India - National Bank For Agriculture and Rural Development (NABARD) Credit Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 5594-IN STAFF APPRAI SAL REPORT I NDIA NABARD CREDIT PROJECT January 24, 1986 South Asia Projects Department General Agriculture Division II This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EOUIVALENTS (As applied in the report) US$1.00 = Rs 12.00 Rs 1.00 = US$0.083 Rs 1 million = US$83,333 WEIGHTS AND MEASURES Metric System 1 hectare (ha) = 2.47 acres = 10,000 M2 1 ton = 1,000 kg CROPPING SEASONS Kharif - June to September Rabi - October to February Summer - March to May FISCAL YEAR Government of India - April 1 to March 31 NABARD and cooperative banks - July 1 to June 30 Commercial Banks - January 1 to December 31 ABBREVIATIONS AND ACRONYMS ACD - Agriculture Credit Department (of the Reserve Bank of India) ARC - Agricultural Refinance Corporation ARDC - Agricultural Refinance and Development Corporation CB - Commercial Bank DCCB - District Central Cooperative Bank DICGC - Deposit Insurance and Credit Guarantee Corporation DRDA - District Rural Development Agency EC - European Community FCI - Food Corporation of India GDP - Gross Domestic Product GOI - Government of India HYV - High-Yielding Variety IDA - International Development Association IRDP - Integrated Rural Development Program LDB - Land Development Bank NABARD - National Bank for Agriculture and Rural Development NCDC - National Cooperative Development Corporation PACS - Primary Agricultural Credit Societies PLDB - Primary Land Development Bank Credit Review - Agriculture Credit Review to be financed under the project RBI - Reserve Bank of India RRB - Regional Rural Bank SBI - State Bank of India SCB - State Cooperative Bank SEG - Senior Expert Group SGO - State Groundwater Organization SOE - Statements of Expenditures SWC - State Warehousing Corporation  FOR oMCIAL USE ONLY INDIA NABARD CREDIT PROJECT Table of Contents Page No. LOAN AND PROJECT SUMMARY ...................................... -iv I. THE AGRICULTURE SECTOR ........................................ Introduction .................................................. 1 Development Patterns by Region .... , , , ., 2 Role of Agricultural Credit ................................... 3 Historical Development ......................., 3 Short-Term Credit from Institutional Sources .................. 4 Term Credit from Institutional Sources ........................ 4 II. AGRICULTURAL CREDIT INSTITUTIONS .............................. 5 A. Short-Term Cooperative Credit Structure ................... 5 B. Land Development Banks ............................, 6 C. Commercial Banks .......................................... 7 D. Regional Rural Banks .........................., 8 E. Staff Training ............................................ 8 F. Integrated Rural Development Program ...................... 9 G. Credit Recovery and Eligibility Criteria . . 11 Overview .. 11 Loan Recovery of Land Development Banks . . 11 Loan Recovery of Regional Rural Banks . . 12 Loan Recovery of Commercial Banks .. 12 Past Measures to Improve Loan Recovery ..,.. . 12 Rehabilitation ...... 14 III. PERFORMANCE OF BANK GROUP FINANCED AGRICULTURAL CREDIT PROJECTS IN INDIA ................. 15 Introduction ................. 15 Results of State Based and ARDC Credit Projects .............. 16 This report is based on the findings of an appraisal mission that visited India during December, 1984, consisting of Messrs. G. van Santen, A. Haji, S. Lieberman, B. Argyle, J. Manickavasagam and M. Baxter (Bank); R. Stoner (Consultant) and P. Vorage (representative of the Netherlands bilateral aid who participated full time in the appraisal). Mr. M. Virdy contributed to the report. The word processing was done by Ms. B. Bolden, Ms. H. Pratt, with the assistance of Ms. G. Burnett, Ms. C. Doden and Ms. A. Miinott. This document has a restricted distribution and may be used by recipients only in the performance of their offca duties. Its contents may not otherwise be discksed without World Bank authorizatimn.  -ii- Page No. IV. THE PROJECT ................................................... 17 A. Development Strategy ...................................... 17 B. Project Genesis ................................... 19 C. Brief Description and Objectives .......................... 19 D. Detailed Features ......................................... 20 Agricultural Credit Review ................................ 2C Pilot Scheme for Strengthening Agricultural Credit Delivery System,.................,. 21 NABARD's Training Program ................................. 23 RABARD-s Refinancing Program .............................. 23 Minor Irrigation .......................................... 23 Dryland Farming ........................................... 25 Land Development and Soil Conservation .................... 26 Livestock ................................................. 26 Plantations and Horticulture .............................. 27 Forestry .................................................. 27 Fisheries ........ 28 Storage ........ 28 Markets ..... 29 Farm Mechanization ........................................ 29 Non-Farm Rural Activities ................................. 30 E. Project Costs ............................................ 30 F. Financing ................................................. 32 G. Procurement ............................................... 33 H. Disbursements ............................................. 34 V. PROJECT IMPLEMENTATION ........................................ 35 A. The National Bank for Agriculture and Rural Development ......................................... 35 B. NABARD's Lending Procedures ............................... 42 C. Lending Terms and Conditions .........................,. 43 D. Lending to Small Farmers .................................. 44 E. Accounts and Audit ........................................ 44 F. Monitoring and Evaluation ................................. 45 VI. PRODUCTION, MARKETING. PRICES AND FINANCIAL RETURNS TO PROJECT BENEFICIARIES ........................... 46 Production .................................................... 46 Marketing and Prices .......................................... 47 Financial Analysis .................... 47  -iii- PaRe No. VII. BENEFITS, RISKS, AND JUSTIFICATION ............................ 48 Benefits .....................,,.48 Beneficiaries .................. . 48 Environmental E'fects ........... . 49 Economic Analysis .. 49 Project Risks .. 49 VIII. ASSURANCES AND RECOMMENDATIONS ................................ 50 SCHEDULES AND ANNEXES Schedule A Lending Terms and Conditions Schedule B Schedule of Disbursement Tranches Schedule C Criteria for Eligibility of Financial Institutions to Receive Refinancing from NABARD Annex 1, Table 1 A. Summary of Indicated Costs for Pilot Project Table 2 B. Nabard Head Office Costs for Pilot Project Table 3 C. Model Cost Table for One Regional Office (NABARD) Table 4 D. Model Cost Table for One District Table 5 Agriculture Credit Review - Cost Estimates Table 6 NABARD's Expected Expenditure on Training Programs - 1986/87 to 1988189 Table 7 Estimated Disbursements Annex 2, Table 1 Investment Models - Financial and Economic Rate of Return Table 2 Investment Models - Sensitivity Analysis Annex 3, Table 1 Summary Balance Sheets - ARDC/NABARD (1977-1985) Table 2 Summary Statement of Income and Expenditure - ARDC/NABARD (1977-1985) Table 3 ARDCINABARD Disbursements by Subsector (FY1980/81-FY1984/85) Table 4 NABARD's Estimated Resource Position for Term Lending FY86/87 - FY88/89 Table 5 NABARD's Expected Refinancing Program FY86/87 - FY88/89 Annex 4 Agricultural Credit Review, Draft Terms of Reference Annex 5, Table 1 Analysis of LDB Demand and Recovery (1981/82-1983/84) Table 2 Summary of Land Development Bank Recoveries (1974-1984) Table 3 Recovery Performance of LDB Branches (as of 6/30/84) Table 4 Recovery Status of RRBs as of 6/30/83 Table 5 Recovery of Agricultural Loans of Commercial Banks by Region/States (1975-1981) Table 6 Demand and Collection of Agricultural Loans of Commercial Banks (1979-1982) Annex 6 Documents in Project File  -iv- CHARTS 1. NABARD Organization Chart of Head Office (World Bank 27140) 2. Flow of Agricultural Credit in India (World Bank 27141) 3. Major Institutional Funding for Agricultural Term Loans (World Bank 27142) MAP World Bank 18832  INDIA NABARD CREDIT PROJECT Loan and Project Summary Borrower: Government of India Beneficiaries: NABARD, for onlending to farmers for on-farm investments through participating banks. Amount: US$375 million (equivalent) Terms: Repayment over 20 years, including 5 years' grace at the standard, variable rate. Onlendnj (i) GOI to NABRED: Interest would be 6.5% (net of a Terms: reduction of 0.25Z for prompt repayment) per annum in respect of NABARD refinancing of terms loans of up to nine years' duration and 7.0% (net of a reduction of 0.25Z for prompt repayment) in respect of NABARD refinancing of term loans of up to 15 years duration. Repayment of principal in lump sums at the end of 9 or 15 years, respec- tively. (ii) NABARD to Participatinjt Banks: (a) For small farmers, 1/ IRDP loans, and all loans for minor irrigation and land develop- ment, not less than: 6.5% (b) All others, not less than: 8.0Z. Repayments to coincide approximately with collec- tions from ultimate borrowers but not to exceed two years beyond initial amortization schedules specified in loans to ultimate borrowers refinanced by NABARD. (iii) Participating Banks to Ultimate Borrowers: (a) For small farmers, IRDP loans, and all loans for minor irrigation and land development, not less than: 10.0%. (b) All others, not less than: 12.5%. Repayments are based on borrowers - repayment capacity and types of investments financed. jj Small farmers are defined as those cultivating land that provides an annual pre-development net return to the farmer and his family of not more than Rs 2,000 based on 1972 prices.  Proiect Description: The purpose of the project is to help increase agricultural production through particiption in KABARD's ongoing loan refinancing program for investments in agriculture, and to strengthen the agricultural credit system. The project includes a component for extending the existing pilot scheme for branch-level training and farm-level credit delivery systews, along with an extensive training program component for the management and operational staff of NABARD, the Land Development Banks (LDBs), commercial banks (CBs) and Regional Rural Banks (RRBs). A thorough review of the workings of the credit system would also be financed under the project. There is a risk that the more strict criteria governing branch access to NABARD refinance may not necessarily induce such branches to take the measures required to improve credit discipline. A series of covenants are designed to have NABARD take a more active role in the rehabilitation of branches to minimize this risk. Secondly, there is a risk of continued general deterioration of loan recoveries in the agriculture credit system. However, the project is designed to progressively reduce NABARD refinancing of poor-performing branches and under either of the above risk scenarios, IBRD funds would not be at risk, as repayments by participating banks to NABARD are generally made on time and NABARD's own financial position remains very sound.  -lll- Estimated Cost: 1/ (US$ millions) Item Local Foreisn Total (a) NABARD Refinancing (i) Schematic Loans: - Minor irrigation and land development 1,413.3 157.0 1,570.4 - Farm mechanization 667.4 117.8 785.2 - IRDP and other schemes 1,522.3 97.2 1,619.4 - Other (including dairy livestock, plantations, fisheries, forestry) 885.8 46.6 932.4 Subtotal 4,488.8 418.6 4,907.4 (ii) Nonschematic Loans 2/ 330.6 82.6 413.2 (b) Pilot Scheme 22.5 2.5 25.0 (c) Agricultural Credit Review 1.6 1.9 3.5 (d) NABARD's Training Programs 5.2 1.3 6.5 Total Project Cost 4,848.7 506.9 5,355.6 ;_=== c = ~==== = = Financing Plan: IBRD 375.0 375.0 Dutch Government 5.5 5.5 KFW - 21.2 21.2 Ultimate Borrowers' equity 426.8 105.2 532.0 NABARD and GOI resources 3,464.2 - 3,464.2 Participating Banks' contributions 957.7 - 957.7 Total 4,848.7 506.9 5,355.6 1/ Excluding taxes and duties. ZJ Comprising loans for crop storage, input stocks, and rural elec- trification.  -lv- Estimated Disbursements: (US$ millions) FY87 FY88 FY89 FY90 (a) RefinancinR First Tranche 100.0 - Second Tranche - 80.0 - - Third Tranche - 80.0 - Fourth Tranche - - - 80.0 (b) Other Disbursements 9.0 15.0 9.5 1.5 Cumulative Total 109.0 204.0 293.5 375.0  INDIA NABARD CREDIT PROJECT I. THE AGRICULTURE SECTOR Introduction 1.01 Agriculture remains the mainstay of the Indian economy. Currently, farming and related activities account for 35% of Gross Domestic Product (GDP) and 29% of exports; agriculture employs 70% of the labor force. Agricultural growth has kept pace with rising domestic demand, thereby enab- ling the country to eliminate foodgrain imports. Agricultural output rose by 2.6% per annum between the late 1960s and the late 1970s, and increased to 2.8Z since the late 1970s. In 1983/84 the combined output of wheat, rice, coarse cereals, and pulses was 152 million tons, 16% higher than the average production level of foodgrains in the previous three years. 1.02 India's agricultural gains over the past 15 years are due, in large part, to the diffusion of high-yielding varieties (HYVs) of hybrid cereals and to the more extensive application of fertilizers and otl.-r inputs. By 1983/84, fertilizer-responsive HYVs, initially introduced iu India in the mid-1960s, were used on more than half the area devoted to cereal production. Fertilizer consumption, which averaged 1.8 million tons in 1967/70, rose to 7.8 million tons in 1983/84. Similarly, purchases of pesticides rose sharply during the 1970s and early 1980s. Overall, the estimated value of all inputs (including labor) rose in real terms from Rs 49 billion in 1970/71 to Rs 74 billion in 1981/82. Expenditure on cash inputs (including hired labor) by public agencies and private individuals grew even faster. 1.03 The agricultural sector has also absorbed high levels of capital investment by public agencies and private individuals since 1970. This has substantially enlarged the area devoted to irrigated crops. The irrigation potential created by publicly constructed dams and canals rose from 18 mil- lion hectares (ha) in 1968/69 to more than 30 million ha in 1983/84; land brought under irrigation, mostly by private expenditure, increased from 19 million ha in 1968/69 to 34 million ha in 1982/83. Overall, gross agricul- tural capital formulation is estimated to have risen in real terms from Rs 13 billion in 1970/71 to Rs 21 billion in 1981/82. Public outlays as a share of total agricultural investment increased from 25% in 1970/71 to 32% in 1981/82. Privately executed and largely farm-based investments such as minor irrigation works, pumpsets, tractors, and livestock increased in absolute terms, but less rapidly than public investments.  -2- Development Patterns by Region 1.04 The scope and quality of investments has varied considerably among regions since the mid-1950s. A geographically differentiated pattern of investment accounts for much of the variation in the pace of agricultural expansion at the State level. Growth has been most rapid in the States of Punjab and Haryana and the lowest in the eastern region of the country. For example, the combined output of foodgrains in Bihar, Orissa and West Bengal grew by a total of 3% between the late 1960s and the early 1980s. Growth rates in other regions have been more in line with all-India trends; only the average annual growth rate of about 2% in foodgrains production in the southern tier, comprising Andhra Pradesh, Karnataka, Kerala and Tamil Nadu was below average. 1.05 The high growth rates of between 3.8% and 5.8% annually attained in Punjab and Haryana are attributable, in part, to the large program of public investment, complemented by extensive farm-level investments. Public resources have been applied to major irrigation and power projects, rural roads, and marketing facilities, and also to the implementation of effective land reform, a reorganization of the cooperative movement, and strengthening of agricultural research and extension. Cultivators have followed up with private investments to improve farm production, control the timing of field operations, and increase the application of inputs needed for successful utilization of HYVs. Household savings in these States have substantially increased and have been used, to a considerable extent, to finance continuous improvements in agricultural production capacity. 1.06 Comparable growth has been achieved in those districts of Rajasthan and Uttar Pradesh that are adjacent to the Punjab and Haryana, and in selected districts of Andhra Pradesh, Gujarat, Maharashtra and Tamil Nadu. Elsewhere, agricultural performance has been poor either because of under-utilization of facilities created through public investment or because of the absence, thus far, of concerted public initiatives in the agricultural sector. Growth rates have been disappointing in the southern region, despite extensive reliance on irrigation and above-average use of fertilizer on a per hectare basis. Utilization of water has been unsatisfactory in western India -here considerable public resources have been invested in irrigation facilities. However, agricultural growth in this region has been stimulated by the widespread dissemination of improved varieties of coarse grains. In the eastern region, reliance on irrigation and use of fertilizers and HYVs is limited, despite the considerable potential for surface and groundwater irrigation. Farm-level constraints, such as the small average size and fragmentation of holdings, make it difficult for cultivators to invest in tubewells and dugwells. Farmers reluctance to invest in improvements is linked to poorly funded and badly executed state-level initiatives on land reform and production.  -3- Role of Agricultural Credit 1.07 In addition to direct public investment in irrigation facilities and other productive infrastructure, the Government of India (GOI) h.s fostered private investment in agriculture. These efforts since the 1960s, have increasingly centered on the expansion of the institutional credit system. Two distinct objectives have underpinned GOI ambitions: First, institutional credit disbursements have been seen as a primary tool in the campaign to eradicate rural poverty. The Government's view is that credit advances can free cultivators from exploitative economic relationships and provide the asset base needed by them to become self-sustaining producers. The most recent manifestation of this view is the Integrated Rural Develop- ment Program (IRDP) intended for small and marginal farmers and landless workers (paras 2.13-2.15). 1.08 Second, the role of institutional finance ir rural areas is inter- twined with the national goal of achieving self-sufficiency in food produc- tion. Institutional credit is regarded as a critical means of improving agricultural production and productivity by facilitating adoption of new agricultural techniques and inputs and by encouraging a shift to irrigated cropping. Based on the estimates of the All-India Rural Credit Reviev Commit- tee, which forecasted the amount of institutional credit needed to reach "required" investment levels, the Sixth Five Year Plan (FY79/80 - FY84/85) set a target of Rs 40 billion of short-term lending and Rs 14 billion of (medium- and long-term) lending for 1984/85. The Seventh Five Year Plan proposes more ambitious goals - Rs 80 billion of short-term lending and Rs 45 billion of term lending for 1989/90. Historical Development 1.09 GOI-s ambitions with respect to rural credit were initially pur- sued through expansion of the cooperative system which, in the mid-1950s, comprised some 110,000 Primary Agricultural Credit Societies (PACSs), with links to 509 District Central Cooperative Banks (DCCBs) and 16 State Cooperative Banks (SCBs), all lending for short-term purposes; for term-lending, there were seven State Land Mortgage Banks, which advanced credit mainly for redemption of non-institutional debt, and several Land vevelopment Banks (LDBs). The State Land Mortgage Banks later converted into LDBs, while additional States introduced LDBs by 1969. Dissatisfaction with the volume and quality of lending through cooperative institutions in the early 1970s led to measures by GOI to involve Commercial Banks (CBs) in agricultural lending (paras 2.07-2.10). Expanded CB financing to agriculture was induced by GOI by stipulating minimum lending levels for priority sec- tors, including agriculture, and was supported by licensing measures, which further induced CBs to increase the number of rural branches from 5,000 in 1969 to 40,000 in 1984. The rural credit system was further expanded with the creation, in 1975, of Regional Rural Banks (RRBs) (para 2.11). Institu- tional credit operations in rural areas are regulated by the Reserve Bank of India (RBI), which ensures that lending terms available through different channels are compatible with the structure of interest rates set out in government directives. Until 1982, the Agricultural Credit Department (ACD)  -4- of RBI was the main source of refinance for short-term cooperative credit institutions, vhile the Agricultural Refinance and Development Corporation (ARDC) served as the primary means of refinancing term lending. In 1982, the National Bank for Agriculture and Rural Development (NABARD) was created, and it assumed the refinancing functions of ARDC and ACD (see Chapter V). Short-Term Credit from Institutional Sources 1.10 Short-term advances rose rapidly in nominal and in real terms between 1974/75 and 1978/79. Cooperative disbursements (including RRBs) rose 60X to Rs 12.1 billion and CB advances more than doubled to Rs 3.6 billion. During this period, short-term disbursement accounted for an increasing proportion of expenditures on agricultural inputs. However, thereafter, short-term credit flows (in real terms) first declined and then remained roughly at the level that prevailed in 1977/78, with CB adiances continuing to grow, reaching Rs 7.0 billion in 1981/82 and 1982/83, vhile cooperative disbursements declined in real terms. Despite the lack of growth of short-term advances in real terms, expenditure on purchased inputs continued to grow rapidly during the late 1970s and early 1980s, as cultivators were able to find other means, including their savings, to finance their working capital requirements. Term Credit from Institutional Sources 1.11 Institutionally funded investment in agriculture increased nearly threefold in real terms between 1971/72 and 1980/81, while the proportion of institutional lending in private agricultural investment increased by one-third during this period, reaching 442 in 1980/81. Hovever, recent investments funded through institutional sources may yield lower returns than would have accrued if investments had been funded through private savings alone. IRDP (para 2.13) has garnered a growing proportion of agricultural lending, and recent studies highlight the low returns and high administrative costs associated with IRDP lending. Also, large allocations of institutional finance to minor irrigation investments in non alluvial area (para 4.21) are likely to be characterized by low marginal economic benefits. 1.12 The efficiency of institutional credit for term lending needs to be assessed with regard to how much financing has augmented or, in some situations, displaced private savings in different regions. Large institu- tional credit disbursements in Punjab and Haryana have been combined with substantial private savings to finance the continuing increases in productive capacity. Allocations to IRDP in these States account for a small proportion of institutional credit advances. By contrast, the growth of agricultural production has remained relatively low in the southern tier of States, despite significant financial disbursements including a large share to IRDP, and substantial public investments. Private savings in this region where average returns to farming investments are low might have been drawn to non-agricultural opportunities, leaving agricultural investment to be financed institutionally. In the eastern region, agricultural investment is low and suffers from an inappropriate mix of investment outlays. Rural  -5- financial institutions in Bihar and to a lesser extent, Orissa have con- centrated more on farm mechanization schemes for larger farmers than on financing minor irrigation projects. IRDP accounts for more than 202 of term credit disbursed in Bihar and Orissa. The relative share of the institutions of NABARD's term credit disbursements has been as follows: Share of ARDC/NABARD Disbursements for Term Investments FY74175 FY79/80 FY83/84 FY84/85 Agency Rs M Z Rs M 2 Rs M x Rs M Z LDBs 771 72 1,640 40 3,140 35 3,140 30 CBs 1/ 279 26 2,300 56 4,500 50 5,700 54 RRBs - 90 2 870 10 1,400 13 SCBs 15 2 90 2 410 5 370 3 Total 1,065 100 4,120 100 8,920 100 10,610 100 1/ CBs- share of financing in agriculture is likely to be higher, because CBs finance up to 60% of their agricultural investments from their own resources without NABARD refinance. II. AGRICULTURAL CREDIT INSTITUTIONS A. Short-Term Cooperative Credit Structure 2.01 The short-term cooperative banking structure in each State com- prises a three-tiered structure with the SCB at the apex, serving DCCBs at the district level, and PACSs being the primary lending units. In 1984, the 27 SCBs, 337 DCCBs, and 94,000 PACSs had the biggest network for short-term lending. All DCCBs within a State are shareholders of the apex SCB and provide short-term credit to village-level PACSs, based on limits prescribed by RBI. Besides serving as bankers to their member DCCBs, SCBs conduct general banking business, arrange for statutory inspection of DCCBs and provide some training facilities. SCBs and DCCBs compete with CBs and RRBs for deposits and are the major institutional source of short-term credit to 'armers. 2.02 At the village level PACSs provide members with short-term credit and marketing services, and since their functioning expanded in 1969, have also provided a number of diversified services, such as stocking and retail- ing agricultural inputs, and foodgrain purchases. Recently, some PACSs have also collected savings deposits from members and offered consumer loans. In many States, the credit performance of PACSs has been inadequate, due to weak management, inadequate supervision and ineffective administrative control, a limited resource base, and narrowly focussed services. Rooted in the local village power structure, PACSs are vulnerable to local pressures. Since  -6- 1973, GOI has encouraged amalgamation of PACSs to improve their operations, management and profitability. Even profitable PACSs are often unable to appraise and supervise a large volume of loans adequately. As a result, the quality of lending is poor, and advances often fail to satisfy farmers' real needs. The level of overdues is higher among PACSs than among DCCBs and SCBs; repayment of PACSs to DCCBs and from DCCBs to SCBs is within reasonable limits; in most States SCB recovery is 90% of demand. 2.03 SCB-s short-term borrowings from NABARD usually constitute up to 50% of the funds advanced to DCCBs. SCBs also provide medium-term finance, particularly through IRDP. SCB's medium-term credit as a share of agricul- tural credit has grown substantially in recent years, although from a very small base; 5Z of NABARD term refinance went to SCBs in FY83184. Although SCBs have been increasing their term lending, more than 952 of medium- and long-term lending in agriculture is shared among CBs, LDBs and RRBs, with NABARD providing refinance to all institutions. B. Land Development Banks (LDBs) 2.04 In 1984, 19 States had LDBs; eight LDBs lend through their own branches (1,007), while 11 LDBs refinance 889 affiliated Primary Land Development Banks (PLDBs), which, in turn, lend through branches. LDBs only provide 5-15 year loans for agricultural investments; they do not make seasonal loans, do not accept deposits, nor lend for non-agricultural pur- poses. LDBs service more than six million accounts. About 30% of LDB share capital is owned by the respective State Government, and the balance by cooperatives and farmers, who are normally required to deposit 5-10% of the loan amount in their share accounts for the duration of the loan. LDBs and PLDBs are managed by Boards of Directors, partly elected and partly State-nominated, but State Governments may and do suspend the boards, some- times for political reasons. The degree of control exercised by State Governments over LDBs varies widely. 2.05 Total LDB lending has more than doubled in nominal terms from FY79180 to FY83/84 (Rs 2.4 billion and Rs 5.7 billion respectively), but the LDB annual share of NABARD refinance has dropped from 40% in 1979/80 to 30% in 1984/85. LDBs lend predominantly for minor irrigation investments and farm mechanization. The main sources of LDB funds are: (i) special develop- ment debentures, which it sells to NABARD and State Governments, guaranteed by the latter, bearing 6.5% - 8.0% interest, and retired in annual install- ments over actual maturities (5 - 15 years) of loans to which they relate; and (ii) ordinary debentures, floated on the wider institutional market at a similar rate, guaranteed by the State Government and repayable in full at maturity. The share of NABARD special debentures of LDBs new borrowings decreased from about 70% in FY79/80 to 55% in FY82/83. 2.06 The performa-nce of a number of LDBs since 1970 has been less than satisfactory. Most visible has been the poor recovery performance of some banks, as detailed in para 2.17. Low recoveries reflect a generally low quality of lending. LDB operating costs, with branches often covering very large areas compared to those of other institutions, are generally high.  -7- With interest rates and margins determined or. an all-India basis by RBI, and with no flexibility to diversify their sources of funds or engage in more profitable lending, LDBs face strong pressures to reduce their operational costs. As a result, contacts between borrowers and LDB field staff are less frequent than desirable. Some LDBs have up to 75% of their loan applications and appraisals processed by State agricultural extension agents, who are paid a fee for each submitted application. A overwhelming array of operational directives also limit staff responsibilitv. Although chartered as independ- ent cooperatives, LDBs are sometimes treated as if part of State Government with Government staff normally posted to top management positions and fre- quently changed. Not all LDBs are so affected however; the Punjab and Haryana LDBs are strong, sound, and vigorous organizations with very low overdues. C. Commercial Banks (CBs) 2.07 CBs are comprised of 28 public sector banks, nationalized in 1969, and 38 private sector banks. The public sector group consists of the State Bank of India (SBI) with seven subsidiaries, and 20 other banks. With more than 9,600 branches, SBI accounts for about 50% of the agricultural loans made by CBs. The other public sector CBs have also well developed branch networks in most parts of India, while private CBs operate on a much smaller scale. The amount outstanding in agricultural loans by CBs increased from Rs 2 billion and 270,000 accounts in June 1969 to Rs 53 billion and 12 mil- lion accounts by March 1982. Of the total outstanding about 142 consisted of "indirect lending" for crop storage, input stocks, and rural electrification. The remainder (86Z) was for direct lending to farmers of which about 55% were medium and long-term loans and the balance crop production loans. The north- eastern States are the smallest recipients of CB loans, accounting for less than 1%. 2.08 Refinance to CBs from ARDC/NABARD increased from Rs 0.3 billion during FY69-74 period to Rs 5.7 billion in FY84/85. The CB share of ARDC/NABARD refinance, which peaked at 56Z in FY79/80, amounted to 50X in rY83/84, and 54Z in 1984/85. Deposits of all CBs grew tenfold frcm Rs 46 billion in 1969 to Rs 458 billion in FY82/83. 2.09 All CB loans to agriculture and small industries are insured by the Deposit Insurance and Credit Guarantee Corporation (DICGC), to the extent of 75% of the outstanding overdues. LDBs can join these schemes only if they have less than 25% overdues; very few LDBs qualify. Claims paid by DICGC are growing very rapidly and amounted to Rs 0.6 billion in 1983, of which 18% were for agricultural loans. The rapid growth of the CB agricultural loan portfolio and continuing high levels of overdues (para 2.20) and an explosive growth of small industries claims have raised concerns about DICGC's ability to meet future claims. DICGC is reviewing whether it should lower the per- centage of the loan it guarantees and/or increase the fee it receives for its services (currently 0.5% annually of loan amount outstanding for small farmers and 0.75% for bigger farmers).  -8- 2.10 The main advantages of CBs over LDBs are: Ci) their ability to provide short- and long-term credit; (ii) greater flexibility in lending procedures and collateral requirements, as CBs increasingly provide loans against hypothecation of assets created with the loan, which LDBs cannet; (iii) their much stronger financial position allows some CBs to offer better terms of service and, thus, attract more qualified technical staff; and (iv) a two-way relationship with clients, who normally are depositors as well as borrowers. Their disadvantages are relative newness to rural operations and the urban orientation of many CB staff. D. Regional Rural Banks (RRBs) 2.11 RRBs provide all banking services in rural areas and each RRB usually confines operations to one or two districts. An RRB is sponsored by a CB that contributed 15% of the share capital and provides management assis- tance; GOI and the State Government contributed 50% and 35%, respectively, of its share capital. Within India's multi-agency agricultural credit system, RRBs were specifically established to lend to the weaker sections of the population, those not having access to CBs. First established in 1975, the number of RRBs has grown to 170 with 8,200 branches in 1984, covering 270 of India's 411 districts. In FY83/84, the RRB loan portfolio totalled Rs 4.7 billion, and RRBs received 10% of NABARD's refinance, compared to 2% in FY79/80. RRB deposits stood at Rs 4.2 billion in FY83/84. E. Staff Training 2.12 The huge size of the agricultural credit system dictates that financial institutions are responsible for their own staff training. The CBs, particularly the SBI group, have the largest training capacity. They operate 27 staff colleges for their own long-term professional staff training needs and 117 regional centers for lower-level staff courses. In addition, they use RBI's College for Agricultural Bankers for technical subjects. Institutions of a sponsoring CB largely train RRB professional staff. The rapid growth of RRBs has necessitated additional training facilities: the College for Agricultural Bankers provides courses for senior staff, while NABARD's Bankers institute for Rural Development trains RRB staff only. Training institutes within the cooperative banking system include a manage- ment training institute in Poona, 17 cooperative training colleges for inter- mediate-level training, and 86 cooperative training centers. In addition, LDBs operate 14 junior-level training centers. SBI has the most satisfactory training system; smaller CBs, RRBs and the cooperative banking structure display serious weaknesses, training does not address needs, most courses are academic and contribute little to practical operations. At the same time, the demand for technical training, refresher courses, and practical "hands ou" training in rural banking at all levels far exceeds supply. ARDC/NABARD has been involved in training of staff of other financial institutions since 1975. The Bank Group assisted ARDC series of projects (paras 3.01-3.02) has financed training for 10,500 senior and 23,155 junior staff of CBs, LDBs and RRBs and 2,600 other officials. Financing of NABARD's training programs will continue under this project (para 4.16 below). The present system of staff training raises the question who should be responsible for the quality and  -9- quantity of training, and who should monitor and control training. ARDClNABARD's past involvement in training was to some extent ad-hoc to satisfy immediate training needs, particularly of RRB staff. As detailed in Annex 4, the agricultural credit review would assess training needs of institutions, and develop a strategy and program to strengthen the capacity, effectiveness and efficiency of staff training. It would also review NABARD-s role in providing training to staff of client banks, and determine NABARD-s role in monitoring and improving training quality and quantity. To gain insight into the implications of a modified role, NABARD would initiate and monitor an experimental training program for branch managers of a selec- tion of banks. Aimed at improving the effectiveness of loan recovery and the efficiency of the branch, it would combine training with simultaneous efforts to address institutional factors impeding effective branch operations (paras 5.10-5.12). If necessary, the program could be modified once the agricultural credit review findings become available. F. Integrated Rural Development Program (IRDP) 2.13 IRDP was set up in FY78/79 as an amalgamation of existing rural poverty-alleviation programs. It has become GOI's principal tool in address- ing rural poverty, and together with the National Rural Employment Program, has been accorded very high national priority. The Central and State Govern- ments share in its expenditure; all rural credit institutions participate in its financing. During the Sixth-Plan up to March 1984, about 12.6 million families, including 4.7 million belonging to scheduled castes and tribes, have been assisted, involving a government subsidy of Rs 11.9 billion (US$1.0 billion) and institutional credit amounting to Rs 22.4 billion (US$1.9 billion). Between FY79/80 and FY84/85, the share of NABARD refinance for IRDP investments increased from about 1Z to 33% of annual NABARD disbur- sements for specific schemes. The Seventh Plan emphasizes GOI's continued commitment to poverty alleviation. IRDP focusses on all sectors of the rural economy - agriculture, small industries, and services, and aims at realizing productive investments, in conjunction with employment creation. 2.14 IRDP's target is to identify every year about 600 families below the poverty line (with an annual income of Rs 3,500 or less from all sources) in each of India's 5,011 blocks (subdivision of a district) and to provide these selected beneficiaries with productive investments, which would allow them to cross the poverty line. At the block level, IRDP is executed by a District Rural Development Agency (DRDA), engaging block and village develop- ment officers and extension workers. At the district level, IRDP operations are headed by the Collector who supervises DRDA staff, and coordinated by a district consultative committee, consisting of representatives of the par- ticipating banks, NABARD, DRDA, and the Planning Department of the State Government, which examines annual plans drawn up by DRDA for each block to ensure availability of resources for planned investments, and to ensure the technical and financial viability of the proposed investments. It then prepares a banking plan and allocations by bank of the number and types of investments. These plans are consolidated at the State level by a State  -10- Consultative Committee involving all participating institutions. In for- mulating such plans, GOI gives a priority to encourage agriculture invest- ments with-.n the less developed states. 4] At the national level IRDP is administered by the Ministry of Rural Development. Selection of beneficiaries is based on household surveys conducted by block development officers who also prepare loan applications. Central and State Governments share the 25 - 50% subsidy that each IRDP beneficiary receives; the remaining investment costs are financed by banks. Small and marginal farmers receive 25 - 33% subsidy with a maximum of Rs 3,000 each. Beneficiaries of scheduled tribes receive 502 subsidy with a maximum of Rs 5,000. During FY83/84, 14% of IRDP funds went to minor irrigation, 27% to dairy develop- ment, 27% to other livestock, and 28Z to small businesses and other categories. NABARD can refinance up to 90% of an IRDP loan. However, banks, especially CBs, have historically been the major suppliers of credit under IRDP. In FY83/84, banks disbursed Rs 7.7 billion, of which NABARD refinanced Rs 2.1 billion, or 27%. NABARD expects its share to increase in the future. 2.15 Experience to date indicates that IRDP has reached a substantial number of farmers, nlthough the target over the Sixth Plan Period--3,000 families per block or about 15 million families for the country for the five years--was not met. Evaluations concluded that due to IRDP's rapid growth, insufficient attention was paid to supervision and quality control. As a result, an estimated 20% of IRDP investments have been misused and incremen- tal income of IRDP investments has been much less than planned; fewer than 40% of beneficiaries crossed the poverty line. Studies point up the need for more realistic selection of investments and planning at block and district levels, a more equitable selection of beneficiaries and a more thorough appraisal and supervision of implementation of investments. Since 1980, GOI has become increasingly aware of the need to improve the institutional per- formance of IRDP. However, GOI's heavy political commitment to IRDP's target of nationwide implementation prevents more modest growth of the program. As DRDA and the banks are involved in implementation, with DRDA primarily addressing volume, and banks concerned about lending quality, IRDP is hard pressed to satisfy both. GOI has strengthened coordination at block, dis- trict, and state levels by instituting the consultative committees, improving the capabilities of the DRDA, and streamlining planning procedures. Further efforts are required. Overwhelmed by pressure to reach IRDP's lending tar- gets, the banks, as a group, have not been able to maintain lending quality in the IRDP program. NABARD's involvement in IRDP has focussed on refinanc- ing operations. It will need to play a more active role in assisting and guiding the banks in maintaining appropriate standards of lending quality and strengthening IRDP planning and supervision of implementation at the district level. Measures to strengthen NABARD's role are detailed in para 5.06. Further measures to improve IRDP's operations fall outside the scope of this project and will be pursued as part of the Bank Group's country dialogue. 1/ Specifically, Arunchal Pradesh, Assam, Hanipur, Meghalaya, Mizoram, Nagaland, Tripura, Sikkim, Bihar, Orissa and West Bengal.  -11- G. Credit Recovery and Eli2ibility Criteria Overview 2.16 A high level of overdues (i.e. potential or actual bad debts) ..mong constituent parts of the agricultural credit system is its most visible problem. Reasons for high overdues, include willful default; weather and natural calamities affecting crops; poor functioning of bank branches, par- ticularly inadequate appraisal of investments, insufficient supervision of implementation, follow-up and collection efforts; low interest rates; and in some instances active discouragement of collection efforts by State Governments and local groups for political reasons. Agro-climatic factors complement institutional weaknesses. A state level comparison confirms an inverse relationship between the general level of risk of agricultural opera- tions and the level of loan recoveries. In addition, CBs which have to fulfill the agriculture sector's share of their lending operations do not press established, reliable borrowers for repayment on time, as this would require further lending to naw borrowers. However, the prevalence of a high level of overdues is specifically due to defective lending policies, ineffec- tive field supervision, and a lack of concerted effort and will on the part of bank boards of directors and staff to recover due loans. Recently GOI has stressed the need to improve loan recovery performance requiring the banking system to adopt firm measures to correct the present situation of overdues. Loan Recovery of Land Development Banks 2.17 Between 1974 and 1984, the average percentage of recoveries of all LDBs ranged from a high of about 62% (1974) of total demand 1/ to a low of 48% (1980) (Annex 5, Table 2). The national average conceals wide variations among States; Haryana, Kerala and Punjab consistently maintained high recovery rates, while Bihar, Gujarat, Maharashtra, Tamil Nadu, and West Bengal had long periods of low recovery. But even within States LDB recovery rates vary significantly. Among LDBs with low recoveries in the five States mentioned above, 165 of their total of 876 branches, or 19Z had recoveries of more than 75%, while 396 branches or 39% had recoveries of less than 40% (Annex 5, Table 3). Although LDB branches in some districts have uniform low recoveries, particularly in Maharashtra, West Bengal and to some extent Gujarat, other districts (Tamil Nadu) display a mixture of high and low recoveries among branches within a district. 2.18 Several State Governments provide substantial financial assistance to their LDB (Annex 5, Table 1). During the ARDC IV project, 1/ Demand being defined as principal and interest due for the current year in addition to overdues of previous years.  -12- (Cr. 1209-IN/Ln. 2095-IN), the Bank Group agreed that four LDBs 1/ with high overdues would be allowed to "block" chronically overdue accounts 2/. During FY81/82 to FY83/84, Rs 609 million were "blocked" of which Rs 370 million were in Maharashtra. Due to "blocked" accounts, States paid Rs 162 million in overdue recoveries to the LDBs in FY81/82, Rs 330 million in FY82/83 and Rs 326 million in FY83/84. Without such financial assistance, reported recoveries for all LDBs for FY83/84 would have been less by 8 percentage points; 51% instead of 59%. Loan Recovery of Regional Rural Banks 2.19 As of June 1983, more than 57% of all RRB branches had recoveries of less than 55% (Annex 5, Table 4). High overdues were concentrated in Bihar, Maharashtra, Orissa, Uttar Pradesh and West Bengal, while Andhra Pradesh and Karnataka had a more satisfactory performance with only 25% and 35%, respectively, of their RRB branches having recoveries of less than 55%. Loan Recovery of Commercial Banks 2.20 CB recoveries (on their entire short-term and term agricultural portfolio) have been constant at about 52% of total demand for the past 10 years, during a period of rapid growth (Annex 5, Table 5). SBI has main- tained a higher recovery percentage than the other banks. All banks in the northeastern region have low recoveries. Of the six major States in terms of volume of CB lending, recoveries in Punjab, at about 74 - 80%, have been significantly above the all-India average. Recoveries in Maharashtra, on the other hand, have been consistently below 50%. In general, recoveries on term loans (about 46%) are lower compared to those on short-term loans (about 59%). Past Measures to Improve Loan Recovery 2.21 Since 1975, successive Bank Group-assisted projects--ARDC I-IV (paras 3.01-3.02)--have attempted to improve loan recovery, primarily among LDBs; ARDC IV introduced measures for CBs and RRBs. Key elements of these efforts were various sets of eligibility criteria, initially applied to LDBs and extended to CBs and RRBs in 1983, which imposed restrictions on availability of all funds from NABARD to branches with inadequate recovery. During ARDC III and IV, these eligibility criteria were augmented by specific measures to improve the institutional performance of weak LDBs. Since LDBs receive NABARD refinance for more than 50% of their funds, eligibility criteria directly influence their total amount of lending. These eligibility criteria, details of which have been subject to frequent revision, have 1/ Bihar, Gujarat, Karnataka, and Maharashtra. 2/ Chronic overdues are entered into a "blocked" account and the State Government pays the LDB any shortfall in collecting the dues, generally over a five-year period.  -13- maintained one common characteristic; a branch/PLDB with recovery of more than 75% has unlimited eligibility, while lower levels of recovery qualify for declining amounts of refinance from NABARD. Between October 1978 and December 1981 eligibility criteria in effect excluded branches/PLDBs from refinance unless recovery exceeded 35% 1/. In 1982, to minimize the drastic impact of cutoff of branches with recovery of less than 35%, the eligibility criteria were modified allowing some refinancing to all branches/PLDBs, even to those with minimal recovery. Recoveries improved initially, but have declined gradually since then. The initial improvement was the result of "blocking" of chronic overdues and their exclusion from annual demands and interest remissions, and of the payments made by State Government to cover the shortfall in recoveries of the "blocked" accounts, which were and con- tinue to be counted as recoveries. The deterioration of recovery has occurred, despite rapidly growing State assistance in some States and a higher incidence of rephasing of accounts affected by natural calamities. 2.22 Under ARDC IV, eligibility criteria for CB and RRB branches closely resembled those applicable to LDBs and provided for unlimited eligibility for NABARD refinancing to branches with 652 recovery on their entire (short- and long-term) portfolio 2/. If the parent CB or RRB had an overall 65% recovery rate, their branches qualified for unlimited refinance, regardless of branch performance, as were branches which were in operation for less than five years and very small branches. Below the 65% recovery level, the amount of NABARD refinancing depended on cash recovery in previous years. Effectively, there was no significant constraint to NABARD refinance. All loans made prior to June 1972 and those for which DICGC had disbursed were not included in the calculation of demand and recovery. The impact of eligibility on CBs and, to some extent, on RRBs differed from LDB because: (i) agricultural operations of CBs constitute only 15% of CB operations; (ii) NABARD refinance is one of several sources of of CB and RRB funds; and (iii) CBs are instructed to lend (including rediscounts) the equivalent of 60% of rural deposits for rural purposes. With data on CB recovery not yet available for FY83/84, the impact of these eligibility criteria is uncertain, but appears to have been limited. By law, all banks are not allowed to make public their write-offs and bad-debt reserves. LDBs generally do not write off loans; they assume that all loans are fully secured by adequate mortgages, although in practice execution of such security is often impossible. Through the "blocking" facility, LDBs have come closest to writing off loans. However, instead of creating reserves, State Governments guarantee repayment, which has counted towards recovery. Eligibility criteria alone will not secure the financial integrity of the financial institutions; adequate reserves to cover 1/ 45% minus 10% notional reduction on account of the State Governments contribution to the LDBs' share capital. 7J/ 75% minus 10%, being the notional reduction on account of the State Governments contribution to the LDB share capital, which was equally applied to CBs.  -14- doubtful debts are also required, while bad debts should be removed from their books. Rehabilitation 2.23 Under ARDC III and IV, eligibility criteria were combined with attempts to rehabilitate weak LDBs. Under ARDC III, execution of a rehabilitation program for each of six LDBs with very low recoveries was the responsibility of the LDB concerned, assisted by ARDC; the program attempted more rigorous action toward long-term defaulters, instituted "blocking" of chronic overdues, rephased 1/ loans, and handling of incomplete investments. State Governments were reluctant to agree formally to the rehabilitation programs, although certain components were executed. Hore broadly based rehabilitation efforts were instituted and agreed to by State Governments during ARDC IV for the LDBs of Bihar, Gujarat, Karnataka, Madhya Pradesh, Maharashtra, and Tamil Nadu which collectively accounted for 60% of LDB loan collection demand and 80% of all overdues. These rehabilitation programs, supervised by a standing committee in ARDC, were drawn up by individual groups of banking specialists and focussed on: (i) measures to reduce the demand or increase recovery of the LDB portfolio, including "blocking" of loans, rephasing and rescheduling 2/ (ii) removal of constraints on legal action against defaulters; and (iii) revisions in operational and technical directives, including those concerning personnel and training policies, financial management, auditing, record keeping, and loan appraisal and super- visory guidelines. Although the rehabilitative efforts had some impact on lending procedures, staffing policies, and operational matters, they fell far short of expectations. Their effectiveness was diluted by the relatively short time period in which rehabilitation took place and by insufficient control and guidance on the part of NABARD. Accordingly, only a portion of the programs recommendations has been satisfactorily executed and effective in improving operations. 2.24 In sum, the experience with the LDB rehabilitation programs indi- cates that limited technical changes by themselves will not cause weak institutions to perform better. The broad issues (institutional, lending policies and procedures, credit delivery system and manpower) affecting LDB operations and, to some extent, those of other banks need be addressed in order to create a banking environment conducive to healthy lending operations and growth. They will be addressed by the agricultural credit review (para 4.04). Meanwhile, measures are necessary to prevent lending quality from declining further and to effectively increase the rate of loan recovery by the entire banking system. Although the effectiveness of eligibility criteria for NABARD disbursements varies by institution depending on its need for NABARD funds, only such criteria are available to influence recovery in the short term. To compel weak institutions into action to strengthen jj Spreading repayment for one year over the remaining installments. II Allowing one or more extra year(s) for repayment.  -15- recovery efforts, the eligibility criteria should allow lending units with low levels of loan recovery a very limited access to NABARD refinance. Such eligibility criteria should also provide incentives for lending units to improve their recovery performance and, at the same time, penalize (through further reduction of refinance facility) those units that do not demonstrate improvement. Also, units should ascertain whether the loans have a realistic repayment schedule and should deal with bad and doubtful debts. Assistance from the parent banking institution is uecessary to accomplish these tasks, and such assistance should be guided and supervised closely by NABARD. Efforts to improve recovery and review the loan portfolios of weaker units need be taken quickly. Such decisions should be made in a timely manner because experience shows that sustained deprivation of NABARD funds without specific assistance to improve recovery is ineffective and counterproductive. 2.25 The bank branch is the unit ultimately responsible for loan recovery, and the logical unit to which eligibility criteria should be applied. In the past, an exception was made for PLDBs, small banks with a limited number of branches operating in one district. PLDBs and RRBs, each operating only at the district level, are sufficiently small to have eligibility criteria applied toward their entire portfolio; for other banks, criteria should apply toward the individual branches recovery performance on term-loan portfolios. New eligibility criteria, combined with safeguards to ensure the financial integrity of branches are detailed in paras 5.09-5.12. III. PERFORMANCE OF BANK-GROUP FINANCED AGRICULTURAL CREDIT PROJECTS IN INDIA Introduction 3.01 The Bank Group has been involved in agricultural credit in India since 1969. During the first five years it financed 10 State-based agricul- tural credit projects, covering the country s major farming regions. Funds were exclusively channelled through the Agricultural Refinance Corporation (ARC) to State level LDBs. From 1975 onwards, the Bank Group ceased lending through State-based projects and initiated a nationwide series of projects through ARC's successor, ARDC; each project financed time slices of about two years of ARDC's refinancing program. In addition, ARDC has taken an active part in the formulation and implementation of another 25 completed Bank Group projects that provided financial support for investments in specific agricul- tural subsectors. The amount committed by the Bank Group for these 39 projects totalled US$2.4 billion. 3.02 The first 10 Credits mainly provided funds for State LDBs, and facilitated close supervision of LDB operations, and individual project conditionality tailored to the State credit system. In shifting to a nation- wide approach by providing a general line of credit through ARDC, the Bank Group assumed that ARDC had greater capability to supervise the operations of the participating banks in view of its proximity and local knowledge, and that a line of credit would allow complete coverage of Indian credit, wherever conditions were appropriate, without being subjected to gaps and  -16- periodic delays of the State-level system. A further advantage was the increased opportunity to refinance CBs. While this opportunity had not been precluded under the earlier, State-based projects, CBs only moved substan- tially into long-term agricultural credit from the mid-70s. Results of State-Based and ARDC Credit Proiects 3.03 Project completion reports and audits by the Operations Evaluation Department have been prepared for the 10 State-based credit projects that were financed by the Bank Group and for the ARDC I, II, and III projects, while the completion report for the ARDC IV project is being prepared. These reports indicate that the projects achieved physical targets, brought about increased agricultural production, and improved agricultural incomes. The ARDC series of projects coincided with rapid growth of agricultural lending and ARDC disbursements, with a trend towards diversified (non-irrigation) lending and an increasing focus on small farmers (through IRDP) and less developed areas. In addition to the Bank Group's support to the ARDC series of projects, bilateral donors and the European Community (EC) provided about US$500 million to ARDC since 1976. 3.04 All Project Performance Audit Reports conclude that the State and ARDC Credit Projects achieved their objectives. The Reports consistently question the adequacy of lending margins and financial viability of some of the smaller branches; the limited availabi'lity of groundwater, particularly in hard rock areas receives also major coverage in all Reports (pars 4.21). Increasing attention is paid to the worsening recovery situation; the ARDC II Report is the first to ask for a review of future Bank Group assistance in view of this issue. The ARDC I and II Reports also point to inadequate ARDC review and supervision procedures and shortages of competent staff. The Reports elaborate on the need for a balance between lending quality and volume, but acknowledge the importance of growth. 3.05 The Bank Group's rapidly growing support to agricultural credit allowed ARDC credit disbursements and agricultural credit to maintain its rapid growth. But insufficient account was taken of the credit system's inability to handle this growth; projects were appraised before the lessons learned from earlier ones were understood and evaluated. As a result, the fundamental factors affecting the credit system's performance received limited attention. Limited supervision was inadequate to provide the data and analysis necessary for investigating the system's problems. Con- ditionality to maintain lending quality and impLvve recovery--eligibility criteria--showed a gradually weakening between ARDC I and IV. Increasingly, criteria facilitated adjustments by branches to improve their eligibility; for example, State contributions to LDB capital were counted as recovery, recoveries were measured six months after the due date and "blocking" of accounts was allowed. The policy under ARDC III, of excluding branches with high overdues from NABARD financing was discontinued under ARDC IV. Although the ARDC projects increasingly emphasized rehabilitation of individual, poorly performing LDBs, execution and results were below expectations, in part, because fundamental variables (lending margins, quality control, train- ing) were inadequately addressed. Wnile successful in supplying substantial  -17- funding for minor irrigation and diversified investments, the projects were unable to stem mounting overdues, and witnessed a deterioration of several formerly strong LDBs and the increasing difficulties of the credit system to maintain adequate control over lending quality. IV. THE PROJECT A. Development Strategy 4.01 Institutional agricultural credit in India is playing an important role in sectoral development, most successfully in the northwest where production potential is high and supporting services adequately available. Despite rapidly expanding operations in other selected parts of India. the agricultural credit system has been less successful than in the northwest in matching growth with the selection of high-return investments. Agricultural potential in such areas is less promising, and risk is higher. The fundamen- tally different nature of agricultural banking in areas of higher risk, more poverty, lower financial liquidity, fewer services, and less potential has been inadequately reflected in the policy framework in which the agricultural credit system operates. Banking practices appropriate to an environment of better-off farmers involved in low-risk, high-return operations are being promoted in such different environments where the policy of low lending margins, high disbursement targets, and mass-lending operations for preselected beneficiaries have different consequences. 4.02 The various components of India-s multi-agency agricultural credit system have reacted differently to rapid lending growth induced by Govern- ment. The short-term cooperative system has been least effective in coping with demand for quslity lending; its relatively slow growth in recent years reflects disbursement limitations imposed because of inadequate recovery. More importantly, ongoing efforts to formulate and implement a concept of a viable and effective cooperative structure for short-term lending have not succeeded. Concerning term credit, the LDB structure has been most con- strained by its inability to provide services other than term-credit, and by its inflexibility in obtaining funds, and has been most exposed to the effects of risk in agriculture production. In some cases, LDBs are increas- ingly dependent upon ad-hoc State financial assistance. CBs have been most successful in dealing with the strain of high growth and a restrictive policy environment. Subsidized by their dominant non-agricultural operations, as well as by various concessions of RBI, and generally assisted by superior management and financial resources, they have expanded very rapidly, while particularly SBI has been introducing innovative lending strategies in selected areas. They have been more flexible in their approach to agricul- tural lending and more realistic in their assessment of and approach to risks. Hovever, having reduced the impact of risks, through the DICGC, they have become somewhat complacent about the level of incremental benefits being generated by new investments, particularly in mass-lending schemes.  -18- 4.03 GOI s commitment to agricultural development and poverty allevia- tion is likely to spur future growth of the agricultural credit system. The challenge is whether the system's growth can be combined with necessary improvements in lending quality, as reflected in selection of investments, institutional performance, and recovery. Approaches to improving a fundamen- tally intricate system require broad and firm but careful action, facilitat- ing reasonable adjustments over time. The thrust of a long-term strategy should be three pronged: in facilitating credit operations of all institu- tions, it should recognize essential prerequisites for healthy operations, taking into account regional differences in the agricultural environment; ir should aim at strengthening and rehabilitating the performance of credit institutions; and it should enhance monitoring, supervision, and quality control functions of the apex level. 4.04 To formulate a strategy, and develop a program to execute it, the proposed project would have GOI carry out an Agricultural Credit Review. The Credit Review would include five studies of all major components of the credit system (para 4.11) under the umbrella of a Senior Expert Group (SEG) consisting of senior Indian and foreign specialists. The studies would bring together existing knowledge of the credit system and add detailed analysis of key components. Draft Terms of Reference agreed during negotiations are detailed in Annex 4. Tne SEC would coordinate the studies, but its main task would be to formulate a specific, timebound program of recommendations to address the problems of the credit system, and ensure that these recommenda- tions are technically sound and take account of India-s socio-political environment. In parallel, the project would support and further expand ongoing efforts to strengthen credit delivery at the village level. A pilot scheme, commenced in October 1984 and involving all credit institutions in three districts (para 4.13) under the supervision of NABARD, would be expanded to 20 districts under the project (para 4.14). The experiences and conclusions of the pilot scheme would complement the Credit Review by provid- ing guidance on how credit operations and loan recovery can be improved at various branch levels. 4.05 The final results of the Credit Review would be available in 1988; thus they would not affect credit operations during the proposed project period. To improve lending quality and improve loan recovery, short-term measures would be required in the meantime, which, although not addressing fundamental problems, would assist in keeping ongoing credit delivery on track. Foremost, this short-term program would comprise of revised eligibility criteria (para 5.09), under which badly performing branches would have limited access to the refinancing facility of NABARD. This would be combined with efforts to improve recovery and the lending portfolio and financial position of weak branches (paras 5.10 - 5.12). The intention would be to restrict new lending by institutions with clearly unacceptable levels of recovery and to have NABARD focus its attention on ensuring that the financial status of these institutions remains within acceptable norms. Secondly, NABARD would streamline and improve its ongoing efforts to rehabilitate weak LDBs (para 5.07). The third component would be aimed at enhancing the lending quality of IRDP by strengthening its planning process and by improving NABARD-s control over implementation, as detailed in  -19- para 5.06. These measures, in conjunction with eligibility criteria, would concentrate IRDP lending on better-performing institutions. The fourth ccmponent would strengthen NABARD-s institutional performance, enhance the effectiveness of NABARD's monitoring and control functions (paras 5.26 - 5.27); and improve NABARD staff training (para 5.08). These measures would strengthen NABARD's capability to control and monitor credit delivery and recovery, and to improve lending quality. B. Proiect Genesis 4.06 In anticipation of the completion of ARDC IV (June 30, 1984). GOI submitted a request for a NABARD Credit Project to the Bank Group in April 1983. However, in view of the unsatisfactory performance of ARDC IV, the Bank Group was unable to appraise the proposed project at that time. In January 1984, the Bank Group embarked upon a critical review of the ARDC series of projects, which concluded that its involvement in agricultural credit was highly desirable if it would lead to improvement in the credit delivery and control system. 4.07 In June 1984, the Bank outlined the major problems of the agricul- tural credit sector and made suggestions for corrective measures to GOI, which gave its reaction in a detailed letter, on the basis of which the Bank Group decided that conditions for its support for the proposed project would be that: (i) GOI agree to a fundamental review of agricultural credit to be undertaken in parallel with Bank Group financial support for NABARD; (ii) where appropriate, necessary studies be undertaken by internationally recruited consultants and be monitored, guided, and reviewed by a panel of internationally recognized local and foreign experts; and (iii) any sub- sequent NABARD Loan/Credit would be linked to satisfactory progress being made on the formulation of a long-term strategy and program to improve India-s system of agricultural credit, based on the findings of the above-mentioned review, and to a start on implementation of the strategy. Based on the above understanding, appraisal followed in December 1984. C. Brief DescriDtion and Obiectives 4.08 The project's main objective would be to strengthen the agricul- tural credit system and improve the quality of credit delivery; specifically, it would initiate a series of institutional reforms in the agricultural credit system, in general, and in NABARD, in particular. In so doing, NABARD's role as apex institution in agricultural credit would be enhanced. Its second objective would be to increase agricultural production, rural income, and employment, while continuing emphasis on assistance to small farmers. GOI regards NABARD lending as a key program in its agricultural development plan and for national policies on poverty alleviation and employ- ment creation. 4.09 The project would, over a three-year period, help finance, in part, NABARD's medium- and long-term refinancing program to participating credit institutions (CBs, LDBs, RRBs and SCBs). The project would also include (i) execution of the Credit Review, (ii) implementation of the pilot scheme  -20- in 20 districts to strengthen credit delivery and improve recovery; and (iii) a component to support NABARD's training activities. 4.10 Contrary to practice under the ARDC series of projects, under the proposed project a portion of the loan for NABARD's refinance program (US$340 million) would be disbursed in tranches, the release of which would be trig- gered by introduction of revised eligibility criteria (para 5.09) and by the progress in execution of the Credit Review (para 4.11). NABARD would refinance minor irrigation, farm mechanization, plantation crops and hor- ticulture, animal husbandry, forestry, fisheries, storage and markets, serv- ices industries, and rural electrification. Bank Group finance would be channelled through NABARD, which would refinance loans made by participating banks in accordance with specific terms and conditions (Schedule A). Final beneficiaries would be individuals or groups of farmers, cooperatives, private and public sector companies and corporations, with more than 80% of NABARD's operations expected to refinance lending to individual or small groups of private sector borrowers. Lo_ns under IRDY, minor irrigation, and farm mechanization would predominate NABARD's refinancing program for FY86/87 to FY88/89. D. Detailed Features Agricultural Credit Review 4.11 The main objective of the Credit Review would be to define a program to strengthen tne agricultural credit system, in order to improve lending quality, effectiveness, and credit services to farmers, while enhanc- ing the effectiveness of supervision and control over agricultural credit operations. The Credit Review would be executed under the general direction of RBI. The complexity of the agricultural credit system in India, and the intricate nature and sensitivity of the issues to be addressed necessitate five separate studies. The studies would be executed by local and foreign consulting groups; GOI has indicated it intends to engage foreign leadership for Studies III and V. GOI would engage the SEG, which would consist of seven specialists to be individually recruited to coordinate and guide the five studies and to prepare an integrated final report. GOI has indicated it intends to recruit about three foreign specialists for the SEG. During negotiations assurances were obtained that GOI would engage consultants for the SEG and consultancy firms for the Credit Review in accordance with Bank guidelines, and would execute the Credit Review. Terms of Reference dis- cussed during negotiations, with details outlined in Annex 4, are summarized below: (i) Study I - Agricultural Credit in General would examine the role and operations of the entire agricultural credit system, and assess how within the framework of poverty alleviation and agricultural development policies that role should evolve and which factors have significant impact on its efficiency, effectiveness, and institu- tional integrity.  -21- (ii) Study II - The Role and Effectiveness of Lendiny Institutions would review a sample of LDBs, CBs, RRBs, SCBs, DCCBs and PACSs, and analyze their effectiveness in their present role, their financial structure, operational systems, and branch expansion policies, and recommend if and how the role, structure, and functions of these institutions should be adjusted. (iii) Study III - The Role and Functions of the Apex Level in Agricul- tural Credit would clarify and define the roles and functions of NABARD and other apex level institutions involved in rural credit and how such roles should be executed. It would formulate recom- mendations concerning NABARD-s capital structure, legal mandate, its relationship with RBI, the National Cooperative Development Corporation (NCDC), and lending institutions, and its role in improving lending quality and the institutional performance of banks. (iv) Study IV - Lending Costs and Margins would assess whether lending margins are appropriate in comparison to the costs of lending, taking into account staffing and other operational costs of properly functioning credit operations and the overdues situation. The study would attempt an assessment of the bad-debt situation and the amount of provisions necessary to remedy the situation and the impact on landing margins. The study rould make recommendations concerning appropriate lending margins for NABARD and the par- ticipating banks and indicate how such margins should be realized. (v) Study V - Organization and Management Study of NABARD would prepare a set of recommendations and a timetable to strengthen NABARD's operations. The study would assess how NABARD's institutional and functional structure should be adjusted to reflect the recommenda- tions of Studies I to IV, how its operational procedures should be strengthened, its staffing and training policies and practices be adjusted, its monitoring evaluation and inspection systems be improved, and its institutional development efforts be enhanced. 4.12 GOI has established a Secretariat to assist the Credit Review and SEG. Assurances were obtained that GOI would maintain this Secretariat with adequate facilities to assist SEG in carrying out the Credit Review. Members of the SEG would be selected by GOI in consultation with the Bank. SEG would prepare a final report of general conclusions and recommendations for a timebound program to rehabilitate the credit system. Implementation of the Credit Review would be directly linked to disbursement of the various tranches under the Loan, as detailed in Schedule B. Pilot Scheme for Strengthening Agricultural Credit Delivery System 4.13 The pilot scheme for strengthening agricultural credit delivery system and to improve recovery was initiated in March 1984 and commenced field work in October 1984, in one district each in the States of Gujarat (Sabarkantha), Madhya Pradesh (Hoshangabad) and Haharashtra (Aurangabad).  -22- Agricultural production in each district is mainly rainfed, although a sub- stantial irrigation command system does exist in Hoshangabad and there is significant irrigated production in Sabarkantha and Aurangabad. The main features of the pilot scheme are the initiation of systematic in-field con- tacts of bank-branch staff with actual and potential borrowers, brought about by provision of some additional staff, special pilot scheme training activities, and supporting resources. The pilot scheme covers all institu- tional agricultural credit operations *n each district. The scheme was designed to overcome the commonly observed problems of limited contact of farmers with bank-branch field staff; overly-ambitious and poorly organized responsibilities of such field staff; their inadequate guidance, supervision, and training; and the lack of focus on satisfying farmers- total credit needs. The objectives of the pilot project are to: (i) increase the volume and quality of agricultural credit operations; (ii) strengthen the agricultural credit delivery system at the branch level; and (iii) ensure a timely and sufficient supply of credit and its proper utilization and repayment, by providing more intense bank branch staff coverage and implementing appropriate flexibility in lending terms and conditions. The pilot scheme attempts to determine if these objectives can be achieved through disposition of seriously overdue accounts; greater flexibility in lending operations; limiting the workload of branch staff to reasonable levels; a systematic, closely supervised program of staff work; and providing regular, need-based training and linkages with lending policy-makers. 4.14 Although the pilot scheme in the first three districts was late in starting, it is recognized by GOI/NABARD and the Bank that key elements of an effective credit delivery system are now being addressed under the scheme. Preliminary results also indicate a positive impact on lending institutions- operations at the village level. Consequently, the proposed project would finance an expansion of the scheme into 17 new districts, i.e., a total of 20 districts in all. Districts would be selected so as to bring a wide range of districts under the pilot program covering as many states as possible. Priority would also be given to those districts that have serious recovery problems. NABARD would submit to the Bank, by June 1, 1986, a detailed proposal which would, inter alia, identify the districts selected, the number of banks and branches involved, and include a timetable for implementing the Pilot program. Tentative total costs (Annex 1) for the pilot scheme are estimated at about Rs 300 million (US$25 million). An amount of USS25.0 million of the proposed loan has been allocated to finance 100% of incremental staff, and operational costs and 95% of the costs of motor vehicles and motorcycles for the staff of participating banks. Most of the funds under this component would be passed on as grant by GOI to NABARD and to the participating banks.  -23- 4.15 NABARD would be responsible for providing guidance to participating banks on field training and for overall monitoring and evaluation of the pilot scheme. NABARD has a pilot scheme cell and has appointed a full-time District Project Officer for each district covered. NABARD would also estab- lish procedures to regularly review the feedback from the pilot scheme. In order to assess the results of the pilot program, assurances were obtained at negotiations that NABARD would, by October 1, 1986, prepare an initial evaluation of the pilot scheme, to be followed by a second review to be completed by October 1, 1988. NABARD-s Training Program 4.16 Training in agriculture credit and NABARD's role within this has been described in para 2.12. As under previous ARDC series of projects, the Bank will continue to support NABARD's programs for training of its own staff and those of participating LDBs, RRBs, CBs and DCCBs. As part of its ongoing efforts in this regard, NABARD has prepared a program for training activities over the next three years, see Annex 1, Table 6. Up to US$6.5 million of the proposed Bank loan will be available to reimburse NABARD (on grant basis) for eligible expenses incurred for training activities during the project period. These funds would be utilized for payment of participantsC per diem and costs, lecturers fees, teaching material, and equipment as well as other administrative costs of running NABARD's training facilities. NABARD's Refinancing Program 4.17 NABARD's tentative refinancing program for the three years 1986187 to 1988/89 and resources to finance this program are indicated in Annex 3, Table 4 and discussed in sections E and F of this chapter. The following paragraphs describe and discuss some of the issues related to the principal activities and subsectors supported by NABARD-s refinancing. NABARD would also submit annual plans detailing the sectoral allocation of its funds to the Bank Group prior to the start of each fiscal year. Minor Irrigation 4.18 There is further scope for carefully planned investment in minor irrigation in India. On an all-India basis, the current estimates are: Ultimate Presently Investment Potential Irrigated Potential (M ha) (M ha) (M ha) Groundwater 40 28 12 Surfacewater 15 9 6 Total 55 37 18  -24- Further investment particularly in less developed States in eastern India appears desirable. Sixth Plan investments (through 1984) have resulted in an annual increase of about 1.4 million ha per year, totalling some 7 million ha of both groundwater and surface water. The Working Group on Minor Irrigation for the Seventh Plan has called for an increase in area of 9.5 million ha, 8.2 million ha from groundwater and 1.3 million from surface water, or more than half of the remaining 18 million ha. If these targets are achieved by the end of the Seventh Plan, 85% of the ultimate minor irrigation potential would be reached, and 90% of the groundwater potential. As detailed in para 4.20, many areas would reach their potential much earlier, and safeguards need be introduced to limit excessive exploitation. It is expected that during the Seventh Plan investment in groundwater would slow down and more attention would be paid to improvements in the efficiency of water delivery and rehabilitation of poorly functioning systems. 4.19 Groundwater development would continue to be an important component of lending in minor irrigation (85%). Investments by farmers would include dugwells, dug-cum-borewells, shallow tubewells, filter points, deepening of existing dugwells and adding bores to such wells, provision and replacement of pumpsets, standby pumpsets, replacing diesel pumpsets by electric sets, providing electric connections, piping irrigation water from the water source to the field, provision of pump houses, provision of storage tanks, and at a minor level, the provision of sprinkler and drip irrigation equipment. Costs vary widely according to local conditions. Surface water schemes would constitute about 15% of lending in the minor irrigation subsector. Schemes would range from provision of a small pumpset to an individual farmer to lift water from a river, canal, tank or lake to large public schemes involving river intakes, pumping stations, storage tanks and a piped or lined channel irrigation delivery system commanding several hundred hectares. 4.20 Performance of electric pumpsets has improved in recent years, but still leaves much to be desired. Delays of two months to two years in obtaining a power connection after a farmer invested in a well reflect a low level of efficiency. A number of State Electricity Boards claim to supply a connection within three to six months but in practice, procedures in sanc- tioning connections are slow, although banks obtain electricity board confir- mation that a connection would be made before a loan is sanctioned. 4.21 By the end of the Seventh Plan the most serious constraint to minor irrigation development would be the danger of over-exploitation of groundwater. Almost all of India