India - Fourth Agricultural Refinance and Development Corporation Credit Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 7925 PROJECT PERFORMANCE AUDIT REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) June 30, 1989 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank auhorization. ABREVIATIONS AND ACRONYMS ARDC - Agricultural Refinance and Development Corporation BIRD - Bankers Institute for Rural Development CAB - College of Agricultural Banking CALCOB - Committee on Agriculture Loans through Commercial Banks CBs - Commercial Banks COTELCOOP - Committee on Term Lending through Cooperatives CRAFICARD - Committee to Review Arrangements for Institutional Credit DCBs - District Cooperative Banks DOM - District Oriented Monitoring EAPD - Economic Analysis and Publications Department (of NABARD) IRDP - Integrated rural Development Programme NABARD - National Bank for Agriculture and Rural Development PACs - Primary Agriculture Cooperatives PLDBs - Primary Land Development Banks RBI - Reberve Bank of India RDPCs - Rural Development Project Courses RRBs - Regional Rural Banks SCBs - State Cooperative Banks SLDBs - State Land Development Banks THE WORLD SANK Wash"*on. D.C. 20433 U.S.A. Ofe of Obse-CGW Openesses ashsonm June 30, 1989 OM)RANDM TO THE E MCUTIVE DlRECTORS AND THE PRBSIDENT SUBJECT: Project Performar . dit ' eport on India Fourth Agricultura Refinance and Development Corporation Cr*dit Project (ARDC IV) (Loan 2095-IP/Czedit 1209-IN) Attached, for informati.on, is a copy of a report entitled "Project Performance Audit Report on India - Fourth Agricultural Refinance and Development Corporation Credit Project (ARDC IV) (Loan 2095-IN/Credit 1209-IN)" prepared by the Operations Evaluation Department. Attachment This document has a restricted distribution and may be ued by recipients only in the perrormance of their official duties. Its contents may not otherwise be disclosed without WGul Bank authorization. FOR O KIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) Table of Cortents Pate No. Preface .................................................... i Basic Data Sheet .......... ..................... ii Evaluation S-immary .......................................... iv PROJECT PERFORMANCE AUDIT MEMORANDUM I. PROJECT BACKGROUND.................... ............ 1 Introduction.................................... 1 The World Bank Group and Agricultural Credit in India .............................. 2 Performance of Earlier Bank-Assisted Agricultural Credit Projects ................. 3 Project Objectives ............................. 4 Project Formulation and Design ................. 4 Project Processing ......5.................... 5 II. PROJECT IMPLEMENTATION AND OUTCOME ............... 7 General ......................................... 7 Achievements ..............7.................. 7 Implementation ............................... 9 III. FINDINGS AND ISSUES .............................. 10 Introduction ................................... 10 The Rural Financial System . .................. 11 The Integrated Rural Development Programme ..... 12 Bank Perfoi-mance ............................... 14 ATTACHMENT I: Comments from the Overseas Development Adminis- tration of the U.K. Government ................. 19 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. PROJECT PERFORMANCE AUDIT REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) PREFACE 1. This is a Project Performance Audit Report (PPAR) on the Fourth Agricultural Refinance and Development Cooperation Credit Project (ARDC IV), involving a loan and credit in the total amount of US$350 million, with the objective of increasing small-scale agricultural production and strengthening agricultural credit institutions. The loan and credit were approved on February 23, 1982, and became effective on May 25, 1982. The loan and credit were closed as scheduled on June 30, 1984. The date of final disbursement was November 20, 1984 when the loan and credit were fully disbursed. 2. The PPAR is based on a Project Completion Report (PCR) issued in June 1989, prepared by the Government's National Bank for Agricultural and Rural Development (NABARD) (with an Overview of the PCR prepared by the Bank's Asia regional office), the staff appraisal report (3629-IN, dated February 1, 1982), the president's report (P-3196-IN, dated February 1, 1982) and the loan, credit and project agreements dated February 24, 1982. Correspondence with the borrower and internal Bank memoranda on project issues as contained in relevant Bank files have been consulted. Bank staff associated with the project have been interviewed. An OED mission visited India in March 1988. Discussions were held at NABARD headquarters in Bombay and visits were made to participating banks and farms in Andhra Pradesh State. The audit acknowledges with gratitude the assistance provided by Government officials, staff from NABARD and participating banks, and farmers who provided information and hospitality to the mission. 3. The PCR provides a satisfactory account and assessment of the project experience, and discusses the performance of the Bank and the project executing agencies. The PPAR agrees with the principal conclusions of the PCR Overview, which is more analytical and critical than the PCR itself. In addition the PPAR elaborates on particular aspects such as (a) the broader rural financial sector context of the Bank's agricultural credit operations in India, (b) the financing of IRDP through this general line of credit and (c) the performance of the Bank, especially the staff resources devoted to the work program and marpgement follow-up. 4. Following standard procedures a copy of the draft audit report was sent to the Borrower and the (verseas Development Administration (ODA) of the U.K. Government on March 27, 1989. No comments were received from the Borrower. Comments received from ODA are included as an Attachment to the PPAR. At the request of the Reserve Bank of India copies of the draft report were also sent to members Gf the Senior Experts Group (SEG) of the Agricultural Credit Review Committee of the Reserve Bank. - II - PROJECT PERFORMANCE AUDTREPORT IDIA FOURTH AOPICULTURAL REFWNANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (;.OAN 2eN5-INICRDIT 1210-IN BASIC DATAINE Key Project Data Appreleal Actual or Actoel as X of Estimate [dimeted Actual Aptralsal Estimate Total Project Cost (USS MIeion) 2,866.2 - Total Project Cost (Rs M) 16,696 - / Eligible for IDA/150 Financing (US M) 1,317.4 Eligible for ID.W/IRD Financing (Re M) 10,54 12,815 119.7 Credit Amount (SDR M) 139.0 139.0 166 Date of Board Approval 2/28/82 Date of Effectiveness 6/25/62 Loan Amount (USS Million) 196.6 19.0 Igo Date of Effectiveness 5/25/82 Date of Credit/Loan Agreement 2/24/62 Closing Date 6/3/84 6/38/84 Economic Rates of Return (N) 6/ aes-5*9 Negative to 654 Financial Rates of Return (S) h/ 20-481 1o-61I Number of Direct Bneficiarles 1.4 million 2.2 million 157 STAFF INPUTS (staff weeks) FY76 FY76 FY77 FY78 FY79 FY39 FY81 FY82 FY63 FY84 FY65 FY8 FY87 FY62 TOTAL Preappraisal .1 1.2 5.7 19.1 26.2 Appraisal 55.8 98.8 92.6 Negotiations 22.6 22.8 Supervision 6.0 29.4 24.6 .3 6.0 .6 66.8 Other .1 .6 2.3 3.2 TOTAL .1 .1 .0 1.2 .0 5.7 75.7 67.9 29.4 24.6 .3 .6 6.6 .6 211.6 CUMULATIVE DISBURSEMENTS FY82 FY83 FY84 FY6 Appraisal Estimate (USS Million) 46 190.0 368.4 36.0g Actual (USS Million) 40.6 156.8 340.6 346.9 Actual as X of Estimate 161.6 82.5 97.3 97.4 Date of Final Disbursement of Loan 11/20/3-, Date of Final Disbursement of Credit 6/26/13 a/ The ARDC IV project was described as "a two-year time-slice" of ARDC's ongoing lending program (para 4.30 of SAR). Total project costs (estimated at Ra 16,690 million) were based on aggregate costs of all field-level investments supported by ARDC refinance including equity contributions of borrowers and portions firanced by other banks. Actual total costs of such investments comparable to SAR estimates, are not available at ARDC; however total ARDC refinance during FY83 and FY84 amounted to RS 15,950 million against SAR estimates of Rs 11,887 million. Project costs do not include large-scale investment subsidies provided by central and state governments (for example, 25% to 33% of capital costs for IRDP beneficiaries--SAR para. 5.15). b/ Based on samples of typical on-farm investments. c/ Both Loan and Credit were fully disbursed. The apparent under-disbursement is due to conversion of SDRs to US$ equivalent. No. of Naadays Spacialisations Perfeorsmse Types of Date Pereces in Field teaseted 3/ Ratios S/ Trend Problem 4 Preparatis 11/79 1 3 A VA NA sh Review of Project 6-7/81 2 10 N,1 MA NA NA .)Prepeaties Appraisal 4-5/81 7 1% A,8,C,D,FC MA NA BA Subtotal W Supervision 1 2/82 2 7 A,5 1 2 N,0 supervision 2 56/82 3 35 A,B 1 2 N,0 Supervision 3 10/82 2 24 B,D 1 2 N,0 SuperviWon.4 2-3/83 2 35 3,C 2 2 N,0 Supervision 5 5-6/83 2 20 B,C 2 3 N,0 Supervision 6 10/83 2 13 a 3 3 M,0 Supervision 7 )/64 3 16 8,r 3 3 N,0 Supervision 8 3-4/84 2 8 a, 3 3 N,0 Subtotal 165 Total 369 OTHER PROJECT DATA Borrowers Government of India Baecuting Agency: Agricultural R nanceand Development Corporatin (now NARARD) Fiscal Year of the Borrove: July 1 - June 30 Name of Currency Rupee () Currency fachange Rate: Appraisal Report US$1.00 - 8.0 Appraisal Year Average US$1.00 Rs 10.2 Intervening Years Average U331.00 Rs 10.5 . Completion Year Average US$1.00 Rs 11.6 Follow-on Projects Meme ABA&D--I Credit Project Loan Number Loan 2635-IN Loan Amount (US$ million) 375.00 Date of Board Approral 2/25/86 3/ Specializations represented: A a Agriculturalist; 8 Agricultural Credit Specialist; C a Financial Analyst; D - Loan Officer; E a Fisheries Specialist; F a Economics; G - Irrigation Spoialist; U a Cooperative Specialist. 4/ Performance Rating: I 1 Proble0-free or minor problems; 2 Moderate problem 3 U Sajor problems. 5/ Trand- U Improving;. 2 a Stationary; 3 - Deteriorating. 6/ Types of Problems: F Financial; - ManaterialT; B Tachnical; P = Political s 0 a Other. - iv - PROJECT PERFORMANCE AUDIT REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) EVALUATION SUMMARY Introduction 1. The Fourth Agricultural Refinance and Development Corpotation Credit Project (ARDC IV) was approved on February 24, 1982 and was one of a long series of Indian agricultural credit projects assisted by the Bank Group since 1970. This was the fourth national project, following a series of ten agricultural credit projects for individual states, all through ARDC, which in turn was followed in 1986 by a further national project- -NABARD I. The project became effective in May 1982 and was closed as scheduled on June 30, 1984. As expected at appraisal, ARDC was absorbed by a successor organization, the National Bank for Agriculture and Rural Dev2lopment (NABARD) in July 1982 but this had little effect on project implementation. 2. The performance of the earlier projects has been described in OED reports as in general highly satisfactory, except that some investments were not realizing expected benefius and poor credit administration, inter ali&, had led to a serious and deteriorating overdues situation in partici- pating banks. 3. The project cost of US$2.1 billion, was supported by a Bank loan of US$190 million and an IDA credit of SDR 139.0 million (a total of US$350.0 million or 17% of the cost). At the time this was the largest operation approved by the Bank. Parallel financing of ARDC's lending program was also expected from four other donors but details had not been finalized. Funds were channeled through ARDC to Participating Banks (PBs), primarily State Land Development Banks (SLDBs) and Commercial Banks (CBs), which made loans directly to farmers. Objectives 4. In common with the three earlier ARDC projects, the project was to finance a two-year time-slice of ARDC's lending operations for calendar years 1982 and 1983. The two main objectives were to increase small-scale agricultural production and to strengthen agricultural credit institutions. As before, lending was mainly for minor irrigation, land development, soil conservation and land reclamation, livestock, plantation and horticultural enterprises, and fisheries. In addition marketing and storage schemes were eligible for Bank funding for the first time. ARDC financed various other categories of lending from other resources. -v - 5. About 1.4 million farmers throughout India were expected to bene- fit directly from the projett and not less than 551 of the credit and loan funds were earmarked for loans to small farmers. About half of lending was targetted to disadvantaged areas. Implementation and Results 6. Throughout project processing the high level of overdues (ap- proaching 50Z) in the PBs and measures to improve the situation were the main issues. Two measures were agreed: revised criteria (relating mainly to credit recoveries) would determine the eligibility of PB9 for ARDC refinance, and a program to rehabilitate the weakest SLDBs. Funds for staff training were also provided. In the event the eligibility criteria were too loose to provide the necessary incentive to improve credit recoveries and in any case were little applied. Furthermore state governments and SLDBs were reluctant to implement the agreed rehabilitation program. Coasequently, despite agreed and widely applied accounting procedures which reduced actual overdues, over the two project years the overduas reported by PBs remained static at 47% for CBs, or rosa to 60% for SLDBs and to 68% for cooperative banks. Institutionally therefore the project performed poorly. 7. In terms of coverage and refinancing volumes, however, the project performed as planned. Data presented in the PCR show that all targets were met or exceeded, including tLat disbursements to small farmers were around 77% (cf. 55% targeted at appraisal), disbursements to disadvantaged areas were as planned about half of total lending, beneficiaries were estimated at 2.2 million (1.4 million), employment creation was estimated at 411 million mandays (134 mmd) and incremental production was expccted to be double the appraisal estimate. The main categor.es of activity financed were minor irrigation at 42% of total lending, and dairying and bullock purchase (for draught) each at close to 20%. Ex-post evaluation studies of a sample of credit schemes estimated financial and economic rates of return at very satisfactory levels, except for public tubewell investments. A major unforeseen feature was that lending for activities under the government's Integrated Rural Development Programme (IRDP) Zor poverty alleviation rose sharply to almost half of total lending. The audit shares th6 doubts about the effectiveness or IRDP as expressed in the Overview and other Bank papers. 8. Rates of return presented in the PCR for a sample of schemes are judged by the Overview to suggest that, with the exception of public tubewells, investments in such schemes are financially and economically viable. The Overview states, however, that NABARD overestimated project impact because poor quality investments and diversion of funds were not taken into account. The audit, furthermore, notes that some of the sample rates of return presented in the PCR raise questions about the consistency of the methodology used and also, like the Overview, raises questions about the effectiveness of the very large disbursements under IRDP for livestock purchase. 9. Towards the end of the project, continued poor recovery perform- ance led to a new feature being introduced in the form of a pilot credit administration scheme for three districts. This was designed to attack - vi - what was perceived by ARDC and the Bank as the root cause of high overdues and has been pursued further under the follow-up project--NAARD I. Findings and Issues 10. Overall, the audit concludes that while there were undoubtedly saccessful investments financed in the rural oreas by project funds, and hundreds of thousands of poor farmers (and others) clearly benefitted, the general impact of the project on agricultural credit institutions and the agricultural credit system in India was largely negative. By making large external resources available in the short run to the syster., without en- suring that the system's serious problems received appropriate attention, the Bank failed to contribute to developing the credit system in the long run. The sustainability of the institutional credit system at reasonable cost was not enhanced by the Bank's participation in the project and in- deed, in the opinion of some knowledgeable Bank staff, the Bank's parti- cipation may have even weakened the system. It is doubtful whether meeting the short-run disbursement targets of the Bank and government's need for foreign exchange during this two-year period, through an operation such es the project, was adequate justification for conti.iuing to support a dete- riorating inscitutional credit system in India. 11. The audit memorandum does not repeat all the detailed points which are adequately discussed in the PCR and Overview, but highlights three issues of important general significance. These are, (a) the broader rural financial sector context of the Bank's agricultural credit operations in India, (b) the financing of IRDP through this general line of credit and (c) the performance of the Bank, especially the staff resources devoted to the work program and management follow-up. 12. The audit suggests tnat a broader approach to the rutal financial sector might have better addressed some of the issues which detracted from the project's performance, although this was not the Bank's customary ap- proach to such projects at the time. 13. The IRDP issue raised by the audit is whether Bank financing of such a program on such a large scale was appropriate without first engaging in a policy and program strategy discussion to resolve some of the wider issues connected with such a poverty alleviation program. 14. Lastly, the audit reviews the Bank's performance in supervising the project, especially given that the nation-wide project approach for agriculture credit adopted by the Bonk in 1975 was specifically designed to reduce Bank staff inputs and the inadequacy of supervision became an issue with senior management during ARDC IV as an explanation of poor institu- tional performance under t.he project. The audit concludes, however, that the poor progress on institutional issues during ARDC IV, and especially the overdues question, canno, be attributed mainly to inadequate super- vision, but is in part explained by the important resource transfer func- tion of the Indian agricultural credit series of projects. A reluctance on the part of staff, including management, and the Board, to unduly disturb the Bank's agricultural credit operations in India goes some way to explaining lack of progress on institutional issues while disbursements continued to flow according to plan. PROJECT PERFORMANCE AUDIT REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) I. PROJECT BACKGROUND A. Introduction 1. The Fourth Agricultural Refinance and Development Corporation Credit Project (ARDC IV) was approved on February 23, 1982 and was the fourth in a series of nationwide agricultural credit projects assisted by the Bank Group in India since 1975. These national projects followed a series of ten state-specific agricultural credit projects, also through ARDC, dating back to 1970 and were followed in 1986 by a further national project--the National Bank for Agriculture and Rural Development (NABARD I). 2. The ARDC IV project, which was expected to cost US$2.1 billion, was supported by a Bank loan of US$190 million and an IDA Credit of SDR 139.0 million (a total of about US$350.0 million, or 17% of the total cost). At the time (February 1932) this was the largest operation approved by the Bank Group and went ahead at a critical time of IDA funds shortage only because the Bank was able to increase the IBRD lending program. Parallel financing was also expected from three bilateral donors (UK, CanEda, Netherlands) and the European Economic Community. 3. Funds were channeled through ARDC/NABARD 1/ to participating banks (PBs) which made loans directly to farmers.2/ Large-scale "wholesaling" of credit by ARDC is facilitated by "schematic lending" whereby participating banks submitted costed credit schemes to ARDC for prior approval. Such schemes are for lending in specified areas and for limited purposes (mostly irrigation development). The other main and growing category of ARDC lending has been refinancing the loans made by Regional Rural Banks (RRB) under government's chief vehicle for poverty alleviation in rural areas--the Integrated Rural Development Programme (IRDP). 1/ It was known at appraisal that ARDC was soon to be absorbed by a proposed new and larger body, the National Bank for Agriculture and Rural Development (NABARD). 2/ ARDC/NABARD ref 'nances the lending of four types of banks: (i) State (cooperative) Land Development Banks--SLDBs; Commercial Banks--CBs; (iii) Regional Rural Banks--RRBs, which are specifically for lending to the poor; and (iv) State/Apex Cooperative Banks (and their branches)--cooperative banks; SLDBs and CBs were dominant at the time of appraisal, taking 98% of ARDC's funds, with CBs, and their young offshoot RRBs, growing in relative importance. -2- B. The World Bank Group And Asricultural Credit in India 4. Institutional agricultural credit in India dates back to the establishment of the credit cooperatives in the early 1900s. More recently the rural banking system has been expanded under the guidance and :egu- lation of the Reserve Bank of India (RBI) with the addition of RRBs and agricultural credit operations by the commercial banks. According to RBI survey data quoted in the Staff Appraisal Report (SAR) this expansion of rural institutional credit sources has achieved th&e primary objective of gradually reducing the share of outstanding debts owed to private money lenders. 5. The Agricultural Refinance Corporation (ARC) was established by the government in 1963 as a subsidiary of RBI and was later renamed the Agricultural Refinance and Development Corporation (ARDC). 6. The Bank Group has been involved in agricultural credit in India since 1969. Between 1970 and 1973 ten state-specific agricultural credit projects were approved and covered the country's major farming regions. All of these projects were completed between 1975 and 1978. Total funding was US$320 million, all channelled through ARDC. 7. Since the ten projects were approved in less than four years, and all but two were not completed until 1977-78, there was a period of over two years (from October 1973 to June 1975) when the complete set of ten projects was under supervision. This proved a great burden on the staff resources of the Bank Group, as well as on Government and APDC resources. In addition ARDC had taken an active role in the formulation and imple- mentation of 25 Bank Group assisted projects that provided agricultural credit to specific agricultural sub-sectors, mainly irrigation. 8. From 1975 onwards therefore, the Bank, after careful conside- ration, ceased lending for state-specific projects and initiated what was to become a series of nationwide agricultural credit projects, once again through ARDC and later through NABARD. Each of the first four country-wide operations through ARDC was a two-year time-slice of ARDC's refinancing program with total Bank Group loans and credits of US$875 million. (At the time of ARDC IV, total Bank Group commitments for the 39 projects approved between 1970 and 1982 in which ARDC was involved were US$2.4 billion.) 9. Thus 1975 is the watershed of the Bank's relationships with Indian agricultural credit with the Bank quite consciously and purposefully decia- ing that, for mainly staff resource reasons, it could no longer participate directly in field supervision of agricultural credit operations throughout India. It was decided that Bank supervision henceforth would be restricted to overriding matters of policy and institutional development, largely at the ARDC level, with supervision and monitoring of participating banks and agricultural credit operations themselves being left to ARDC and the par- ticipating banks. (Although the Bank's longer term management of its large agricultural credit operations in India is a broader subject than this audit report addresses, during ARDC IV the inadequacy of Bank supervision was identified by senior management as one of the reasons for the poor institutional results being achieved through Bank lending. This point is discussed further below.) - 3 - C. Performance of Earlier Bank-Assisted Atricultuxal Credit Prolects 10. The Operations Evaluation Department (OED) issued a report in 1981 reviewing the experience with the first nine of the ten state-specific projects, based largely on the Project Performance Audit Report (PPAR) of each project. The review concluded that, "Experience with these projects was on the whole highly satisfactory", and that "this set of projects was timely, generally received strong institutional support, and achieved an impressive measure of success". The major qualification was that, "Overdues have been the most critical problem of participating banks, both LDBs and CBs. Causes are many: deficiencies in lending procedures, effects of inclement weather, inadequate short-term financing available to farmers and political pressure in various forms. States where loan recovery has been good show strong bank management, state government support of loan recovery efforts, and minimal political interference in banking operation." 11. More recently, the PPAR of the Third Agricultural Refinance and Development Corporation Credit Project (Cr. 947-IN) concluded as follows: "Thp project was implemented largely as planned and the IDA Credit wa. fully disbursed before the Closing Date. Actual expenditure under the project was about 10% higher than forecast at appraisal and most of the lending targets were achieved, including those for lending to small-scale farmers and less developed states. Loans were provided to about one million farmers, compared with a target of 1.1 million farmers set at appraisal. However, not all of these farmers derived worthwhile benefits from the project, for some of the investments were unsuccessful, while completion of others was delayed. Serious problems have also been encountered with credit administration, especially with the very high incidence of overdue loan repayments." 12. Other main conclusions of this audit included: - that the complexity and large-scale of the project made it not feasible for IDA to provide effective supervision; - although all credit institutions kept their accounts up to date, little data was available on physicel and financial performance and zhere had been no progress with computeriz- ation; - high overdues were caused by ireffective follow-up by lending institutions, willful default and the difficulty of banks' foreclosing on land; and - loans provided under the Integrated Rural Development Program (mostly for livestock) were relatively unsuccessful because of haphazard credit administration and preoccupation with meeting lending targets. This PPAR was issued in September 1985, however, some 15 months after ARDC IV had closed, thus illustrating discontinuity in lesson learning (at least through formal evaluation channels) that resulted from the rapid succession of ARDC projects. D. Prolect Obiectives 13. The objectives of the ARDC IV project in the context of the government's country and secto: objective are described in the staff appraisal report as follows (para. 4.03): 'Main objectives would be to help increase agricultural production, rural incomes and employment; give continuing emphasis to assisting small farmers and less developed areas; continue institution building through further strengthening of ARDC, improving CB agricultural lending and giving particular emphasis to rehabilitation and strengthening of LDB; and continue programs for improving quality of investments. GOI regards ARDC lending as a key program for its agricultural development plans, and for implementing national policies on Integrated Rural Development, designed to improve the income of the rural poor." Thus in short, the project had two main objectives: increased small-scale agricultural production and improving the institutions providing formal agricultural credit. By more recent standards there was a notable absence of reference to the broader aspects of rural financial sector development. Much of the institutional emphasis was on arresting the deteriorating overdues situation, rather than on developing the rural credit system more generally. 14. About 1.4 million farmers throughout India were expected to benefit directly from the project and about 60% of the beneficiaries were to be small farmers with annual per capita incomes of less than US$114 (the project thus qualified as "rural development" or "poverty-oriented" for the Bank's portfolio monitoring purposes). Not less than 55% of the credit and loan funds wetz earmarked for loans to small farmers. It was also intended that by the end of the project period atout 20% of ARDC's leading would be for the disadvantaged areas of the northeastern states and that eventually half of lending would go to disadvantaged areas, including those in the northeast. 15. Institutional objectives included training programs for 22,000 staff members of participating banks; further strengthening of ARDC to improve the quality of lending; rehabilitation and strengthening of weaker SLDBs, and the continued application of eligibility criteria (as revised and agreed) for PBs to exclude funding those with poor loan collection performance until they met certain minimal standards. E. Project Formulation and Design 16. The fourth ARDC project departed little from the approach esta- blished by the first three nation-wide operations through ARDC. The project thus financed a two-year time-slice of ARDC's lending cperations (calendar years 1982 and 1983) for the following activities: minor - 5 - irrigation (dugwells, dlg-cum-bore wells, shallow bores and filter points, improvement of existing wells and provision of pumpsets); land development (land shaping and levelling, and iield distribution and drrinage channels); soil conservation and land reclamation; livestock (mainly dairy cows, but also abattoirs, and draft animals and carts); plantation and horticultural enterprises, and fisheries. The lending program included support for the Integrated Rural Development Program (IRDP) through the refinancing by ARDC of loans made for this purpose by RRBs. In addition, marketing and storage schemes became eligible for reimbursement by the Bank for the first time. In parallel with these types of loans, ARDC was to finance, from other resources, loans for forestry, farm mechanization, pumpset electrification and various other categories of lending. Excluding these categories was partly influenced by the need not to crowd out other external sources of funds, but also because of doubts about the social impact of farm mecha- nization. 17. The disbursement categories of the Loan and Credit allocated 77% of the Bank Group funds to minor irrigation, 22% to other productive activities and the balance of less than 1% to training. Reimbursement by the Bank under the three previous ARDC projects had been at the rate of 50%, but was reduced to 34% under ARDC IV to preserve the two-year time slice in the absence of additional Bank funds.3/ The Bank Group funds were thus intended to be absorbed by ARDC's planned expansion of its lending program (for schematic loans) in the Sixth Five-year Plan of 15% and 13% during the two project years, although the SAR was not specific in demonstrating ARDC's need for the proposed level of Bank Group funding. 18. With respect to the impact on the project of the establishment of NABARD, the SAR stated that: "Because integration of activities of NABARD's constituent institutions would take time, AROC would most likely continue to operate as at present over most of the pioject period." (SAR para. 5.06.) Even so the Bank took a close interest in NABARD's esta- blishment through a legal covenant referring to consultations on financial and organizational plans (SAR para. 5.09).4/ F. Project Processing 19. The project was appraised by the Bank in April/May 1981, and was approved by the Board in February 1982. The dominant issue throughout project processing was the high level of credit overdues and measures to irprove the situation. These were mainly the incentive eligibility cri- teria for participating banks and the rehabilitation program for weak 3/ In the event when Bank Group disbursements lagged behind schedule because of slower than expected demand for minor irrigation credit and rupee devaluation, as had been agreed at negotiations the reimbursement percentage was returned to 50%. 4/ The status of NABARD at the time of appraisal may be summarized as follows from the SAR (para. 5.07): "A draft legislative bill proposing the establishment of NABARD, has been tabled in the Indian Parliament. GOI officials project that NABARD may be established by mid 1982. The new institution would be formed by an amalgamation of ARDC and the Agricultural Credit Department of RBI." - 6 - "LDs. (The use of eligibility criteria, however, dated to befote the Baik Group became involved in agricultural credit in India.) After intensive review and prolonged discussions with G0/ARDC during project formulation, appraisal and negotiations, including an analysis of the impact of existing criteria on loan recoveries and lending operations of SLDB branches, revised eligibility criteria were agreed under the project. As an incentive for improving loan collections, these criteria generally required that annual loan recoveries in SLDBs, together with state government con- tributions, should be at least 752 of collectiblis (demand less blocked accounts) in order for SLDBa and their branches to qualify for ARDC refinancing. 20. Except in a small number of states, repayments by farmers to PBs had always been poor, although the overdue problem had little if any impact on ARDC since the PBs normally repaid ARDC on schedule, while losses at the state level were made up by state governments. The seriousness and conti- nued deterioration of the overdues situation is illustrated by data quoted during the Board discussion. The overdues of the SLDBs had increased from 282 of loans in June 1974 to 44% in June 1981, while the condition of commercial banks was even worse with overdues on agricultural loans being between 47 and 50% for the previous five years. The SAR noted that "Willful default and unfavorable political conditions for loan collection are the major factors causing high overdues". The SAR especially attri- buted the recent (at that time) deterioration in loan collection perform- ance to the severe drought in 1979 coupled with the several national and state electimns that occurred during that period, siice it was the practice for political parties and candidates to curry favor with lectorates by actual or promised forgiveness of debts. With respect to climatic factors it was significant that since lending discounted with ARDC was mainly for minor irrigation there was a correlation between good loan collection performance by PBs and the existence of reliable ground water resources in the areas served by such banks. 21. With respect to improving agricultural credit operations, the project .%rticularly emphasized rehabilitation and strengthening of the poorer performing SLDBs (the rehabilitation effort had begun under the previous project, ARDC III, but the application of the linked eligibility criteria, which would have excluded two SLDBs, was deferred under that project in agreement with the Bank). The Government undertook under ARDC IV that in addition to financial measures, such as rephasing of overdues and setting aside chronic overdues in "blocked" accounts, a Standing Committee was to be established to guide and supervise the reha- bilitation of six identified SLDBs which were especially weak (overdues of 70% or more), plus any other weak SLDBs in due course. Organization and management studies were to be carried out for these SLDBs, and timebound rehabilitation programs, to be approved by state governments, were to be implemented. 22. The staff appraisal report was commended by a central project staff review for its "extremely high standard of disclosure" and that great care was taken by the drafters to be objective. The reviewer, however, also recorded a reservation that "although the project appraisal methodo- logy ane emphasis are generally consistent with established practice related to credit projects within the Bank, this current practice does not . 7 - fully address the issue of the financial health of the rural credit system". 23. Despite the arrangements to Improve credit operationa, at the Board presentation a major concern remained the continuing problem of high and rising overdues. Bank staff responded that this project was very much a turning point with ARDC, and that either the trend would be reversed or it va unlikely that a further project in the series would be presented to the Board. (In the event, a follow-on project, MABARD I, Loan 2653-IN, wa approved in 1986 for a credit of US$375.0 million, although the overdues situation had not improved. Paragraph 29 below comments -n the trends during AMDC IV's life. This follow-on project was approved on the understanding that the government would conduct a major study of the agricultural credit system and that stricter eligibility criteria would be introduced.) II. PROJECT IMPLEMENTATION AND OUTCOME A. General 24. The project became effective in May 1982 and was closed as scheduled on June 30, 1984, with the last disbursement being made in November 1984. The PCR notes that NABARD took over the functions of ARDC in July 1982, but that for project purposes NABARD's operations were the same as ARDC's. B. Achievements 25. According to the PCR prepared by NABARD, as confirmed by the Region's Overview, most of the quantified targets in the SAR were met or exceeded. Details are summarized belows SAR Tariet Actual Achieve- Re. m. 2 Re.m. 2 went % Investments Minor Irrigation & Land Dev. 7,950 73 5,317 42 67 Non-IRDP 4,165 33 IRDP 1,152 9 Diversified 2,552 27 7,268 58 285 Non-IRDP 2,228 18 IRDP 5,040 40 TOTAL 10,500 100 12,585 100 120 (Sub-total IRDP 6,192 49) Other Landint Targets Lending to small farmers 6,012 55 9,690 77 140 Lending to disadvantaged areas 5,465 50 6,110 49 100 Other Targets Beneficiary Farmers (No.) 1.4 m. 2.2 m. 157 Incremental production (Ra.m.) 5,200 10,400 200 Employment creation (million mandays) 134 411 307 Training (No. of staff) 22,000 5,911 27 - 8 - Notable features of these achievements are that lending under IRDP expanded faster than expected, mainly for livestock under the diversified category. Overall minor irrigation was the single largest type of investment at 42% of the total, followed by dairying at 20% of all lending and 35% of diver- sified lending. (Lending for dairying and bullocks (and buiiock carts) totalled 83% of all IRDP lending.) Lending to disadvantaged areas and to small farmers was on target or exceeded the target respectively. Estimates in the PCR of beneficiaries, incremental production and employment creation suggest that each of these targets was substantially exceeded. The major shortfall was that the ARDC training program did not meet its target (only 27% of the SAR plan was achieved). 26. Financial and economic rates of return from ten ex-post evaluation studies of typical individual credit schemes are listed in the PCR. The Overview suggests that with the exception of public tubewells, these re- sults show that investments in such schemes are financially and economi- cally viable. For successful investments FRRs ranged from 15% to above 50% (20-48% at appraisal), and re-estimated ERRs were 29% to above 50% (32% to over 50% at appraisal). The Overview states, however, that NABARD over- estimated project impact (when the results of such small samples are aggre- gated) because incomplete or infructuous investments and diversion of funds are not taken into account. 27. The audit notes further that results presented in Annex V of the PCR (taken together with a number of scheme impact reports seen) cast dcubt on the methodology or presentation used in some of these rate of return estimates. Thus, for example: (a) a minor irrigation scheme in Bihar State with costs of Rs. 6,500 (per acre) has incremental income of only Rs. 324/acre, indicating a rate of return much below the 50% shown In the PCR; (b) the three dairy schemes for which data is shown have inconsistent data- -the one for Punjab has the highest milk yield (by over 40%), yet lowest net income (by over 20%), but the highest rate of return; and (c) the two electrical and diesel powered pump schemes in Utter Pradesh State have different investment costs (diesel is a third lower) and net incremental incomes (diesel is half of electric), yet rates of return are shown as exactly the same figures in all cases. (These estimates, however, do not suffer from a frequent drawback of some rates of return estimates in the past for agricultural credit schemes, namely attributing benefits to only the credit part of a package of investments). Also, for all schemes, different economic rates of return are shown in the PCR for 'with' and 'without' subsidy cases, which is theoretically inappropriate. The audit notes, even so, that high demand for credit indicates good returns to farmers--although this is linked in many cases to the subsidies available and the "increasing) expectation of not having to repay credit. On balance, however, the audit concludes that rates of return are most likely above the opportunity cost of capital for the sample activities (other than public tubewells), but not as high as presented in the PCR and Overview. 28. Institutionally the project performed poorly. Apart from training falling below target, there was little progress in rehabilitating weak SLDBs, eligibility criteria were hardly employed for the purpose intended and the overdues situation did not improve. Details are well presented in the PCR Overview. Only the pilot credit administration scheme (which was not in the project design--see below) showed some promising commitment and implementation towards the end of the project. - 9 - 29. Trends in recovery performance for the period are shown in extensive data in Annex 1 of the PCR, remembering that such data reflect the position after accounting adjustments, such as blocking, have been made and (subsidy) payments from state governments have been deducted from overdue amounts. These data show that, compared with the pre-project year (1981), reported overdues status (a) of CBs--remained static at 47%, (b) of LDBs--remained static at around 402; and of state cooperative banks--rose from 552 to 68% by 1983 (and 712 in 1984). Thus, far from reversing the trend (and not forgetting accounting adjustments), even the apparent overdues situation showed no improvement or deteriorated during the project. (Reviewers of this report have noted that the overdues issue is an extraordinary complex problem. Part of the problem, they suggest, may be technical, relating inter alia to accounting procedures and repayment dates vis-a-vis crop cycles.) C. Implementation 30. NABARD appraised schemes submitted by PBs and undertook monitoring and evaluation reviews and field studies. District-oriented monitoring studies fell short of targets set for 1982 (95 out of 135 projected--70Z) and 1983 (87 of 162 projected--54%), due to staff constraints. The quality of these reports, however, was generally good. NABARD offered financial assistance to SLDBs for setting up technical monitoring and evaluation cells, but this was not taken up. Some 50 ex-post evaluating studies of completed schemes were undertaken by NABARD as planned. 31. The new eligibility criteria became effective from January 1982 and were liberal enough to permit fresh lending by all LDBs and their branches, even those with minimal loan recoveries. Meeting the criteria in some cases was aided by the agreed linked measure segregating chronic overdues into blocked accounts. This procedure automatically increased "apparent" recovery rates as "demand" was reduced by the amount of overdues which had been transferred to blocked accounts. 32. Plans were drawn up by NABARD for the rehabilitation of the six identified weak SLDBs, as well as nine other LDBs, and submitted to the Standing Committee and to state governments. The plans were quickly approved by the Standing Committee. The eligibility criteria sanction was then applied (for the only time and for only six weeks) in late 1982 by NABARD withholding refinance until the LDBs and state governments confirmed their acceptance of the rehabilitation programs. In the event, however, LDBs and state governments were slow to implement the plans and NABARD did not exercise its right to withhold refinancing according to the agreed eligibility criteria. Furthermore, the "blocking" of overdue loan accounts was used as an accounting ploy to retain eligibility. Consequently the Overview recorded that, "Although 3ank/IDA supervision missions devoted considerable time to following the ARDC IV provisions regarding rehabili- tation, virtually no improvement was reported in the recovery situation and in the overall financial health of the LDBs targetted for rehabilitation". Any reported, but temporary, recovery improvements were more the effect of rescheduling and "blocking" than of borrowers reducing their arrears. This possibility was foreseen much earlier and accurately by a concerned regional manager in comments on the post-appraisal Issues Paper when he - 10 - wrote tha., "the financial rehabilitation measures proposed are not rehabilitation but a change in accounting methods". (The apparent reasons for the Bank not pursuing these issues--with the measures available to it under the project--are discussed further below.) 33. Commercial banks, state cooperative banks and primary cooperative societies were also subject to eligibility criteria, although for commer- cial banks these criteria were not introduced until July 1983. Refinancing of CB disbursements, however, was not greatly restricted by the criteria, since they provided for unlimited refinancing where recoveries were not less than 65%, there were some loopholes and some categories of new branches and overdues were exempted from the requirement. At the time that these criteria were introduced the Bank had proposed strengthening some provisions but these changes were not accepted by the government. 34. Partly in reaction to this poor progress, however, a new initia- tive was eventually developed to attack what was seen as at the root of the overdues situation--the poor quality of lending, compounded by inadequate follow-up supervision, leading to low incremental income. At the beginning of 1984, in consultation with the Bank, NABARD sponsored (as an addition to agreed project activities) a pilot project covering all PBs in three dis- tricts. The objectives were to strengthen credit delivery and improve recoveries through intensified field staff training, structured work pro- grams for field and supervisory staff and improved systems and procedures. Results of this initiative were not reported in the PCR, but this new approach was pursued in the next project in the series--NABARD I. III. FINDINGS AND ISSUES A. Introduction 35. It is difficult to restrict the audit's findings to project- specific matters in the strict context of this audit of only a two-year time-slice of funding. Most of the main issues have to be seen in the broader light of a long-term financial relationship between the Bank Group and the very large, complex and diverse Indian formal agricultural credit system. In any case the PCR, and especially the thorough and frank Overview by the Region, adequately detail the project's achievements and the more general operational issues arising (inadequate credit supervision and monitoring, high overdues, negligible impact of eligibility criteria, failure of the SLDB rehabilitation program, interest rates, on-lending margins for PBs, incomplete or infructuous investments, diversion of funds, IRDP's effectiveness problems and the financial difficulties rf RRBs). The audit notes, however, thst in general the Overview is more g'Aarded than the borrower's PCR with respect to some achievements of the project, and, as befits staff concerned with the later NABARD project, is mure concerned with the implications of some of the unresolved issues. 36. Overall, the audit concludes that while there were undoubtedly successful investments financed in the rural areas by project funds, and hundreds of thousands of poor farmers (and othe-- clearly benefitted, the general impact of the project on agricultural credit institutions and the - 11 - agricultural credit system in India was largely negative. By making large external resources available in the short-run to the system, without ensuring that the system's serious problems received appropriate attention, the Bank failed to contribute to developing the credit system in the long run.5/ The sustainability of the institutional credit system at reasonable cost was not enhanced by the Bank's participation in the project. It is doubtful whether meeting the short-run disbursement targets of the Bank and government's need for foreign exchange during this two-year period, through an operation such as the project, was adequate justification for continuing to support a deteriorating institutional credit system in India. 37. There are three other areas which the audit highlights as of im- portant general significance. These are (1) the broader rural financial sector context of the Bank's agricultural credit operations in India, (2) the financing of IRDP through this general line of credit and (3) the performance of the Bank, especially the staff resources devoted to the work program and management follow-up. B. The Rural Financial System 38. The Overview by the Region raises the question (para. 30) of "whether significant and fundamental changes in agricultural credit insti- tutions can be achieved through an all-India apex organization such as NABARD since agricultural credit falls in the jurisdiction of state govern- ments". (This theme has been a concern for some years and appears as a question raised by a manager in 1984.) This Overview comment is placed in the context of the ARDC series of projects as a whole, but certainly the failure under ARDC IV to apply eligibili:y criteria and to rehabilitate weak SLDBs supports the Overview's conclusions. In terms of processing ARDC IV, however, this question would have appeared to be irrelevant in the early 1980s as it is apparent that the staff involved were in effect required to find ways and means of channelling through ARDC a further time- slice of Bank Group funding for agricultural credit. Thus there is no suggestion that broader financial sector issues were on the agenda, and working level staff, if not their managers, were entitled to assume a fait accompli with respect to funding channels and the broader financial sector issues. The most telling evidence for this conclusion is the absence in the SAR of any explicit demonstration of ARDC's need for the unprecedented level of proposed Bank Group funding. (The evidence now is that the need may not have been too critical, given that despite poor loan recoveries, external funding over the early 1980s, as a percentage of total ARDC/NABARD resources, was relatively small at around 18% overall. It was only later during formulation of the follow-on project, NABARD I, that external funding requirements were examined more closely, found to be limited, and the matter became an issue.) A broader more circumspect review at this time of rural financial sector operations might also have raised queries 5/ in the opinion of some knowledgeable Bank observers of the Indian agricultural credit scene, the Bank's involvement at this time may have even contributed to weakening the credit system. Other observers suggest that there is no evidence one way or the other. - 12 - about the large and expanding component of Bank resources going into poverty alleviation under the Integrated Rural Development Program (see further below). Prima facie, working through ARDC without a broader perspective of the financial sector as a whole, or even the major rural part of the financial sector, may not result in the right questions being asked on the structure and functioning of the agricultural credit system. The question posed at the beginning of this paragraph is just such an issue. 39. It is also questionable, in addition to whether NABARD had the power (legally or in practice), as to whether the Bank Group, once it became committed to two-year time-slices of large-scale funding, had the leverage or incentive to fo :e the needed changes. In terms of the rela- tive scale of the Bank's Indian lending program and agricultural credit operations in India, there is at least the possibility that the Bank's lending program needed the Indian project series more than India needed those particular funds (Bank funds were after all only a small part of ARDC's resources). In such a situation a project-related dialogue on matters such as recoveries and incerest rates, which have strong political interests involved, are not very effective. More recently such matters are more appropriately handled in connectlon with financial sector or agricul- tural sector policy-based loans (recent operations for Indonesia are a good example). C. The Integrated Rural Development Programme 40. The India agricultural sector lending program has featured rela- Uively few Bank-assisted projects directly aimed at alleviating rural poverty, especially given the large size of the country and low per capita income. To some extent most Indian agzicultural sector projects should alleviate rural poverty, but there were few of the Bank's archetypal rural devel.pment projects financed in India during the zenith of the Bank's rural development strategy in the 1970s. Instead, through the ARDC series of project , the Bank has financed a part -f the government's Integrated Rural Development Programme by disbursements to ARDC for credit refinanc- ing. IRDP credit totalled about 38% of refinancing eligible for reim- bursement under the project (about US$136 million of the Bank Group funds). Thus in effect the Bank has been indirectly assisting India's rural poverty alleviation efforts in a major way through a "side door", or "credit win- dow". This has not given the Bank the opportunity to address, with the government, the rural poverty question from the broader policy and program perspective. 41. The PCR and Overview report that NABARD's own field reports indicate that at least 20% of IRDP loans were misutilized and therefore benefits are less than expected. The main problem is reported to be targetted lending to families below the poverty line, but in such large numbers that the credit system could not properly process the loans, with the emphasis being on meeting the commitment targets rather than careful loan preparation and follow-up to ensure that intended benefits are realized. - 13 - 42. It is not possible, however, in the audit of such a large ope- ration as ARDC IV, to form an independent judgement on the quality of IRDP investments as a whole, but the main concerns are expressed well in the Bank's 1984 report on the Indian economy as followas "..... a number of concerns have been raised by observers of IRDP. First, administrative procedures for allocating funds apparently assume a uniform geographic distribution of the poor when in fact there are significant disparities. Second, there are problems with the identification of beneficiaries; these are of particular relevance considering IRDP's mandate to benefit the "poorest of the poor". This objective often conflicts with targets for numbers of beneficiaries to be raised above the poverty line. The latter criterion clearly leads to the selection of the "better- off" poor as beneficiaries. Third, the scope and quantum of assistance have bcen questioned, in terms of numbers of beneficiaries and irnestment per beneficiary. The grosL yield of the investment oftca appears too low to allow a significant rise in consumption. This is rte ir. part to loan repayment, working expenses, and administrative costs and fees. Fourth, there appears to be t.o much conccntratioa on livestock in spite of its unsuitability for large numbers of beneficiaries. Last, but not least, the IRDP program puts strain on existing rural development and agricultural institutions, in terms of tailoring assistance to the needs of individual households as well as in terms of meeting IRDP targets. These strains are dCLracting from the capacity of rural administration to carry ouf its normal functions. The strain on rural credit institutions aeems particularly great."6/ 43. The aud.t is concerned that the predominant investment under IRDP was livestock purchase. The largest single investment finax,ced under IRDP was dairy cows (42%), with bullocks and bullock carts a close second (41%). These activities involved some 1.6 million animals (PCR, Annex 1, Tables 3 and 8(b)). While such a vast transfer of ownerships of livestock may at first sight have attractive social benefits, there are doubts about the economic merits of such credit-dr. -en livestock sales. Furthermore, any perceived social benefits of such transfers of productive -apital (cattle) can probably be achieved at lower economic cost through other means. The main issue has been analyzed and discussed in other similar cases. It is that the transfer of ownerships of livestock is not of itself a justifiable investment unless the new owner can raise output from the animal above the level that might have been achieved otherwise. When output does not rise, or even falls, only the transfer of the site of production of an animal has been achieved, while in the process possibly raising livestock prices beyond the reach of the most needy.Z/ (The PCR Overview records that in a significant number of cases dairy enterprises were not successful. This 6/ Situation and Prospects of the Indian Economy--A Medium Term Perspective, Report No. 4962-IN, April 1984, Vol. I: Executive Summary, para. 44. 7/ A reviewer has noted, however, that livestock credit does contribute to the expansion of the subsector since higher prices should stimulate the production of more animals. - 14 - was because technical follow-up services were inadequate and other related investments were not available.) D. Bank Performance 44. Supervision of the project was unusually intensive with eight missions over 26 months. These missions, however, were also involved in preparation of the follow-on project, NABARD I, and there is coment in the files that there is "no clear distinction between supervision and prepara- tion" of ARDC projects. Indeed one regional manager wrote that the general "malaise" of the credit system was known to regional management for some time, but tended to be dealt with in the context of the preparation of the next operation, rather than thLough supervision of the current operation. 45. In a partial reversal of the nationwide approach adopted from 1975 (primarily to reduce Bank staff input)A/ more intensive supervision of ARDC IV (than previous ARDC projects had received) was premeditated by Bank project staff working on ARDC IV. Thus a memo by the appraisal mission leader in reply to a central projects memo refers to "greatly intensified IDA monitoring" of the project, and later the division chief of the res- ponsible division was instructed (on file) to raise supervision intensity, and did so.9/ Even so, the adequacy of the Bank's supervision of the pro- ject later became a source of great concern to senior management, espe- cially in connection with the lack of results on the overdues problem. 46. In an 4xtraordinary intervention in late 1983, a Bank senior manager telexed from India to regional senior management in strong terms in connection with the poor recovery performance--then running at around fifty percent. In referring to reasons for continuing poor credit performance the telex notes that ARDC's "supervision is inadequate but ours does not seem to have been very much better". This intervention led later to the appointment of a consultant to review the effectiveness of the Bank's appraisal and supervision of all four ARDC projects, including whether the Bank's decision in 1975 to switch to countrywide lines of credit was 8/ Although the size range of the IDA credits was within the normal range for operations at that time, and therefore each operation might have extracted--on a Bank wide basis--their share of supervision resources, the fact that the operations were all in one country, however large that country may be, did not apparently lessen concerns about this mode of Bank lending in India. Such decisions appear to be overly influenced by the Bank's practice of allocating budget funds for project work, and therefore staff positions, mainly against individual project operations with only secondary attention to the size and complexity of operations. 9/ Time recording data for FY83 and FY84 indicates about double the Bank average staff input for supervision for these ARDC IV years, and over double the preceding two years, but because supervision activities overlapped with preparation of NABARD I it is not certain that this properly reflects the time spent on AELC IV work. Moreover, since then supervision input for NABARD I has at first doubled then tripled ARDC IV annual levels. - 15 - premature. The ensultant selected was a retired senior Bank staff member, a credit specialist, and indeed the divisional manager who had responsibility for the program before and through the mid-1970s perio4 when the decision to go to nation-wide projects was taken. 47. The report was completed in April 1984 and concluded, on the positive sides - adopting the line of credi. approach (i.e. nationwide projects) in 1975 was correct since it achieved a substantial increase in lending; but on the negative side: - the expansion of Bank lending was too great (from $75 million in 1975 to $350 in 1982) since this did not take L.ito account the capacity of ARDC to control and administer such large increases; - successive time-slices (projects) were appraised and approved before lessons had been learned from previous projects; - too much "optimism and pious hopes" were expressed with respect to the reduction of overdues; - there was undue pressure from the Bank and the Government on ARDC to increase lending and to take on responsibilities for which it was not yet ready. 48. Given the directly related management role of the consultant at the time the decision was taken in 1975, the conclusion that moving to nationwide projects was correct could be viewed as merely self-serving. It is rescued from that judgement, however, by the undeniable fact that a substantial increase in lending did occur as intended. To that extent the conclusion is a statement of fact only, but it is doubtful whether that was the only reason or justifica-ion at the time for going nationwide--insti- tutional growth in India was also an important factor. To the extent that the main institutional problem--overdues--got further out of hand and tne Bank was unable or unwilling to apply corrective pressures across the board through national projects, the decision can now be regarded as unfortunate, if not incorrect. The Bank's potential for resolving the overdues problem on a state-by-state basis should have been greater, at least by exercising the sanction intended but not applied under ARDC IV of not lending to the poorest performing banks. (The selection of that particular consultant to review the issue of whether the 1975 decision was correct now seema questionable, with the manageable limits on insider bias and self-interest having been exceeded). 49. Comments on file note that while a (regional) consensus had not been reached on the report's recommendations--"Many a-e quite acceptable but the broad conclusions on future lending (e.g., small amounts, limited objectives) conflict with our intentions for NABARD I." 50. The question arises as to why regional and other senior Bank management did not take stronger action to force the Government and ARDC to - 16 - follow through on their obligations under thi project with respect to improving loan recoveries and rehabilitating weak SLDBs. Howiver, almost any action would have reduced the Bank's disbursement rate and thus delayed the programed next phase of lending--NABARD I. A regional manager offered another explanation at the time, writing that "supervision missions seem to have been bogged down in a morass of detail, unable to give a clear, concise warning of the few fundamental issues". Examinatica of the files, however, suggests that the position was clearly articulated to all levels of management and therefore that a tacit concensus must have existed nct to disturb the lending relationship. In particular the "chronic and serious overdue payments problem" was discussed at length at the ARDC IV Board discussion in February 1982, with the rising rates since 1974 being quoted (see para. 20 above). Furthermore the 1984 Bank economic report on India has a clear statement on the situation.10/ 51. One peculiarity of supervision mission reporting is related per- haps to a clear message not getting through sufficiently often to manage- ment. That is the performance rating of ARDC IV and its predecessor- -ARDC III. Although ARDC IV was rated a 3/3 ("major prcblems/deterio- rating") for the last of its three supervisions over seven months (and just prior to the regional manager's comment quoted above), in an earlier period of 6 missions over 27 months (from July 1980 to September 1982), while overdues were rising towards the 50% level, ARDC III and IV were always rated 1/1 ("problem-free or minor problems/improving"). Furthermore at no time during supervision of either ARDC III or IV was "Financial" listed as the type of probler in the standard reporting format. This seems a remarkable failure c, r a long time to place the overdues problem on the record properly in the project status reporting system. 52. Responsibility for rating projects at supervision falls to the division concerned in the first instance, but in this case alert senior managers should have queried why the well known poor recovery performance was not reflected in the project ratings, and required changes. Patterns such as this, of consistently "over rating" projects with known problems has been noted by OED before. Apart from inertia favoring not changing ratings downwards, the explanation in terms of disincentives seems to be that lower ratings attract attention (the main purpose in the first place) and create work, since "problem projects" are the subjects of special management reviews. The extent to which the apparent health of the Banks' 10/ "Overstrained managerial capacIty, inadequate and poorly trained staff and lack of effective procedures for supervision, monitoring and evaluation of credit disbursements have resulted in extremely poor loan recovery and high overdues. Political interference by State Govervments in loan recovery efforts has exacerbated this problem. Overdues have been particularly high for long-term credit, hovering Pround 50% for the last several years. Both commercial and co- operative banks are affected, but the latter have suffered more acutely since their loan portfolios are concentrated on agricultural lending, particularly for private wells. Many co-operatives have been forced to restrict lending and sizeable State Government subsidies are needed to keep them solvent. The impact of these constraints is evident in the virtual stagnation of real credit disbursements since the start of the Sixth Plan". Op. cit., Vol.II, para. 6.14. - 17 - total portfolio is skewed by such bias is an issue worth deeper analysis, and if verified, some corrective action. - 19 - ATTACHMENT COMMENTS OVERSEAS DEVELOPMENT ADMINISTRATION Bland House, Stag Place, London SWl 5DH Tele: M03MA FAX: 01471 Teephoe: 01473 r -*w014733G00 0TN: 273 Dat 5 May 1989 Mr Graham Donaldson DED World Bank Room 6045 600 - 19th Street NW Washington DC 20433 USA I refer to your letter of 27 March enclosing a copy of the PPAR on the 4th Agricultural Refinance & Development Corporation Credit Project in India. I enclose a copy of a letter from James Copeatake, an economist whom we have recently funded to study financial institutions in rural India. I hope this will be helpful* have nothing to add to Mr Copatake's comments. C Roth Internatio Inancial Institutions D rtment - 20 - HARVEST HELP 3 NaY 1989 Tis FOY Reem 8702 ODA Ssad House Stag Place LODON SW1 ARDC-4 PROJECT PERFORMANCE AUDIT 1. It cannot be assumed that the actual incremental effect of NABARD rerinance (and hence external funding of NABARD) is increased agricultural lending. Commarcial banks lend up to 40% to agriculture and othe.' priority sectors as they are required to by Government, but rarely above this. Financing this lending with NABARD funds frees their own resources (mainly deposits) for more profitable business and commercial lending. nnf has, however, claimed an increasing proportion of their own funds by raising statutory deposits with the RBI. 2. The PPA makes clear that the problem of overdues got worse during the period and not better. This is serious becauses X I it suggests economic returns to loan investments were low; b) recovery of overdues Is costly in terms of staff time and a small percentage inevitably has to be written off - reducing bank profital C) recirculation of bank capital is reduced. The report is surprisingly critical of the effectiveness of eligibility criteria for receipt of refinance as a way of improving recovery. Surprisingly, because it is the only measure With teeth" that NABARD/WB have been able to use. The Pilot Project to strengthen the credit Delivery System does appear to have some positive effect on loan recovery in the districts where it operates - but it is very expensive in terms of incremental staff costs. I would argue that it has been successful in large part by making recovery an indicator of branch -manager perfnrmanCt in addition to deposit mobilisation and POP" ftmpoWe help, to grow nu Ied. / Tn Foy, ODA - 21 - 3 Ma 1989 and meeting lending targets. 3. I as coming round increasingly to the view that low recovery rates are inevitable so long as banks know that they will alw.-ye be bailed out - by state governments in the case of co-op institutions, Government in the case of PR8a and commercial banks. This is particularly true in the case of government sponsored programmes (especially IRDP) under which banks have little control over borrower selection and appraisal. 4. A final point worth taking into account is the tension that exists between bankers and government over agricultural credit policy. Training of middle and senior managers in the commercial banks is, I think, important to ensure that bankers hold their own, and make best use of their available room for manoeuvre. These remarks are, I fear, very rushed. If you want more, I suggest you go to my Escor Report (No 3905) entitled : Evaluation studies of VABARD refinanced credit schemes in India : a review. Chapter 3 is a straightforward description of how NABARD refinancing works; Chapter 4 lists the problems of evaluation and Chapter 5 attempts to draw conclusions from the evaluations that have been done. Yours sincerely James 0 CoVestake