area") required attention, but that the role of external assistance "will be mainly catalytic, in view of the small amounts of external funding likely to be involved relative to annual loan volumes..." It concluded by observing that "Assistance in strengthening ABC's institutional capabilities is ultimately likely to be as valuable as the financial resources which are provided" [by donors]. This project was to be the first international loan taken by ABC. 8. In the sector mission's back-to-office report it had noted that "Among the provincial offices visited, ABC/Guangxi seemed to be the most interested in the possibility of tapping World Bank financial resources and technical assistance..." "Based on our quick assessment... orange production for processing (juice or canned sections) appeared to be financially attractive" (files, 12/8/81). Such details were not included, however, in the sector study and it was to be almost a year before Guangxi and Hubei Province were identified in a telex to ABC of November 1982 as the target project areas. During this period, the sector study was discussed with the Chinese authorities and ABC was advised in May 1981 that the choice of one or two provinces for the first project was the next critical step. The selection of Guangxi and Hubei was apparently made by ABC in its project proposal (according to a later project brief), although there was nothing found on the files in this respect. Delay on the Chinese side was attributed, in the files, in part at  least, to ABC's concerns on whether IDA funds were going to be available. Given IDA's replenishment uncertainties, the Bank had advised ABC that neither the IDA credit amount nor precise timing of the project's Board presentation could be predicted. 9. In January 1982, IFAD had advised the Bank that it was expecting ABC to formulate a project for IFAD finance, possibly for appraisal by the Bank. It was then agreed to process a single project for joint financing by IDA and IFAD, and IFAD contributed to the various missions which followed. However, the government was concerned that joint financing might create difficulties since IFAD funds were channelled through the Ministry of Agriculture whereas IDA funds were channelled through the Ministry of Finance. In July 1983, the government suggested splitting the project, with IDA funds (US$50 million) going to Guangxi, and IFAD funds (US$25 million) to Hubei. The Bank resisted the split initially, preferring to resolve the coordination problems instead, but eventually agreed in late 1983, after appraisal, to the joint request of IFAD and government. The IFAD project (IFAD Loan No. 153-CH) was then processed and super- vised by the Bank as a separate project, but with the same staff and missions. (The Bank's official files, however, were apparently not separated at this time, causing some confusion to readers. See further below under Findings and Lessons Learned.) 10. Throughout the project preparation there was concern on both the government and Bank sides over the credit amount. To the Bank, China was expected to be an increasingly large borrower, with amounts in the lending program of US$0.9-1 billion for FY84, US$1.4-1.6 billion for FY85 and about US$2 billion for FY86. US$100 million was originally allocated in the lending program for the project, but the loan officer noted with implicit concern in January 1983 that the two rural credit projects in the pipeline would require small amounts of IDA funds compared with the average project size planned for the whole China lending program. Hence there was pressure to keep up the amount of funds allocated to the project. 11. The files record, however, that on the government's side there was a conflicting concern that "US$100 million is too large for ABC to effectively handle in a first phase project." The government was also concerned that US$200 million equivalent of counterpart funds would be needed, given the proposed 1:2 financing ratio. In the course of discussions with government on the size of the credit, Bank staff record on the files (7/83) advising senior staff of the Ministry of Finance that a credit of less than US$50 million would not be worth the Bank's processing effort. (This project was not the only one at this time experiencing upward pressure on borrowing amounts, e.g., the Bank "reluctantly accepted the limitation of US$100 million set by [govern- ment]" for the University Development II Project - file letter, February 1, 1984.) 12. In the event the credit amount was negotiated downwards to US$50 million by the Chinese authorities in mid-1983.jj To begin the process, however, files record that a senior MOF official was advised that the Bank was "thinking of a $70 million IDA credit," because IFAD's contribution was not certain and demand was there. This provoked a response from 1/ The prudent management by government of foreign borrowing, despite pressure from eager lenders, was to pay off handsomely in due course, as compared, for example, with the predicaments of several countries in Latin America (see "How Did the Asian Countries Avoid the Debt Crisis?", World Bank/PRE Working Paper, October 1991).  - 4 - government staff querying the need for any on-lending funds for ABC!V This did not divert the preparation process, however, but perhaps contributed to the comment a little later in the Decision Memorandum in November 1983 that, "The major rationale for Bank Group involve- ment is therefore to assist ABC in developing and implementing improved project appraisal techniques. In addition, Bank Group involvement will enable ABC and the technical bureaus at a provincial and local level to gain access to foreign experience and ideas in new or specialized agricultural subsectors where such experience would be particularly useful." 13. Another important issue during processing was the interest rate spread needed to give ABC/ Guangxi sufficient income to cover its costs and risks. The Bank's contribution to this important decision seems to have been a modest one, starting with the observation that ABC would need 3 percent based on the Bank's experience in other countries. In the event the decision was taken by ABC with the Bank concurring, apparently in the absence of any strong independent view. Thus the appraisal mission suggested 1.9 percent, ABC/Guangxi proposed 1.7 percent and ABC/HQ came up with 2 percent, and 2 percent was agreed. Senior Bank manage- ment queried the adequacy of 2 percent in view of ABC's processing costs for this new type of operation, but it was agreed that ABC and the Bank would review lending rates periodically in the light of relevant other factors. It had been noted earlier that ABC's administrative costs were "low by international standards, and reflect generally low salaries" (Project Brief, 10/83). 14. In an early insight but one which, based on file evidence, was not apparently to be given much attention, the Initiating Memo of October 1981 asked: "Who bears the foreign exchange risk and should subproject loans be made and repaid in currencies other than Renminbi [local currencyl?" (underlining added). This was not raised as an issue again and no record was found of any response to the possibility of ABC lending foreign currency to some subborrowers (presumably those that could bear the risk safely). The Bank's regional office has explained that at the time, the standard Bank practice was to make loans to the Ministry of Finance (MOF), with onlending by MOF in RMB (local currency) to the implementing agency. This practice was in line with the government's program for foreign exchange retention and was not negotiable at that time. Consequently, at negotiations, according to the PCR (para. 4.3), it was agreed that the Ministry of Finance would bear the foreign exchange risk. Project Objectives and Description 15. As described in the SAR (para. 4.01), the major objectives of the project were to (a) support a three-year lending program of the ABC which was part of a high priority government development effort to increase agricultural and rural production and incomes through the diversification and modernization of agricultural production and processing, and (b) assist ABC in developing and implementing improved procedures and methodology for appraising agriculture and rural development projects. Other objectives aimed at expanding rural employment and making fuller use of the ample rural labor force, alleviating rural poverty, raising the value added of agricultural production, improving production efficiency, and increasing the supply and quality of agricultural produce, including processed foods, to consumers. 2j The proposed credit funding was for development investment and not working capital subloans. In view of the scarcity of working capital funds which eventuated, however, the justification for this query is doubtful, even though the scarcity was partly associated with the unforeseen massive expansion in development investments associated with the credit (para. 36).  16. The project had the following components: (a) subloans for investments in: (i) orchard (mainly citrus) and other tree crops; (ii) culture of freshwater and marine fisheries; (iii) livestock, comprising beef/draft cattle, poultry and pigs; (iv) agroprocessing; and (b) technical assistance for the institutional strengthening of ABC. The project was therefore a fairly standard agricultural credit design, with identified subprojects and institutional strengthening through technical assistance, but also with clear diversification and modernization objectives. The modernization objective was a natural and obvious reaction to the low level of technology in the industry following several decades of relative isolation from the world economy. Total project costs were estimated to be US$142.4 million and credit funds were to be committed over a three-year period. 17. IDA funds were to be channeled through ABC which would onlend to collective units (communes, production brigades and production teams), individual farmers, households and state- owned enterprises. Although much of onlending was to be to collective units, the "production responsibility system" through "household contracts" ensures that individual families benefit from the investments in line with their own efforts according to production incentive formulae adopted by their units. The PCR notes that IDA played an important role in ABC increasing its support for individual households (even if indirectly through loans to communes and other collective units) relative to state farms and other state enterprises. 18. ABC was to onlend funds at its prevailing interest rate of 5.04 percent and determine repayment period of up to 20 years on the basis of the cash flow projections of the subprojects. Government and ABC agreed with IDA to ensure that the cost of project funds to ABC, including ABC's counterpart contributions and a 2.3 percent interest rate for IDA funds to start with, would provide ABC with the 2 percent spread needed to cover costs. 19. A major objective of the project and the main rationale for Bank involvement, since funding was relatively small for an institution of ABC's size, was the "modernization of agricultural production and processing" (SAR, paras. 1.09 and 4.01). At the time the government was particularly concerned to "upgrade technology" employed in agroindustries. One version of the project brief had referred to agricultural development policy giving high priority to, inter alia, "raising the efficiency of production through improved technology (including production inputs)," although this phrase was not carried into the SAR. Thus, although present in Bank staff thinking during preparation, technology aspects were not well defined in the SAR. 20. Procurement for ABC's subborrowers of a wide range of machinery and equipment was envisaged: for irrigation, livestock watering, and crop spraying, and hand tools, fishing gear, boats, incubators, vehicles, and a range of office equipment. However, it was agreed that procurement of inputs and equipment was to be by the local non-competitive procedures from existing designated government marketing organizations, and from local producers. Hence, although imported items were not excluded if the customary organizations had them available, no specific arrangements were made under the project to obtain modern equipment from foreign suppliers. Bulk procurement was considered impractical and project investment items were not considered  - 6 - suitable for international competitive bidding. In support of this decision the SAR quotes user preference for "locally manufactured over imported items." Customary arguments concerning standardization, spare parts availability and user familiarity were also given to support local procurement of locally manufactured items. It was believed that "agricultural inputs and other investment items required by the project would be readily available" (SAR, para. 4.15) and that there were "manufacturers producing machinery and equipment of acceptable quality and performance" (Issues Paper, para. 5, see further below on this issue). This judgement proved to be in error as reported in the PCR (Part I, para. 5.7) in connection with agroprocessing equip- ment, and also according to several comments made in the field by users and ABC staff with respect to tillage equipment and incubators. Implementation Experience 21. Introduction. The project was completed six months ahead of schedule and far exceeded the scale of operations agreed at appraisal. All sub-components were greatly expanded, especially planting of orange orchards, the largest single enterprise type, where the 28,000 ha planted were three times the appraisal plan of 9,000 ha (further details of the greatly expanded scope of the project are presented below under Results). This expansion resulted from devaluation of the Yuan by 100 percent against the US dollar such that local funds generated by the IDA Credit rose from the planned figure of Y 100 million to Y 193 million, and, with the agreed 2:1 ratio for counterpart funding, Yuan funding more than doubled from Y 255 million to Y 585 million. This unexpected turn of events is especially ironic, given the astute position of the Ministry of Finance on the size of operation that ABC/Guangxi could handle efficiently when pressing the Bank to cut the IDA Credit amount in half (paras 10 to 12 above). Even though ABC/Guangxi staff had benefitted from training and a new appraisal manual (see further below), the Ministry of Finance's judgement during preparation that a US$100 million loan (at the 1983 Yuan exchange rate of 2.0/US$1) was too big for ABC/Guangxi to handle proved correct. It is clear now, however, that in the process both the capability of ABC/Guangxi was stretched beyond safe limits and the demands on local funds rose to levels which could not be supported (PCR paras 5.4 and 5.6). IDA staff seemed to suspect a problem quite quickly, since as early as the first supervision mission in November 1984, when 43 percent of the total loan funds had been approved, IDA asked that they be reviewed to insure appraisal in accordance with procedures and criteria set out in the Appraisal Manual (para. 32). Devaluation shaped and influenced project events and problems in various ways, the full consequences of which are yet to appear (see below, para. 38). 22. As large areas of citrus were being financed, the potential marketing problem was identified during project supervision by the Bank. By October 1985, ABC had agreed that lending to orchards and tree crops would slow down. Appropriate measures were agreed between the Bank and ABC, to concentrate on other enterprises, including financing post-harvest storage and processing facilities. This resulted in the construction of 43 fruit stores at a cost of US$6.7 million equivalent, as well as 13 processing factories for fruits and vegetables. A second reaction during supervision to this great expansion of the citrus area was the suggestion to ABC that a random sample of 40 citrus subloans be surveyed to examine how this dominant enterprise was performing. Unfortunately, however, this study was not as useful as intended (see further below). 23. To keep interest rates low to farmers, ABC did not on-lend funds at its prevailing interest rate as agreed with IDA and consequently the agreed 2 percent margin to cover ABC's project costs was not achieved until 1989, over a year after the IDA credit was closed, when the rate was raised. Apart from the obvious cost advantage to farmers of low interest rates, ABC staff also  -7- noted that a principle was at stake: what was the point of taking funds from IDA at preferential interest rates if the benefits were not passed on to farmers through comparable rate discounts? 24. Project Size and Subloan Processing Problems. The effect of devaluation on implementa- tion was to signal immediately to ABC and its project staff that the project was becoming much larger than originally planned and, in particular, that the rate of subloan processing had to increase greatly if the original implementation schedule was to be met. Inflation would take care of only some of the increased funding available. ABC remained committed to the original subloan processing schedule,3] and, as far as can be seen from the files, there was no suggestion made or considered that the subloan processing period should be extended to spread the workload. Instead, the main relevant issue was the suggestion to increase the IDA disbursement share from 35 percent to 50 percent, which would have had the twofold effect of reducing both the processing pressure and the generated demand for government counterpart funds. However, no change was made on the grounds that the IDA-financing share should not be increased. (Raising the reimbursement rate was considered again under the Second Rural Credit Project, but too late to be worth making the change, although a 50 percent rate was adopted for the third project.) 25. To keep up with the required commitment rate to stay on the same schedule as before, processing was sometimes skimped and, as the PCR states, "some marginal quality subprojects" were thereby approved (PCR para. 5.4). According to ABC, skimping of subloan appraisals was most evident for subloans made to individual families, because so many such loans were made. (Hence, some ABC staff attribute part of the problem to the Bank's insistence on a large proportion of the loans going to individual families, whereas their own ABC funds are "whole- saled" through the RCC system to individual borrowers.) Moreover, and perhaps even more importantly, follow-up by ABC staff to assist subborrowers was unavoidably limited, given the expanded commitment workload and the widespread dispersion of subprojects throughout the region. Thus, a major characteristic feature of supervised credit programs was neglected. In the circumstance more than the usual amount of follow-up may have been desirable, given that working capital shortages were a problem and both the investment plan and operating costs were being affected by cost increases. Conversely, however, ABC/Guangxi's reaction to the increased processing pressure seems to suggest that there was no great concern about staffing inadequacy, since there was an increase of only 95 staff (12 percent) to the roll of 799 staff on the project in 1986. 26. On the input side, another result of the expansion of commitments was a shortage of good-quality planting material and of agricultural inputs, which was exacerbated by shortage of working capital for the purchase of inputs by subborrowers (see below). This planting material problem will show up for many years in terms of production levels in the many orchards financed. 27. Counterpart Funding and Incremental Working Capital. Shortage of government funding was a constraint throughout implementation (and remains a constraint on ABC's activities in Guangxi), resulting especially in inadequate working capital for subprojects since. In part at least (see below), this resulted from ABC/Guangxi's not having sufficient own funds available. The [/ Apparently as a matter of policy, ABC and government wanted to commit IDA funds rapidly.  -8 - amount of local funding 41 required increased, of course, because IDA funds were so much larger in local currency than originally planned. Shortages of working capital (from both subborrowers and ABC) especially hurt those many subprojects where several years passed after investment before production began, including the large area of perennial orchard crops. ABC/Guangxi funds for this purpose were constrained by, firstly, limited branch funds and, secondly, by allocations from HQ being pre-determined by a sharing formula for all provinces under the annual credit plan which could not be readily altered. Shortage of counterpart funding and working capital could have had the beneficial effect of slowing subproject processing down, but this did not apparently happen as the alternative of skimping on matching funds was preferred. 28. One of the Bank's contributions to the project's success was in the form of appraisal team case studies and here ABC staff (para. 17(b) of the PCR) note a deficiency which contributed to the lack of incremental working capital (IWC) for subborrowers. Thus in the early years of the project, ABC relied on the case studies, which, ABC says in the PCR, did not show "specific financing arrangements" for incremental working capital. The audit confirmed, on the contrary, that small working capital loans were included in the SAR tables, but that the misjudgment was over the amount of "own funds" which the subborrowers would be able to contribute. ABC staff were in a position to realize, however, that a shortage of "own funds" implied that an increase was needed in IWC financing. 29. Procurement. As already indicated, the project design, despite the project objectives, did not tackle adequately the "modernization through technology" issue. In particular, procurement arrangements did not provide for direct imports. Consequently, when ABC, before project effectiveness, asked whether bulk fertilizer supplies could be financed, the Bank rejected the request on the grounds that bulk procurement from abroad was not a provision of the legal documents. This would have given the impression that there was no possibility to use project funds for imports, which would not have been so, especially if it was confirmed that domestic items were inadequate as has now been confirmed (see PCR para. 5.7). 30. Training, the Role of EDI and Technical Assistance. As noted in the SAR, implementa- tion of the project began before the credit was approved or effective with respect to important institutional and training activities. The Bank's Economic Development Institute (EDI) played a valuable role in organizing training courses in China on rural credit for 32 ABC staff, especially on appraisal methodology. The training of these initial 32 ABC trainers was then the basis, with further Bank staff assistance, for numerous training courses during project implementation for both ABC/Guangxi staff and ABC staff from other provinces. By the end of the project EDI courses had been attended by 120 ABC staff from around the country and ABC/Guangxi staff had participated in the training of over 11,000 ABC staff from elsewhere. ABC also found Bank staff assistance valuable in training of auditors and in setting up the project's evaluation and monitor- ing system, with the help of foreign short-term consultants. 31. Funds allocated for hiring foreign consultants for training assistance were not used because ABC and government preferred to allocate IDA funds solely to investments. Also, training received in EDI courses and Bank staff contributions during appraisal resolved many 4/ Both for counterpart funds to match (2:1) IDA's capital investment funds and for the associated working capital included in subloans.  -9 - technical issues. ABC relied in addition on advice received during Bank supervision missions (although more frequent and better qualified missions were expected--see below). Other reasons noted for not engaging foreign consultants were that conditions in GZAR are varied from place to place and therefore the advice of local experts was likely to be more relevant, that there were many specialist research institutes which could provide advice when needed and previous experience suggested the quality of advice given by visiting experts was mixed. On this last point, the practicality in local conditions of the advice of visiting experts was felt to be limited on occasion. 32. Bank staff also assisted ABC to draft a subproject appraisal manual and with arrangements for study tours for ABC management staff to Bank-assisted rural credit programs in other countries, beginning with Pakistan and Mexico. The "issuance" of the "Appraisal Manual for Rural Credit Projects" was a condition of effectiveness of the IDA Credit. ABC reports that the manual is now used widely on other ABC credit operations following the distribution of 150,000 copies throughout ABC branches. ABC/Guangxi has since expanded the appraisal manual with further guidelines on World Bank processes and the results of the project, including examples of all main documents used. Also ABC has produced an instructional volume of experiences under all the donor-assisted rural credit projects, describing problems and their solutions. 33. Supervision by the Bank. Throughout the project's five-year life, despite the large-scale expansion and resulting implementation issues, performance ratings by the Bank's five supervision missions were consistently a # 1--"Problem free or minor problems," with a status of # 1-- "Improving." However, implementation problems were a little more than "minor," as both the PCR and the discussion above show. 34. The fact that the working capital problem was not reported until the fifth Bank supervi- sion mission in late 1988, at the end of the fourth project year, and then only as a result of reviewing the study of 40 subprojects, adds fuel to ABC's contention that Bank supervision was not adequate (over the main four years of implementation, supervision input exceeded the standard coefficient of about 13 staff weeks at 15.5 only in the first year, but averaged only 6.5 staff weeks thereafter--page iv). Such a shortcoming in the credit packages should have been picked up earlier by supervision staff, especially as, according to ABC, the problem was more acute during the earlier years of subloan processing. 35. Similarly the 40-farm citrus study was a good idea, but was poorly executed. Apparently the concept the IDA staff had in mind was never fully understood by ABC staff, although this communication problem could have been overcome if terms of reference for the study had been prepared and discussed. Partly in consequence, therefore, the study output was meager and, perhaps understandably, no substantive Bank response was traced on the files.5/ 5/ The phenomenon of "disappearing studies" has been noted in OED before, whereby studies are proposed and even launched with good intentions, and then disappear without trace, or, as in this case, make only a weak showing on the record at a later date, compared with their apparent importance. A study of the utility of studies in Bank-assisted operations suggests itself.  - 10- Results 36. Some 1,240 subprojects were completed, with only an insignificant number of approved subprojects being later dropped.