India - Fourth Agricultural Refinance and Development Corporation Credit Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No.7951 PROJECT COMPLETION REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) JUNE 30, 1989 Agriculture Operations Division Country Department IV Asia Regional Office This document has a restricted distribution and maN be used by recipients only in the performance of - - - > vw be W  ABBREVIATIONS 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 Coi_ercial Banks CBs - Comnercial Banks COTELCOOP - Comittee on Term Lending through Cooperatives CRAFICARD - Comittee 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 Programe NABARD - National Bank for Agriculture and Rural Development PACs - Primary Agriculture Cooperatives PLDBs - Primary Land Development Banks RBI - Reserve Bank of India RDPCs - Rural Development Project Courses RRBs - Regional Rural Banks SCBs - State Cooperative Banks SLDBs - State Land Development Banks  IO oUVICIAL U ONLY Tlo WOftL BANK Wahwqton. D.C. 20433 U.S.A. Oke Wd 0w.- Gw4 OpseaImM IEUbauA June 30, 1989 DIORANDUK TO THE MECUTIV DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on India Fourth Agricultural Refinance and Development Corporation Credit Project (ARDC IV) (Loan 2095-IN/Credit 1209-IN) Attached, for iaformation, is a copy of a report entitled 'Project Completion Report on India - Fourth Agricultural Refinance and Development Corporation Credit Project (ARDC IV) - (Loan 2095-IN/Credit 1209-IN)" prepared by the Borrower, with an Overview prepared by the Asia Regional Office. No audit of this project has been made by the Operations Evaluation Department at this time. Attachment This document has a rstricted distribution and may be used by recipients only in the performance of their official duties. Its :ontents may not otherwise be disclosed without World Bank authorization.  AL . PROJECT COMPLETION REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (LOAN 2095-IN/CREDIT 1209-IN) Table of Contents Pare No. Preface ..# .. . ................................................ i Basic Data Sheet ............................................. ii Evaluation Sulnmary ........................................... iv OVERVIEW General .......................................... 1 Project Objectives and Achievements .... .......... 2 Financial and Economic Returns ..... .............. 4 Loan Recoveries and Eligibility Criteria .... ..... 5 Lending to Small Farmers under IRDP .... .......... 6 Regional Rural Banks (RRBs) ..... ................. 7 Rehabilitation Programs for Weak SLDBs .... ....... 8 Project Impact on Institutional Development ...... 10 Financial Performance of NABARD .................. 11 Lending Procedures, Supervision and Monitoring by NABARD ...... ..................... 11 Bank Group Performance ...... ..................... 14 PROJECT COMPLETION REPORT I. INTRODUCTION ..................................... 15 II. PROJECT IMPLEMENTATION ........................... 18 III. PROJECT BENEFITS ................................. 28 IV. ROLE OF PARTICIPATING INSTITUTIONS .... ........... 31 Annexes I. Table 1 - Cost of the Fourth ARDC Credit Project Appraisal Estimate ..... .............. 43 Table 2 - Cost of Project - IDA/IBRD Slice: Appraisal Estimates .................... 44 Table 3 (a) and (b): State-wise and Purpose-wise Disburse- ments Under the Project ................ 45 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.  Pate No. Table 3 (c) State-wise and Purpose-wise Disbursements Under IRD Program ...... 47 Table 4 - Cumulative Disbursements of Funds Under the Project... .................. 48 Table 5 (a) and (b): State-wise Details of Actual Project Costs ............ 49 Table 6 (a) U.ats of Minor Irrigation Financed Under Non-IRDP Schemes ............... 51 Table 6 (b) Units of Minor Irrigation Financed Under IRDP . . . ...... ... . 52 Table 7 - Groundwater Potential in States (19S2). 53 Table 8 (a) Diversified Investments - Units Financed Under Non-IRDP Schemes ...... 54 Table 8 (b) Divwrsified Investments - Units Financed Under IRD Program ........... 55 Table 9 - Project Cost and ARDC Disbursements in Less Developed Areas ................ 56 Table 10 - NABARD Disbursements in Less Developed Areas ........................ 57 Table 11 - Small Farmers Coverage ................. 58 Table 12 - Staffing Pattern of National Bank Composition of Staff ................... 59 Table 13 - Composition of Staff - Technical Service ...................... 61 Table 14 - Staff Position of National Bank as of June 30, 1984 .................... 63 Table 15 - Resources Mobilization by ARDC/NABARD ............................ 64 Table 16 - Cash Flows ............................. 65 Table 17 - Condensed Statement of Income and Expenditure .............. 66 Table 18 - Condensed Balance Sheets ............... 67 Table 19 - Financing Agency-wise Position of NABARD Disbursements ................... 68 Table 20 - Total Borrowings of LDBs - Debentures Floated During the Financial Year (April-March) ..................... 69 Table 21 - Lending by Land Development Banks Cooperative Year (July-June) ........... 70 Table 22 - Land Development Banks - Recovery Performance at Primary Level.. 71 Table 23 - Summary of Land Development Bank Recovery (1974-1984) ................... 72  Pat* No. Table 24 - State Cooperative Banks and Central Cooperatives Banks Sumary of Overdues of Agricultural Advances... as at the End of June ... 73 Table 25 - Direct Institutional Finance for Agriculture ... ..... 74 Table 26 - Indirect Institutional Finance for Agriculture .......... . 75 Table 27 - Bank Group-wise/Population Groupwise Distribution of Comercial Bank Offices in India .... 76 Table 28 - India Scheduled Commerclal Banks: Agricultural Advances Overdues - All Loans .............................. 77 II. Pilot Project Studies for Quality Control of Agricultural Pumpsets ........................... 78 III. Training .......................................... 92 IV. Important Policy Decisions During the Project Period ............................................ 103 V. Table 1 - A Study of Minor Irrigation Scheme in Allahabad District of U.P .... 119 Table 2 - A Study on the Working of the Deep Tubewells in West Bengal, Hooghly and West Dinajpur .......................... 120 Table 3 - Public Tubewells and River Lifts in Orissa ................................. 122 Table 4 - Public Tubewells and River Lifts in Orissa ................................. 123 Table 5 - Minor Irrigation Scheme in Rohtas District of Bihar ...................... 124 Table 6 - Evaluation Study of Dairy Development in Malerkotla Area of Sanorur District, Punjab ................................. 125 Table 7 - Evaluation Study of Dairy Scheme in Surat District Gujarat ................. 126 Table 8 - Evaluation Study of Dairy Development in Bharatpur District, Rajasthan ....... 127 Table 9 - Evaluation Study of Marine Fishery Scheme for Mechanized Boats in Bulsar District, Gujarat ...................... 128 Table 10 - Evaluation Study of Inland Fishery Scheme in Krishna District of Andhra Pradesh ................................ 129 Map - IBRD 15972R1  - i - PROJECT COMPLETION REPORT INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION CREDIT PROJECT (ARDC IV) (Ln. 2095-IN/Cr. 1209-IN) PREFACE This is the Project Completion Report (PCR) of the Fourth gricultural Refinance and Development Corporation Credit Project in India, for which a Loan and a Credit in the total amount of US$350 million was approved on February 23, 1982. The Loan and Credit were closed as scheduled on June 30, 1984. They were fully disbursed and the final disbursement was on November 20, 1984. The PCR was prepared by the Borrower (NABARD, previously ARDC) and is supplemented by an Overview prepared by the Agriculture Operations Division, Country Department IV, of the Asi_ Regional Office. The Overview draws heavily on NABARD's report, and is based also on reviews of the Staff Appraisal Report, correspondence with the Borrower, internal Bank memoranda on project issues as contained in relevant Bank files, as well as on interviews with officials both in the Bank and in NABARD who have been associated with the project. This PCR was read by the Operations Evaluation Department (OED). The draft PCR was sent to the Borrower on March 27, 1989 for comments by May 8, 1989, but none were received.  - ii - PROJECT COWLETION REPORT uSDIA FOURTH AORITCLTURAL REFDMANCE AND DEVELOPMENT COWORATION CREDIT PROJECT (AROC IV) (LOAN 25-IN/CREDlT 1209-IN) BASIC DATA SHEET Key Prol-et Date Appraisal Actual or ActmaI as X of Estimate Estimated Actuol Appraisal Estlmte Total Project Cost (US3 Million) 2,JJB.2 - 5/ Total Project Cost (Re M) l6,d96 - !/ Eligible for IDA/11RD Flnnncing (USS U) 1,317.4 Eligible for IDA/IBRD Financing (Ru U) 1,5C4 12,61C 119.7 Credit Amount (SON U) 139.6 139.6 191 Date of Board Approval 2/23/82 Oate of Effectivene 6/25/82 Loan Amount (USS VillIon) 199.0 196.4 100 Dats of Effectiveneos 5/25/82 Date of Credit/Loan Agreemnt 2/24/62 Closing Dats 6/30/84 6/36/64 Economic Rates of Return (%) t/ 30Y-61Ub Negative to 5641 Financial Rates of Return (X) t/ 2ex-48x lox-6ex ubuebr of Direct Beneficiaries 1.4 million 2.2 million 167 STAFF INPUTS (stff wobs) FY75 FY76 FY77 FYt7 FY7t FY6 FY61 FY62 FY8S FY84 FY85 FY86 FYO7 FY6" TOTAL Preapproisal .1 1.2 6.7 19.1 26.2 Appraisal 66.8 36.6 92.6 Negotiations 22.6 22.6 Supervision 6.0 29.4 24.6 .3 6.6 .68 .$ 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 .5 6.6 .6 211.6 CUtMLATIVE DISBURSEMENTS FY62 FY6S FY64 FY6S Appraisal Estiate (US3 Million) .0 190.t 3U.6 34a.4 St Actual (US2 Million) 40.6 156.0 34.5 348.9 Actual as X of Estimate 161.5 82.6 97.3 97.4 Oate of Finai Disburssemt of Loan 11/20/94 Date of Final Disbursemet of Credit 6/26/63 aI Tle ARDC IV project was described as la two-year time-slice" of ARDC's ongoilg lending program (par& 4.30 of SAR). Total project costs (estimated at Rs 16,690 million) were based on aggregate costs of all field-level investments supported by ARDC refinance including equity contributions of borrowers and portions financed 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 inv-stment subsidies provided by central and state goverlments (for examp'.e, 2j2 to 331 of capital costs for IRDP beneficiaries--SAR par&. 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 USS equivalent.  -1ii- No. of badays specializations Portooc Types of Date poerons in field Represncted 3/ Rating 5/ Trend Probl 4_ ( olr) Pr peratiem 11179 1 3 A VA A yA Review of rto3jet 6-7/81 2 10 3,3 VA A VA pr ratioo Appraisal 4-S/81 7 19 A,5,C,D,F,C MA I I Subtotal m Supervision 1 2/82 2 7 A,5 1 2 1,0 Supervision 2 5-6/82 3 35 A,B 1 2 1,0 Supervision 3 10/82 2 24 8,D 1 2 1,0 Supervision 4 2-3/83 2 35 B,C 2 2 1,0 Supervision 5 5-6/83 2 20 B,C 2 3 1,0 Supervision 6 10/83 2 13 B 3 3 1,0 Supervision 7 1/84 3 18 8,E 3 3 1,0 Supervision 8 3-4/84 2 8 B,E 3 3 1,0 Subtotal 160 Total 369 OTHER PROJECT DATA Borrower: Government of India cuiiing Agncy: Agricultural aef mmnce And Develop ent Corporttion (now NALA3I) Fiscal Year of the Borrower: July 1 - June 30 Nam of Currancy Rupee (is) Currencr FschOtge Rate: Appraisal Reprt US$1.00 - Rs- 8.0 Appraisal Year Average US$1.00 * Rs 10.2 Intervening Years Average US$1.00 R is 10.8 Completion Year Average US$1.00 - Rs 11.6 Follow-on Project: ono SeeABARD-I Credit Project Loan u br Loan 2635-IN Loan Ahouat (US$ million) 375.00 Date of Bard Approval 2/25/86 3/ specializations represented: A a Agriculturalist; B * Agricultural Credit Speciali:c; C u Financial Analyst; D a Loan Officer; e - Fisheries Specialist; P - Economics; C - Irrigation Speialist; H - Cooperative Specialist. 4/ Performance Rating: I - Problem-free or minor problems; 2 * Moderate problems; 3 a Major problems. 5/ Trend: 1 a Improving; 2 - Stationary; 3 a Deteriorating. 6i Types of Problems: F a Financial; M - Managerial; T * Technical; P - Political; O a Other.  - iv - PROJECT COMPLETION 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 Corporation Credit Project (ARDC IV) was the fourth in a series of national agricul- tural credit projects supported by the Bank Group in India. The project was intended to finance a two-year time slice of ARDC's lending operations covering the period July 1, 1982 to June 30, 1984. Total project cost, estimated at US$2,086 million, was supported with an Ir Credit of US$160 million and a Bank Loan of US$190 million. Objectives 2. AF mately 45Z of ARDC's lending during the project period was expected to i for minor irrigation systems, 33% for pumpset electrifica- tion and farm mechanization and the remainder for a wide variety of on-farm investments, including plant.tion crops, dairying, poultry, fisheries, storage and market-yards. Bank Group funds under the project were to sup- port these investments with the exception of pumpset electrification, farm mechanization and forestry which were not eligible for reimbursement under the ARDC IV project. Fifty-five percent of the Bank/IDA funds were ear- marked for loans to small farmers and about 1.4 million farmers were expected to benefit under the project. 3. In order to address the principal problem of poor recoveries in agriculture credit, the project contained provisions for: (i) stricter criteria for eligibility of refinance from ARDC based on recovery perfor- mance of commercial and cooperative banks, it was hoped that such stricter measures would provide the incentive for the banks to focus on their re- covery performance, and (ii) measures to implement rehabilitation programs to strengthen weak LDBs in six states (Maharashtra, Karnataka, Bihar, Madhya Pradesh, Tamil Nadu and Gujarat). US$4.0 million of the IDA/Bank funds was also allocated to provide continued support for ARDC's training programs in agriculture credit. Implementation Experience and Results 4. The project was implemented largely as planned; Bank and IDA disbursements were substantially completed by the original closing date of June 1984 (although some final disbursoments were allowed until November 1984 under the project). As indicated in the PCR prepared by NABARD, except for the training component, most of the targets under the project  - v - were met or exceeded. ARDC's refinance for small farmers amounted to 77X of its total lending and the number of beneficiaries under the project is estimated at about 2.2 million. Analyses by NABARD of a sample of scheme- lending models indicate that vith the exception of public tubewells, investments supported by NABARD were financially and economically viable in the range 152 to over 50X. 5. Results on overall institutional developments, however, were less favorable: agriculture credit continued to suffer from poor recoverics and the enforcement of revised eligibility criteria for ARDC refinance had no sustained or perceptible impact in this regard. While ARDC and the LDBs went through the formal procedures to prepare rehabilitation programs, it became clear that State Governments were unwilling to bear the full burden of financial rehabilitation nor to take politically unpopular measures required to instill better financial disciplines in the banks and borrowers. ARDC/NABARD also found it difficult to deny its refinance to non-performing banks for lending under Government priority programs and schemes. 6. ARDC's targets (for loans to small farmers and total benefi- ciaries), for example, were exceeded mainly because of a rapid and signi- ficant expansion of lending under IRDP during the project period. Under the interest rate structure (fixed by GOI), banks are allowed a margin of 3.5% on IRDP lending which is grossly insufficient to carry staff-levels necessary for effective supervision of a large number of loans. While the involvement of GOI district officers combined with procedures for "automatic-loaning' enabled the banks to meet IRDP lending targets, this was achieved at ,he cost of overall "quality" considerations in agriculture lending. Sustainability 7. Lending for agriculture in India is supported essentially through: (i) commercial banks which are required to lend at least 16% of their total advances to agriculture, and (ii) State Cooperative Banks (SCB/DCCBs) which mobilize deposits (and i.-ceive state government resources) to finance mainly short-term (crop) loans. NABARD's role in this structure is mainly to supplement these resources by channelling GOI/RBI funds at reasonable costs to refinance part of the agriculture term lending of commercial banks, Land Development Banks (LDBs) and Regional Rural Banks (RRBs) and to provide lines of credit for short-term operations of SCBs/DCCBs. The Bank Group funds in turn provided part of the resources required for ARDC/ NABARD's refinance of term-lhiding operations. 8. NABARD obtains most of its rerources from RBI and through the recycling of previous advances to the banking system and credit remains an important instrument for the Government to accelerate capital formation in agriculture and to expand food production. Purely in terms of sustaining its total resource base, IDA/Bank funds while important are not critical. Greater impact of Bank Group involvement has been to give support to NABARD's efforts to bring about improvements in the institutional framework for agriculture credit.  - VI - Findinits and Lessons 9. As an apox-level institution NABARD has been effective ir ensuring that GOI and Bank Group funds are utilised for lending in agriculture through the existing network of commercial and cooperative banks. Nost of the physi-al targets under the project were set or exceeded. 10. The main lrassons under ARDC IV (and indeed previous ARDC projects) has been that ARDC/NABARD alone cannot bring about sustained or permanent improviments in pelicy and institutional environment for agriculture credit through its powets to sanction refinance facility. This requires a full support and a h'gher level of commitment by the Central ard State govern- ments and top -.anagement of institutions engaged in agriculture credit. It also impinger on some difficult and sensitive issues which need to be tackled, suzh as irterest rate structure for agriculture, margins allowed for credit administration and supervision, formulation and implementation of government sponsored programs for poverty alleviation, role Ct credit and subsidies in such programs, lending and supervision procedures of NABARD and primary lending banks and functions of various government and banking institutions in agriculture credit as a whole. 11. In the context of the above lessons, the follow-on project (NABARD-I: Loan 2653-IN) contains provisions for GOI to undertake a thorough review of agriculture credit and credit institutions and to intro- duce key policy and institutional reforms to strengthen the agriculture credit in India.  PROJECT COMPLETION REPORT INDIA FOURTH AGRICULTURAL RYFINANCE AND DEVELOPMMNT CORPORATION CREDIT PROJECT (ARDC IV) (LOUAN 2095-IN/CREDIT 1209-IN) Bank/IDA Overview of the Project Completion Report Prepared by National Bank for Agriculture and Rural Development (NABARD) 1/ General 1. The Project Completion Report (PCR) for the Fourth Agricultural Refinance and Development Credit Project (ARDC IV) was prepared by NABARD.2/ This overview, prepared by the Bank's Regional staff, supple- ments NABARD's review of the project with the Bank's assessments and observa- tions, focussing on areas not fully covered in the review. 2. The ARDC IV project was the fourth in a series of Bank/IDA projects beginning with ARDC I (IDA Credit 540-IN, for US$75 million, 1975-1977), followed by ARDC II (IDA Credit 715-IN for US$200 million in 1977-1979), and ARDC III (IDA Credit 947-IN for US$250 million during 1980-1981). ARDC IV was fol- lowed by NABARD-I (Bank loan of US$375 million), which became effective in September 1986. 3. PCRs for the first three ARDC projects have been reviewed by OED; the latest Project Performance Audit report for ARDC III was issued by OED in September 1985 (Report No. 5857). 4. These agriculture credit projects followed the Bank Group's decision to provide country-wide lines of credit through an apex organization instead of establishing separate credit lines to individual states, ten of which were fi.:nced by IDA between 1969 and 1975. The decision took into account the high manpower requlremenLa for supervising separate state projects and the fact that ARDC/NABARD, as the apex-level organization, could perform the supervisory functions for channelling credit funds to states. 1/ This supplement has been prepared by South Asia Regional staff. The original version of the PCR prepared by NABARD (previously ARDC), with minor editorial changes, is attached to this overview. 2/ ARDC IV was financed by an IDA Credit (US$190.0 million) and an IBRD Loan (SDR 139.0 million). References in this overview to the Bank include both IBRD and IDA.  -2- S. The PCR reports favorably and adequately on utilization of Bank/IDA funds and attainment of physical targets under the project: the funds were disbursed over two years as scheduled; investments financed by NABARD during the project period were in line with appraisal targets with significant support to small farmers (77% of projects funds). In fact, financial assis- tance to small farmers was in excess of appraisal estimates of 55Z. However, as in the case of ARDC III, the PCR does not deal adequately with the problems facing agriculture credit in India during the project period: con- tinued high level of overdues, insufficient supporting services to agricul- ture production, inadequate staff and infrastructure at primary lending units/branches for effective supervision and administration of credit to farmers, further weakening of financial viability of lending units, par- ticularly the RRBs and cooperative banks, and a decline in credit discipline generally in many rural areas. 6. While an attempt was made to address these problems under ARDC IV, NABARD's efforts in this regard proved to be less effective than envisaged. Later, these chronic problems became the principal issue in the follow-up project (NABARD-1), which included as a significant component financing of a major review of the Agriculture Credit in India. The review is expected to analyze, among others, the above problems, and outline specific recommenda- tions to strengthen the credit system in the country. Many of the Bank Group's concerns and principal issues during implementation of ARDC IV were very similar to those under ARDC III. Bank's attempts to address these problems through measures such as eligibility criteria, rehabilitation of weak lending units, etc., should be viewed as a continuous process under these two projects. Moreover, under tne ongoing NABARD I, steps are being taken for institution building and for improving the performance of the banking system dealing with agriculture. Project Objectives and Achievements 7. By and large, the project achieved its overall objectives by expand- ing investments in agriculture and by supporting small farmers. Hcwever, in the area of institution building, project results were less than expectation at the time of appraisal. The project aimed at supporting a two-year time slice of ARDC's ongoing refinancing program in minor irrigation and other diversified lending in farm mechanization, dairy, livestock, horticulture, fisheries development, and for storage and market yards for agriculture products. 1/ 1/ NABARD's refinance of agriculture loans proceeds on the basis of invest- ment schemes which are prepared by participating banks (commercial banks, Land Developments and Regional Rural Banks--See para 33-41 below for discussion of schematic lending).  -3- 8. Total investment in agriculture supported by NABARD over the two-year period was expected to amount to Rs 16.7 billion (SAR para 4.30). Of this, Rs 10.5 billion was forecast as eligible for refinance under ARDC IV (Rs 7.9 billion in minor irrigation and Rs 2.6 billion in other diversified activities). NABARD's own disbursements for investment eligible under ARDC IV were projected to amount to about Rs 8.2 billion. 1/ Against these targets, actual investments for eligible purposes amounted to Rs 12.6 billion (201 higher) and NABARD's disbursements for such loans of Rs 9.1 billion were about 11% more than SAR estimates. About 78X of NABARD's disbursements were expected to be for minor irrigation and 222 for diversified lending whereas actual total disbursements (Rs 9.1 billion) were about equally divided between these two categories. The increase in lending for diversified pur- poses was partly due to rapid expansion of lending under the Integrated Rural Development Program (paras 17-19). 9. Other quantitative targets set under the project included the follow- ing: (a) About 1.4 million farmers were to benefit from the program throughout India; (b) About 60% of the beneficiaries were small farmers and at least 55% of Bank Group funds was earmarked for reimbursing loans to such farmers; (c) About half of NABARD's lending was to be directed to less developed states (Assam, Bihar, Himachal Pradesh, J&K, Madhya Pradesh, Orissa, Rajasthan, Uttar Pradesh and West Bengal); (d) Investment supported by NABARD was expected to result in incremental production (mostly of foodgrain), valued at about Rs 5,200 million, and to generate additional employment of 134 million mandays per year; (e) NABARD's training programs were expected to benefit 22,000 staff of participating banks. 10. As stated in the PCR, except for training, these targets were sub- stantially met or exceeded: (a) total number of beneficiaries under the project was estimated at 2.2 million (of which 890,000 were for minor irriga- tion and rural development projects); (b) 77% of NABARD's refinance was in respect of loans to small farmers; (c) disbursements to less developed states 1/ NABARD's lending for pump electrification, farm mechanization and forestry projects was not eligible for financing by Bank Group funds under ARDC IV.  -4- amounted to 49Z of NABARD's refinance during the project period; (d) total value of incremental production, in 1983-84 prices, was estimated at Rs 10.4 billion (Rs 3.6 billion for minor irrigation and Rs 6.8 billion for other diversified projects). These investments have generated permanent employment equivalent to 411 million mandays; and (e) NABARD continued to improve its training programs; however, the total number of staff participating in its training courses (5,911) was short of the appraisal target and NABARD util- ized only US$2.8 million of the project funds for training against the original allocation of US$4.0 million under the project. 11. Given the large number of loans financed and limited staff capacity at the primary lending units, participating banks did not submit regular progress reports on physical completion of investments financed. In the absence of such reliable and complete data on physical completion of invest- ments financed, the above statistics on impact of projects refinanced by NASARD are derived partly from samples in ex-post evaluation studies under- taken by NABARD. Where such sample data were not available, estimates of investment costs, production and employment impact were based on unit costs and "norms" established under each scheme at the time of approval of schemes by NABARD. 12. On balance, it would appear that NABARD's estimates of project impact are probably overestimated since they do not take into account invest- ments not fully completed or funds which were diverted for other purposes. In many of NABARD's monitoring and evaluation studies, incomplete or infruc- tious investments and diversion of funds are often given as the main reasons for a scheme's failure to achieve its objectives. Financial and Economic Returns 13. SAR estimates of financial and economic rates of return of invest- ments financed by ARDC were based on 18 investment models built from data collected in NABARD's ex-post evaluation studies. These models indicated FRRs of 23% and 48% for investments for minor irrigation, and FRRs in the range of 20Z-43Z for other models (coffee, dairy, fisheries and land develop- ment). A sub-set of 11 of the 18 models indicated economic rates of returns of between 32% and over 50%. 14. The PCR attempts a comparison of SAR estimates with FRRs and ERRs as estimated for sample beneficiaries in ten other es-post evaluation studies, with the following results:  Scheme State FRI ERR Ainor Irrigation Scheme Uttar Pradesh 25X 392 Deep Tubewells West Bengal above 50Z above 502 Public Tubevells Orissa negative negative to 4Z Minor Irrigation Uttar Pradesh 40X 472 Minor Irrigation Bihar above 502 above 50X Dairy Development Punjab 35X 362 Dairy Scheme Gujarat 252 36Z Dairy Development Rajasthan 202 292 Maritime Fisheries Gujarat 15% n.a. Inland Fisheries Andhra Pradesh 282 n.a. These results indicate that with the exception of public tubewells, investments supported by NABARD were financially and economically viable. Loan Recoveries and Eligibility Criteria 15. As under previous ARDC projects, poor loan recoveries remained one of the principal concerns during the appraisal of ARDC IV. Criteria for eligibility of participating banks for access to NABARD refinance were first introduced under ARDC I and continued through subsequent projects. The main objective of such criteria on a sliding scale, was to provide an incentive to banks to improve their recovery performance by linking future refinance to their collection performance. At the time of appraisal it was felt that the criteria applicable at that time had not had the desired impact and were therdfore revisad under which: (a) lending units were allowed unlimited access to NABARD refinance if collection were 752 or more of demand, 1/ (b) refinance to branches with recoveries of 60X-75Z was limited to the level of their average lending in the preceding five years or average collection in the preceding three years (whichever was higher), and (iii) refinance to branches with recoveries of less than 60X was not to exceed actual recoveries 1/ Defined as all repayments overdue at the beginning of the year plus principal and interest falling due during the year.  -6- in the previous year or average annual recoveries in the three preceding years (whichever was higher). 16. Application of the above criteria was expected to focus greater attention by the banks on improvement of their recovery performance. Experience under the project, however, showed that expectation in terms of improved performance was too optimistic. While the PCR claims that there was some improvement in recoveries, this was at best, short-lived and probably tha result of rescheduling of loans, and/or reduction in demand through State Government contribution or "blocking" of old overdues. There was no evidence of sustained, significant improvement of actual cash recoveries from bor- rowers and under the follow-up project (NABARD-I), the criteria were further revised to make them more restrictive for branches with poor recoveries. Lending to Small Farmers under IRDP 17. Under ARDC IV, NABARD continued its emphasis on Lending to small farmers. In fact, as noted above, about 77% of NABARD's refinance during the project period was for loans to small farmers (against the target of 60%). Total number of beneficiaries (and total employment created) also exceeded SAR targets mainly due to rapid expansion of lending to small farmers. As under ARDC III, the main driving force for this expansion continued to be through lending under the Integrated Rural Development Program. Commenced in 1979 through amalgamation of several poverty alleviation schemes, IRDP has since remained GOI's principal national level program for reduction of rural poverty. Share of IRDP lending in ARDC/NABARD's total disbursements increased from 1% in 1979 to 16% (under ARDC III) and then to 26% by the year 1983/84. IRPD currently absorbs about a third of NABARD's total refinance. NABARD's refinance of IRDP loans constituted 38.8% of total refinance eligible for reimbursement under ARDC IV. While this rapid expansion has enabled the implementing authorities to meet lending targets set for IRDP, it has at the same time raised concerns as to the overall quality of lending. It is unclear whether such loans are being implemented and supervised as intended, and whether the program is yielding the full benefits as intended. 18. IRDP's objective is to identify each year, 600 families below the poverty line in each of India'-s 5,011 blocks and to provide these selected beneficiaries with productive investments which would enable them to cross the poverty line. Block, District and State plans are drawn up annually identifying types of investments, costs and share of financing by participat- ing banks and NABARD. Selection of beneficiaries and preparation of individual investment proposals are handled by the District Rural Development Agency (DRDA). Investment proposals and simplified loan applications are presented to local branches of participating banks which theoretically have the right of refusal; in practice, however, bank branches do not have the staff resources to examine each application in detail and meeting the targets often becomes the overriding factor in banks' approvals. IRDP borrowers generally receive subsidies of 25Z-33% (50% for scheduled tribes) which is  -7- channelled through the participating banks. Banks are allowed a margin of 3.5% on IRDP lending which is insufficient to carry staff-levels necessary for effective supervision of a large number of loans. Branch field officers typically handle 1,500-2,500 active borrower-accounts. Several evaluation reports of NABARD and GOI indicate that many IRDP investments do not yield the full benefits as envisaged because of inappropriate investments selected, insufficient financial assistance for the selected investments, or inadequate supporting services for the investment. For example, the most important investment financed under IRDP is dairy development. Viability of IRDP loans for this purpose is based on borrowers purchasing at least two milch cattle. Many of NABARD's field reports, however, point out that a significant number of borrowers did not or were not allowed loan for a second buffalo. Findings such as these have in fact led to the conclusion by NABARD that at least 20% of IRDP loans have been misutilized and incremental benefits have been less than envisaged under the program. 19. On the positive side there is now a greater consciousness of the problems of IRDP on the part of GOI, NABARD, and participating banks. These problems which became increasingly obvious during implementation of ARDC IV were addressed during NABARD-1 appraisal. NABARD is also now instituting internal procedures to ensure a greater participation by lending banks in formulation of di.trict plans, selection of borrowers, etc. It is also expected that the Agriculture Credit Review under NABARD-I will make recom- mendation in this regard. It should be noted that unless lending units are allowed required staff to administer and supervise a large number o! bor- rowers, improvement of overall quality of lending under IRDP is likely to be impaired. Regional Rural Banks (RRBs) 20. The status of RRBs 1/ demonstrates the adverse effects on financial intermediaries resulting from a rapid expansion of small-farm credit without the required financial support for their operations. As of June 1980, there were 73 RRBs operating through 2,678 branches. ARDC refinance to RRB's in FY1979/80 amounted to RS 90 million, about 2% of ARDC's disbursements in that year. The SAR for ARDC IV noted that while small, RRB participation was growing rapidly and that "as it grows, it may become a major channel for development loans to small farmers" (SAR para 2.37). 1/ RRB is sponsored by a: Commercial Bank which contributes 50% of its capi- tal and provides management assistance; GOI and State Government provide 35% and 15% respectively of its share capital. Inspired by the Gramin Bank model, RRBs are set up specifically to provide banking services to small-farmers and weaker section of the rural population.  -8- 21. RRBs expanded rapidly in the first half of 1980s and by June 1985, there were 183 RRBs with over 12,139 branches, 14.8 million deposit accounts and 6.3 million borrower accounts. In FY1984/85 NABARD's disbursements to RRBs amounted to Rs 1,400 million, about 132 of its total disbursements in that year. While RRBs proved to be fairly successful in mobilizing rural deposits, they were facing severe financial difficulties: unable to cover their recurrent costs, about 130 RRBs were incurring losses, and 49 of these had eroded their entire share capital by December 1984. Loan recoveries by RRBs were also below national averages with about 55X of all RRBs recovering less than 60% of demand. 22. To address the problem of RiBs, GOI established a 'Working Group on RRBs' in 1986 (The Kalker Cooittee). The Comittee has made several recom- mendations to strengthen the management and viability of RRBs. These include: increase in share capital of RRBs, amalgamation/consolidation of non-viable branches, changes in interest rates charged to RRBs to allow them greater margins, and setting procedures for appointment and supervision of RRB managers. Recomendations of the Comittee are presently under con- sideration by GOI. Rehabilitation Programs for Weak SLDBs 23. Special attention to rehabilitation of weak SLDBs with poor recovery performance was first commenced under ARDC III and rehabilitation measures were originally intended to be tied to the revised, stricter eligibility criteria. Application of these criteria would have made LDB of Gujarat and Maharashtra ineligible for ARDC finance and performance of LDBs in four other states (Bihar, Himachal Pradesh, Karnataka and Tamil Nadu) was only mar- ginally better. In order mainly not to cut off fresh financing to such LDBs, the eligibility criteria contained provisions that would allow ARDC refinanc- ing if such LDBs had a rehabilitation program agreed with ARDC and in cases where 50% of branches (or PLDBs) of SLDB were inaligibLe, the state govern- ment had agreed a time-bound rehabilitation program with ARDC. As it turned out, application of these provisions (together with the stricter eligibility criteria) was deferred in September 1980 at the request of GOt. Rehabilita- tion programs were, nevertheless, prepared although their effectiveness remained doubtful. 24. With the above background, more concerted and focussed measures towards rehabilitation were again attempted under ARDC IV. Principal provi- sions under the project included the following: (a) setting up of a "Standing Comittee" by GOI to oversee and guide efforts for the six LDBs initially, with others to be added at a later date. The six LDBs for immediate atten- tion were Maharashtra, Karnataka, Bihar, Madhya Pradesh, Tamil Nadu and Gujarat; (b) appointment of six 'expert teams' one for each LDB (including suicable consultants) to carry out organization and management studies of the six LDBs and to recommend improvement programs; (c) based on these studies, ARDC was to prepare detailed, time-bound rehabilitation proposals to be  -9- agreed upon by the Standing Committee and the state governments; (d) eligibility criteria under ARDC IV included special provisions applicable to the six LDBs under which ARDC was to withhold refiaance assistance unless the LDB had undertaken to carry out financial and management improvements; (e) as part of financial rehabilitation measures, Bank/IDA agreed to the transfers of specified amounts identified as chronic overdues 1/ to "blocked accounts" in four of the six LDBs. 2/ Under the eligibility criteria, mounts blocked were not to be taken as a bank's annual demad. 25. As indicated in the PCR, serious attempts were made to implement these provisions: organization and management studies by expert groups and consultants were commissioned for the six LDBs originally targetted. In fact nine additional LDBs (Orissa, West Bengal, Rajasthan, Himachal Pradesh, Uttar Pradesh, Jainu & Kashmir, Assam, Tripura and Pondicherry) were identified and management studies undertaken for these; based on these studies, NABARD formulated rehabilitation programs for approval by the Standing Committee (Committee on Term Lending through Cooperatives - COTELCOOP) and state governments. 26. However, as it became progressively apparent during project implemen- tation, the LDBs and state governments were very slow or reluctant to imple- ment financial and organizational changes required. The provision in the project agreement enabling ARDC to withhold refinancing to LDBs not complying with the requirements of rehabilitation program was used only once for a very brief period (between September 1, 1982 to October 15, 1982) to obtain con- firmation from the LDBs that they (and the state governments) had accepted the rehabilitation programs approved by COTELCOOP. NABARD was reluctant to use this sanction to enforce the implementation of the measures under the programs. Without the effective implementation of other corrective measures, the procedures for "Blocking" of old overdue loans became by and large an exercise to circumvent the minimum recovery performance under the eligibility criteria. Although Bank/IDA supervision mission devoted considerable time in following the ARDC IV provisions regarding rehabilitation, virtually no improvement was reported in the recovery situation and in the overall finan- cial health of the LDBs targetted for rehabilitation. 27. These disappointing results (and indeed the lessons learnt) regarding rehabilitation efforts under ARDC IV are in many ways similar to those under ARDC III and the following conclusions in the ARDC III PCR are generally applicable to the ARDC IV also: "The proposed programs in some instances dealt only with symptoms, without addressing the more important underlying 1/ Generally over 5 years overdue. 2/ Bihar Rs 32 million, Cujarat Rs 185 million, Karnataka, Rs 54 million and Maharashtra Rs 229 million.  -10- problems. They dealt with removal of obstacles towards taking coercive action towards long-term defaulters, blocking of chronic overdues, rephasing of loans and handling incomplete investments. Limited attention was paid to basic institutional problems, like lack of field staff, inappropriate super- vision and appraisal practices. State governments were reluctant to imple- ment the programs. The experience during ALDC III raises questions concern- ing ARDC's ability to design and institute effective rehabilitation programs, which raises the wider issue of the de facto supervisory role of ARDC in agriculture credit and their effectiveness in instituting adjustments in basic credit operations cf its client banks. That role, and ARDC's effec- tiveness were never questioned in the appraisal and the subject does not appear in supervision reports. One can raise the question whether in India's political system it is realistic to assume that a more dominant role of a "center" organization would and will be accepted by banks controlled by the States." (ARDC III PCR, page 31, para 31). 28. The Institution Development Department (IDD) of NABARD continues its efforts to rehabilitate weak LDBs and NABARD-I project includes covenants towards strengthening of IDD. Given the past experience, the magnitude of the problem and the fact that overall policy and political environment in which NABARD operates has not changed materially, it remains unclear whether results under NABARD-I will change dramatically without key credit reforms. It should also be noted that under NABARD-I, emphasis has shifted back to stricter eligibility criteria. It is hoped that more severe restric- tions on availability of NABARD's financial assistance would prompt LDBs with poor recovery performance to take corrective measures for better credit delivery and administration. The role of NABARD vis-a-vis client banks will also be examined under the Agriculture Credit Review by the Senior Experts Group. Project Impact on Intitutional Development 29. The discussion above indicates that while the project succeeded in contributing to NABARD's resources and that while such investments probably dad a positive impact on production, rural employment and incomes, the project's impact on overall strengthening of institutions handling agricul- ture credit was at best only marginal; the results under the principal com- ponent for institutional development in the project (i.e. the rehabilitation of weak LDBs) were well below expectations. The enforcement of the eligibility criteria had little impact on overall recovery performance nor on greater credit discipline by participating banks. 33. The experience under the last two ARDC projects in fact raises the question whether significant and fundamental changes in agriculture credit institutions can be achieved through an all-Tndia apex organization such as NABARD since agriculture credit falls within the jurisdiction of state governments. For social/political reasons varying in each state, the state governments make policy decisions (including allocation of resources out of  -ll- state budgets to state banks, interest remission, subsidies, etc.) which directly or indirectly affect discipline and institutional environment for credit. Both the commercial banks, and the cooperative banks look to NABARD primarily as a source of low-cost funds for their agriculture lending. Thus, virtually the only leverage which NABARD has in order to bring about changes is through its refinance facility to participating banks. In the past, NABARD had been extremely reluctant to utilize this leverage and in cases where it has done so (e.g., through eligibility criteria) the results have not necesuavly been entirely satisfactory. 31. Difficulties of bringing about major institutional changes through the Bank loan to NABARD were recognized at the time of appraisal of NABARD-I project. As mentioned above, the Bank is attempting a different approach under this project by having the Reserve Bank of India undertake a thorough review of the agriculture credit and credit institutions in India. The review will allow GOI to introduce key policy and institutional reforms required to strengthen the credit system in the country. Financial Performance of NABARD 32. Despite prvblems related to recoveries of agriculture loans of par- ticipating banks, overall financial status of ARDC/NABARD has never been at issue. Loan repayments due to ARDC/NABARD continue to be repaid promptly and loans rescheduled by NABARD have been insignificant. In comparison to its predecessor (ARDC), the financial position of NABARD was in fact significantly strengthened at the time of its establishment in 1982. Its share capital was almost doubled (to Rs 1.0 billion) and with the transfer of credit funds from the R3I, total resources of NABARD increased to Rs 45.2 billion (compared to Rs 21.4 billion of ARDC). Since the resources from RBI bear no interest, NABARD's profitability is also significantly better than that of its predecessor; net profits of NABARD increased from Rs 0.9 billion in 1982/83 to Rs 2.4 billion 1985/86. Most of NABARD's profits are retained to augment the resources available for lending to participating banks. Lending Procedures, Supervision and Monitoring by NABARD 33. In order to analyse lending procedures of NABARD, it is first neces- sary to distinguish NABARD's lending for IRDP from its lending for schemes and purposes other than IRDP. As discussed in paras 17-19 above, IRDP is an all-India, poverty alleviation scheme in which the District Rural Development Authorities (DRDAs) play a key role in identification of borrowers, purposes for which loans are to be given by banks and release of subsidies to approved borrowers. Role of participating banks (mainly commercial banks and Regional Rural Banks) is confined to providing resources for the program by way of loans and then refinancing these through NABARD. Key focus in this process is often the desire to meet pre-established targets. Given the large number of loans, and limited capacity at rural branches, banks are barely able to  -12- undertake appraisal of borrowers nor follow-up and effectively supervise loans under IRDP. 34. At the time of appraisal of ARDC IV (May 1981), IRDP was still in the early stages of implementation and NABARD's total lending for IRDP in FY1980/81 of Rs 132 million constituted less then 3X of its total disburse- ments of Rs 5.0 billion in that year. During implementation of ARDC IV, however, IRDP increased very rapidly and by 9Y1983/84 it absorbed Rs 2.3 billion (or about 27X) of NABARD's total disbursements of Rs 8.6 billion. 1/ Expansion of this magnitude of a program in which the banking institutionsT role is limited essentially to provision of resources obviously has had an adverse effect on the overall quality of loan portfolio. Although it is virtually impossible to measure (or quantify) the effect of rapid increase in IBDP-lending, it is reflected in commercial banks' inability to improve their recovery performance and a deterioration in financial status of Regional Rural Banks (see paras 20-22 above) which deal almost exclusively in Lending to IRDP target groups. 35. There are no evident answers to the issues related to IRDP lending. One possible solution would be for lending units to increase their field and rural branch staff to handle more effectively the appraisal and supervision of IRDP borrowers and to improve the quality of investments. But this is constrained by the fact that the interest rate structure and margins allowed to banks on IRDP lending (of 3.5%) is barely sufficient to cover the banks' present cost of the program. 36. Nevertheless, under NABARD-I, steps are being t ken to involve NABARD's regional offices and participating banks in tne process of formula- tion of programs for IRDP lending at the district levels. It is also expected that IRDP lending will be examined under the Agriculture Credit Review (NABARD-I) and appropriate recommendations made to GOI in this regard. 37. NABARD's refinance of participating banks' loans for purposes other than IRDP is undertaken on "schematic lending". The underlying purpose under this approach is to handle relatively large numbers of loan applica- tion/appraisal and yet maintain some degree of "quality-control" on utiliza- tion of NABARD's resources. Participating banks prepare "schemes" for approval by NABARD, in which loans are defined (e.g., schemes for pump sets or tractors). The scheme also stipulates technical criteria for individual investments and minimum conditions which borrowers under the scheme would be required to satisfy. Once approved, the banks accept and approve loan ./ Increase in IRDP has continued, though at a slower pace in recent years; by 1985/86 IRDP disbursements by NABARD amounted to Rs 3.8 billion--about 32Z of total disbursements of Rs 11.9 billion  -13- applications and refinance these through NABARD within the total amount allocated for a scheme. 38. Apart from checking that refinance applications generally conform to the approved schemes, NABARD does not involve itself in detailed monitoring of individual borrowers or physical completion of investments financed. It does, however, carry out evaluation of a few selected schemes usually 2 or 3 years after the investments under a scheme have been completed. 39. Under ARDC IV, NABARD was to undertake 50 such evaluations. Although this target was not achieved during the ARDC IV project period, NABARD has now completed the 50 evaluations. The quality of these evaluation reports has been very good. In addition, under ARDC IV NABARD undertook "dis- trict-oriented monitoring" (D.O.M. studied) under which selected beneficiaries and participating banks' performance under various schemes is reviewed within a selected district. 40. As indicated in the PCR, borh evaluation and DOM studies have brought to light several problems in financing of schemes and constraint: as a result of which full benefits envisaged under the schemes do not materialize (see summary at para 2.29 of PCR). The principal problem within NABARD, however, was that there was no effective mechanism to translate lessons learnt from evaluation and DOM studies into actions directly related to approval of fresh scheme proposals. Some efforts towards this are now being implemented under NABARD-I; a unit within the operations complex has been set up to sift through evaluation/DOM studies and to prepare recommendation for staff engaged in scheme-approval indicating major types of problems in a given district or under a scheme. It is not yet certain, however, if NABARD would be willing to withhold its approval of a scheme based on findings of its evaluation/DOM studies of similar schemes implemented previously. 41. The schematic lending procedure increasingly adopted (and often required by NABARD) has never beLn questioned by the Bank. There is no doubt that schematic lending has been very effective in enabling the banking system to handle a large number of loans and yet keep the administrative cost down. The offsetting element in this approach, however, is that the overall quality of lending cannot be maintained at the same level as when a bank branch deals directly and individually with each borrower. Once a scheme is approved, there is a tendency to attain the scheme targets, and individual loan applications do not get as close attention as otherwise; bank branches ha"e complained for example that applicants do not always "fit exactly" the norms and criteria estabLished under a scheme, and that if such applicants are considered credit-worchy, the bank has either to accept them contrary to some of the scheme-criteria or reject chem. 42. The Agriculture Credit Review is also expected to review the schematic lending procedures and in particular assess whether or not the schematic lending should continue with or without any modifications.  -14- Bank Croup Performance 43. By financing the agricultural credit system through NABARD, the Bank has been able to provide neede resources for investments in the sector. However, considering the problems of the system, including the high levels of overdues and the deficiencies in the performance of MABADD and the par- ticipating banks, more intensive and frequent Sank supervision missions were necessary. In the implementation of the project, the Bank: (a) should have pushed more strongly in getting GO to increase the level of margins allowed to the participating banks; (b) should have questioned more vigorously the below-market rates of interest charged by the participating banks; and (c) should have given special attention to the increased level of NABARD's lend- ing under the IRDP program in terms of investments and impact on agriculture. On the positive side, the Bank maintained dialogues with GOI and its action in the introduction of the pilot project (para 12) was a positive first step to resolve some of the problems of che agricultural credit syscem.  -15- PROJECT COMPLETION REPORT INDIA INDIA FOURTH AGRICULTURAL REFINANCE AND DEVELOPMENT CORPORATION C .EDIT PROJECT (ARDC Iv) (LOAN 2095-IN/CREDIT 1209-IN) I. INTRODUCTION 1.01 The Fourth ARDC Credic Project (ARDC IV), envisaging financial assis- tance of US$350 M was approved by the Board of Directors of the World Bank on February 23, 1982. The total financial assistance to be provided under the project included retroactive finance of US$15 M in respect of disbursements made during January 1 to January 23, 1982. The financial assistance under the project comprised IDA credit of US$160 H and IBRD loan of US$190 M. 1.02 The main objectives of the project were to: help increase agricul- tural production, rural incomes and employment; give continued emphasis to assisting small farmers and less developed areas; continue efforts for institution building through further strengthening of NABARD; improve agricultural lending by banks and give particular emphasis to rehabilitation and strengthening of LDBs; and continue programs for improving quality of investments. The objectives were similar to those of the three general lines of credit completed/implemented earlier by the ARDC. 1.03 NABARD's association with the World Bank group dates back to the formulation of state specific agricultural credit projects between 1969 and 1974. The state-specific agricultural credit projects were succeeded by three general lines of credit and a number of activity-specific IDA/IBRD projects by channelling term finance to support investments in minor irriga- tion including land development, tree crop development, marketing, horticul- ture, processing, inland and marine fisheries, seed production and sericul- ture. As of June 30, 1984 financial assistance commitced by IDA/IBRD in respect of various agricultural credit projects amounted cumulatively to US$2.41 Billion against which US$1.46 Billion had been disbursed. 1.04 'The General Line of Credit' approach, with its national focus, involved a major departure from the state oriented projects earlier supported by the Bank Group. The enlargement of scope from state-specific projects to the general lines of credit indicated, inter alia, a growing confidence of  -16- the gank in the ability of erstwhile ARDC and NABARD in implementing nation- wide projects with greater coverage, flexibility in fixing targets and sanc- tioning of various types of investments, etc. The general lines of credit help save considerable time for the state governments, and the Bank in preparation and appraisal of individual state-specific projects. 1.05 Encouraged by the successful implementation of the state agricultural credit projects and the three general lines of credit, GOI submitted the proposal for ARDC IV in July 1980 to the Bank Group for approval. 1.06 In the context of the Sixth Five-Year Plan (1980-85), ARDC undertook a detailed review of its perspective lending program and adopted an indica- tive lending program as detailed below: Projected Lending Program Year (Schematic-Term Loans) (July-June) (Rs Million) 1980-81 4,500 1981-82 5,200 1982-83 6,000 1983-84 6,800 1984-85 7,500 30,000 1.07 The indicative lending program for the years 1982-83 and 1983-84 formed the core of the project report for the Fourth General Line of Credit. The Project sought a credit allocation of US$350 M (including training) which was to support NABARD's lending program of Rs 12,800 M. Issues during Appraisal and Negotiations 1.08 The project was appraised by the Bank in April/May 1981 on the basis of the project report submitted by ARDC, through GOI, in July 1980. The major issue at the time of appraisal related to the high level of overdues in the rural development banking system and designing of rehabilitation programs for six weak SLDBs, viz., Bihar, Gujarat, Karnataka, Madhya Pradesh, Maharashtra and Tamil Nadu. Other issues during appraisal related to the introduction of eligibility criteria for commercial banks and State Coopera- tive Banks, and submission of subloans amounting to US$1 M for post-facto review by the Bank. 1.09 The project was negotiated in Washington from November 9 to 25, 1981. The discussions during negotiations centered around the issues listed above. In addition, the organizational plans for setting up of NABARD, financial arrangements between G)I aid ARDC, loans to small farmers, setting up of Committee on Term Lending through Cooperatives (COTELCOOP), and flexibility  -17- regarding the percentage of reimbursement by the Bank were also discussed. It was also agreed during negotiations that not less than US$190 M (about 55 percent of the total Bank assistance of US$346 M excluding the training component under the project) would be utilized in respect of disbursements to small farmers. One of the improvements over the Third ARDC Credit Project was the enlargement of eligible investments by inclusion of storage and market yards. The Indian delegation canvassed for widening the scope of diversified investment categories by including financing of biogas, plough bullocks, bullock-carts, slaughterhouses, sericulture, mushroom development, etc. These were recorded in the Minutes of Understanding. 1.10 As a result of the reduction in the allocations of IDA credit to India during fiscal year 1981-82, the financial assistance to support ARDC IV comprised a blend of IBRD loan and IDA credit. 1.11 Under the three general lines of credit, successfully implemented earlier, the World Bank had been reimbursing at 50% of the disbursements made by ARDC. However, under ARDC IV the proportion of reimbursement was reduced to 34% in view of the larger size of the lending program and the inclusion of new categories of investment, viz., storage and market yards. As there was no possibility of increasing the amount of Credit/Loan beyond US$350 M, the rate of reimbursement was kept at the reduced level of 34Z. 1.12 The Indian negotiating team had indicated that in view of disburse- ment of sizeable proportion of refinance for investments in farm mechaniza- tion (around 22% of ARDC's annual schematic lending) and the latest trend in the investment-mix for fiscal year 1980-81, the composition of ARDC's lending would undergo a change necessitating an appropriate revision in the percent- age of reimbursement. The reimbursement percentage stipulated at 34% was increased to 50% of (National Bank) disbursement under the project with effect from April 1, 1983 in view of the depreciation in the value of the Rupee vis-a-vis U.S. dollar. The Project at Appraisal 1.13 The investment to be financed under the project were estimated to cost Rs 16,690 H (US$2,086.2 H) at the ground level (Annex I Table 1). As against this, the cost of the investments eligible for reimbursement under ARDC IV was estimated at Rs 10,528 M. 1/ It was estimated at the time of appraisal of the Project that about 75.44% of the cost of investments will be 1/ The estimates are based on May 1981 prices. The price contingencies of 9% for 1981, 8.5% for 1982 and 7.5% for 1983 onwards compounded annually have been applied to the base cost. Total contingencies work out to 11% of the base cost (Annex I Table 2).  -18- towards minor irrigation and Land development, 24.222 towards diversified purposes, and the remaining 0.34X on training (Annex 1, Table 2). II. PROJECT IMPLEMENTATION Disbursements and Financing 2.01 The disbursements under various project components other than train- ing were completed by March 31, 1984, and the project was closed as scheduled on June 30, 1984. Disbursements by NABARD under the project aggregated Rs 9,087.4 M as against the anticipated amount Rs 8,227 M. Claims worth Rs 3,488.7 M were sufficient to draw the entire credit/loan (corresponding to disbursements by NABARD of Rs 8,476 M). Under the project, NABARD supported investments at the grass root level (project cost) of Rs 12,614.95 M as against the appraisal estimate of Rs 10,528.35 M. ARDC/NABARD IDA/IBRD Project Cost* Disbursement Credit/Loan A praisal Actual Appraisal Actual Appraisal Actual (Rs M) (Rs M) (Rs M) (Rs M) (Rs M) (Rs M) MI & LD 7,943.16 5,316.95 6,389.0 4,483.6 270.0 162.7 Diversified 2,549.67 7,267.50 1,802.0 4,573.0 76.0 184.5 Purposes Training 35.52 30.50 36.0 30.5 4.0 2.8 /a Total 10,528.35 12,614.95 8,227.0 9,087.1 350.0 350.0 * Excluding investments not eligible for IDA/IBRD funding. /a The GOI have, however, released the entire rupee equivalent of US$4 M as grant to National Bank. 2.02 The pattern of investments actually financed under the project was somewhat different from that anticipated. Minor irrigation and land develop- ment together accounted for 49.34% of the project disbursements as against 77.66Z estimated at appraisal. The disbursements under diversified invest- ments, consequently, were higher at 50.32% as compared to anticipated 21.90%. The factors which contributed to the increase in the disbursement of refinance for diversified purposes are discussed in a subsequent section. Disbursement on training at Rs 30.5 M was lower than the appraisal estimate of Rs 36.0 M. The expenditure on training accounted for 0.34% of total  -19- disbursements under the project in comparison to 0.44% envisaged at the time of appraisal. 2.03 The pace of disbursements by National Bank during the project implementation period was generally as per schedule drawn in the appraisal report (Annez 1 Table 4). The disbursements, as per the appraisal schedule, helped GOI in drawing reimbursement from the World Bank on a continuing basis. Minor Irrigation 2.04 The minor irrigation structures under the project covered investments both by individuals and state groundwater development corporations. The cost of these structures amounted to Rs 5,194.67 H (Rs 4,064.5 M under non-IRDP schemes and Rs 1,130.62 M under IRDP) (Annex 1 Table 5(a) and 5(b)) as com- pared to the appraisal estimate of Rs 6,702.0 M (excluding price contingen- cies). The National Bank provided refinance of Rs 4,379.0 M, which con- stituted 84.3Z of the cost of minor irrigation works. 2.05 The number of minor irrigation units financed 1/ under the project was estimated at 0.80 M, of which 0.63 M, were financed under non-IRDP schemes and the remaining 0.17 M, under IRD Program. For ensuring efficient use of the irrigation water, lining of water courses and construction of water distribution system and sprinkler sets were covered under the project. 2.06 The state/union territory-wise details of the completed unit equiv- alents are given in Annex 1 Tables 6(a) and 6(b). A summary of the same is given below: Units Investment Non-IRDP IRDP Total Dugwells 126,112 57,S60 183,672 Dugwells and Pumpsets 37,389 1,086 38,475 Pumpsets 182,986 62,779 245,755 Renovation of wells 61,035 20,091 81,126 Persian wells 49447 - 4,447 Shallow tubewells 203,487 26,906 230,393 Deep tubewells 1,371 - 1,371 River Lift Irrigation Units 623 - 623 Sprinkler sets 7,443 - 7,443 Others including water distribution system 7,373 1,313 8,686 Total 632,256 169,735 801,991 1/ Relates to completed unit equivalents of units financed.  -20- 2.07 The number of beneficiaries of the minor irrigation investments under the project is estimated at 0.85 M. The investments in minor irrigation are estimated to have added to the irrigation potential by 1.27 M ha. About 75% of the National Bank's disbursement of refinance for minor irrigation was to the small farmers (Annex 1 Table 11). 2.08 The state-vise distribution of investments in respect of minor irrigation indicates that the states of Andhra Pradesh, Madhya Pradesh, Maharashtra and Uttar Pradesh accounted for about 61% of the total disburse- ments by the National Bank under non-IRDP schemes. In the case of IRDP, the states of Orissa, Andhra Pradesh, Uttar Prade.h, Bihar and Maharashtra in their order of importance) accounted for 81% of the National Bank's disburse- ments for minor irrigation. In the nine less developed states, the share of minor irrigation disbursements by National Bank was 57%. 2.09 A number of steps were taken by ertswhile ARDC and National Bank in the minor irrigation discipline during the project implemencation period. These included measures aimed at increasing efficiency of the pumpsets, relaxation for issue of loans by SLDBs to SEBs for energization of pumpsets, and setting up of standing committees at the regional levels to review, periodically, the unit costs of various items of investment. In view of power shortage in certain areas of the country, financial assistance has been extended to farmers who were already in possession of electric pumpsets, to install diesel pumpsets of up to 5 hp as standby arrangements. The details of measures initiated by National Bank in the sphere of energization of pumpsets are indicated in Annex II. Quality Control of Agricultural Pumpsets 2.10 In the first phase of studies undertaken during ARDC III, a publica- tion entitled "Pilot Project Studies for Quality Control of Agricultural Pumpsets" was published in December 1980. Under Phase II of the program and in compliance with Section 3.04 of ARDC IV, similar studies were taken up in nine states of Assam, Gujarat, Haryana, Karnataka, Kerala, Madhya Pradesh, Orissa, Tamil Nadu and West Bengal. The objectives of these studies were similar to those of the first phase studies, viz.: (1) to study the efficiency of a few existing pumpset installation and (2) to prepare a block/area-wise matrix based on hydrological and agronomic aspects for proper selection of agricultural pumpsets in future. The main findings of these studies are as follows: (1) improper pumpset installations resulting in high suction Lift; (2) improper selection of suction and delivery pipe; (3) installation of oversize diesel engines/electric motors; and (4) inadequate maintenance of the equipment/installation.  -21- Loans for Effecting Improvements in Pumpset Installations 2.11 With a view to improving the efficiency of the agricultural pumpseets, a decision was taken to extend refinance facilities for providing a sup- pLementary loan to the farmers for replaceent of undersized/defective piping system, foot valve and correcting installation/other defects in existing puapsets (both electric and diesel). This facility was made available to farmers who had installed pumpsets earlier, with or without bank loan. Groundwater Evaluation 2.12 The primary unit of planning in India is a 'block'. Groundwater resource evaluation under AB.DC IV was carried out by the respective state groundwater department on a block-wise basis and they were categorized as white, grey or dark depending upon the extent of groundwater development in each block. The blocks are categorized as white, grey and dark as per "guidelines for evaluation of groundwater resources" given in Annex I to Schedule l of the Project Agreement. According to the estimates made by the state groundwater departments, the groundwater balance on all India basis has been estimated at 35.86 H ha m of which about 9.79 M ha m are estimated to have been developed by 1982 leaving 26.07 M ha m (72.7Z still available for future development). Although considerable scope exists for groundwater development in most of the states, the states of Haryana, Punjab, Tamil Nadu have reached a relatively high degree of groundwater development. These estimates are, however, subject to revision keeping in consideracicn a large number of surface irrigation schemes under execution in these states and adoption of refixed norms for groundwater resource evaluation. 2.13 At the initiative of National Bank, the State Governments of Punjab, Haryana, Karnataka and West Bengal have considerably strengthened the staff position in their groundwater directorates. The control on development of groundwater is exercised only in respect of such investments as are taken up with the help of bank loans. Depending upon the availability of groundwater resources and theextent of mutual interference of minor irrigation works, the spacing between these works (as are supported by institutional finance) is decided, and control on development of groundwater made effective. Number of additional minor irrigation units to be financed by banks are cleared in accordance with the availability of groundwater resources. In respect of the 'dark' blocks, the concerned State Governments have been requested to arrange for micro level surveys to firmly establish the availability of groundwater resources before minor irrigation schemes are cleared by National Bank. 2.14 The second control relates to spacing between two minor irrigation works determined as per pump tests data collected by the State Groundwater Development. The spacing between two-minor irrigation works is kept equal to twice the radius of the influence. At present the minimum spacing between dugwells with pumpsets and shallow tubewells is fixed at 150 m and 180 m respectively. These spacing criteria are stipulated by NABARD as part of  -22- terms and conditions to be observed by banks while sanctioning individual loans. The above controls have worked very well in ensuring planned develop- ment of groundwater resources in the country (see Annex 1, Table 7). 2.15 In terms of Section 3.05 of the Project Agreement, a Model Compensa- tion Scheme for Failed Wells have been drawn up by NABARD and circulated to all the State Governments and participating institutions for adoption. The 'Scheme' has been adopted by the Governments of Karnataka, Maharashtra, Kerala and Tripura. The matter is being vigorously pursued with the other State Governments. Land Develooment 2.16 The total cost of land development financed under the project amounted to Rs 122.28 M (Rs 101.6 M in respect of non-IRDP schemes) which constituted about 39Z of appraisal estimate of Rs 317 M. The major reasons for low disbursements for land developments were high initial cost of land development works, ineligibility of some of the farmers to avail of loans, and reluctance to undertake the work due to loss of crop income. The disbur- sement of refinance for this activity, under the project, amounted to Rs 91.4 M and Rs 13.2 M under non-IRDP schemes and IRD Program respectively. Total number of schemes sanctioned was 298 with refinance commitment of Rs 417.3 M (excluding IRDP). The State-wise analysis of disbursements under Land Development, including disbursements for this item of development made under IRDP, indicates that 24% of the invesement was made in less developed States. Andhra Pradesh and Punjab accounted for about 67% of the disbursement of refinance under non-IRDP schemes while under IaD Program the states of Bihar, Hi