Report No. 30740-IN India Scaling-up Access to Finance for India's Rural Poor December 2004 Finance and Private Sector Development Unit South Asia Region Document of the World Bank ACKNOWLEDGEMENTS This report was prepared by a WorldBank team led by PriyaBasu, under the overall guidance of Joseph D e l M a r Pernia, Manlou Uy and Michael Carter. T h e peer reviewers were Anjali Icumar, Asli Demirguc-Kunt and Abayomi Alawode. Major contributors to the report were Niraj Verma, Ulrich Hess and Leora Idapper of the World Bank, and Pradeep Srivastava and Rajesh Shukla of the National Council of Applied Economic Research (NCAER), India. T h e report draws on a Rural Finance Access Survey (RFAS) for India, designed and undertaken jointly by the World Bank and NCAER, India. I t also draws on the following background papers: (i) "Microfinance: Analytical Issues for India" by Jonathan Morduch and Stuart Rutherford; (ii) "Scalmg up Microfinance in India" by Priya Basu and Pradeep Srivastava; (iii) "Microfmance in India: Banyan Tree and Bonsai" by Vijay Mahajan and Bharti Ramola Gupta; (iv) "Review of Rural Finance Institutions in India" by Ramesh Deshpande and Niraj Verma; (v) "The Use of N e w Products, Processes and Technology in the Delivery of Rural Finance in India" by Ulrich Hess and Leora Idapper; (vi) "Informal Finance in Rural India" by Jock Anderson and Niraj Verma; and (vii) Agricultural Commodity Markets in India: Policy Issues for growth by Susan Thomas. Connie Bernard, D i n a Umali-Deininger andJames Hanson provided helpful comments. T h e report incorporates suggestions from representatives of the Government of India, Reserve Bank of India, National Bank for Agriculture and Rural Development, Small Industries Development Bank of India and commercial banks, as well as microfinance practitioners and other key stakeholders, who attended a workshop in N e w D e l h i in February, 2004 at w h c h an earlier draft of the report was presented and discussed. Administrative support was provided by Anita D'Souza, Heather Fernandes and Maria Marjorie Espiritu. TABLE OF CONTENTS EXECUTIVE SUMMARY .......................................................................................................................................... I I.INTRODUCTION ................................................................................................................................................. 1 INDIA'SLANDSCAPE RURALFINANCE ........................................................................................................................ 3 What are thefinancial needs of the ruralpoor? 3 Ruralfinance service providers 4 11. ACCESS T O RURAL FINANCE ININDIA THE EVIDENCE .................................................................. 7 SUPPLY SIDE INDICATORS OF ACCESS FINANCE TO .................................................................................................... 7 ACCESS RURALFINANCE:EVIDENCE TO FROMTHE DEMAND SIDE .......................................................................... 11 Access to savingddeposit accounts 12 Payments Services -Limited use, high cash economy 14 Access to Credit 14 Access to Insurance 16 TheImportance of Informal Finance 16 111.WHAT CONSTRAINS ACCESS TO FINANCE FOR INDIA'S RURAL POOR? ..................................... 18 WHY BANKS ARERELUCTANT TOLEND RURALCLIENTS.................................................................................. TO 18 Uncertainty and default risk 18 Lack of credit information. 19 The tyranny of collateral 19 Transactions costs 20 Weaklegalframework and enforcement issues. 20 Governmentpolicy 20 WHYDoSMALL, RURALBORROWERS RURALBANKS FIND UNATTRACTIVE? ........................................................ 25 Absence offlexible products and services 25 Transactions costs 25 Collateral 26 IV. RECENT EFFORTS ININDIATO IMPROVERUR4LACCESS TO FINANCE: THE ROLE OF FORMAL-INFORMAL LINKAGESAND NEW PRODUCTS ....................................................................... 27 SHG-BANKLINKAGE APPROACH: LINKINGCOMMERCIALBANKS GRASSROOTS TO BORROWERS .......................... 27 How Effective Has SHG Bank Linkage Been in Targeting the Poor? 28 Impact on vulnerability ofpoor households 31 Some Lessonsfrom SHG Bank Linkage 32 Key concerns: Limited outreach and scale of lending and issues infinancial sustainability 33 MICROFINANCE INSTITUTIONS (MFIs)..................................................................................................................... 34 What has constrained the outreach and scale of MFIs? 35 THE"SERVICE PROVIDER" MODEL MICROFINANCE .............................................. 36 THEKISANCREDITCARD(KCC)............................................................................................................................. OF PILOTEDBY PRIVATE BANKS 37 RECENTINNOVATIONSINMICRO- AND WEATHER INSURANCE ................................................................................ 38 COMPOSITEFINANCIALSERVICE PROVIDERS........................................................................................................... 39 PORTFOLIO SECURITIZATION .................................................................................................................................... 40 PRICEINSURANCEAND RISKMANAGEMENT PRODUCTSFORFARMERS ................................................................... 40 THESTRATEGICUSEOF INFORMATIONAND COMMUNICATIONTECHNOLOGY (ICT) ................................................ 41 V. MEETINGTHE CHALLENGE OF SCALING-UP ACCESS TO FINANCE FOR INDIA'S RURAL POOR THE POLICY AGENDA ............................................................................................................................ 43 MAKING FORMALFINANCIALSECTORBETTER BANKING THE AT THERURALPOOR.............................................. 43 Low-cost ways of reaching the ruralpoor through theformal sector: Role of B a n k and Government 43 Improving the incentive regime, andpromoting competition: Role of Governmentpolicy 44 SCALING-UP MICROFINANCE ................................................................................................................................... 48 REFERENCES ......................................................................................................................................................... 52 ANNEX 1:STRUCTURE OF INDIA'S FINANCIAL SYSTEM ......................................................................... 55 ANNEX 2: WORLD BANK/NCAER RURALFINANCEACCESS SURVEY (2003) ....................................... 56 ANNEX 3: SUMMARY REVIEWOF THE PERFORMANCEOF RETAILRURAL, FINANCE INSTITUTIONS (RFIS) ......................................................................................................................................... 58 ANNEX 4: APEX RURALFINANCE INSTITUTIONS .................................................................................... 64 ANNEX 5: SUMMARY OF RECOMMENDATIONS OF RECENT GOVERNMENT ................................. 68 COMMITTEES ON FORMAL RFIREFORMS ................................................................................................... 68 ANNEX 6A REGULATION OF RURALBANKING SELECT INTERNATIONAL REFERENCES - ......71 ANNEX 6B: SUPERVISIONOF RURALBANKING SELECT INTERNATIONAL REFERENCES - .....72 ANNEX 7: RURAL BANKING CRISESAND POLICY REFORM RECENT INTERNATIONAL ........................................................................................................................................................ - EXPERIENCES 73 ANNEX 8: ENABLINGFRAMEWORKFORMFIS.DEMANDS FROMTHE SECTOR ............................ 74 EXECUTIVESUMMARY I Background Since the ear4 nationalplans, successivegovernments in independent India have emphasixedthe link between improving access tojnance and reducingpover& a stance that has had inflence global4.l T h e need to improve financial access for India's poor, the overwhelming majority of whom are concentrated in rural areas, 2 motivated the nationalization of commercial banks in the late 1960s, and an aggressive drive through the 1970s and 1980s to expand rural banking, coupled with policies mandating banks to provide subsidized credit to rural households. The 1990s saw the partial deregulation of interest rates, a gradual reduction in Govemment's stake in commercial banks, and increased competition in the banking sector. But political compulsions have dictated the persistence of government intervention in 'priority' sectors such as rural finance. Government still dominates the rural finance institutions, and duected credit and subsidies, albeit lower than before, continue to distort risk-retum signals in m r a l finance. Access to jnance for the mralpoor has improved somewhat over the past decades, with the public sector commercial banks being the dominant players in the fomal mral jnance market. Govemment's policies of branch expansion of public sector commercial banks in rural areas, particularly in the 1970s and 1980s, certainly helped banking outreach. Today, India has over 32,000 rural branches of commercial banks3 and Regional Rural Banks (RRBs), some 14,000 cooperative bank branches, 98,000 Primary Agricultural Credit Societies (PACS), not to speak of the thousands of mutual fund sellers, several Non-Bank Finance Companies (NBFCs) and a large post office network with 154,000 outlets that are required to focus on deposit m o b h a t i o n and money transfers. Not surprisingly, India compares favorably with other developing countries in terms of the average population served per bank branch, and the average geographxal area served per branch. But the vast major& ofIndia's ruralpoor still do not have access tofomaljnance. T h e World Bank-National Council of Applied Economic Research (NCAER) Rural Finance Access Survey (RFAS), 2003, prepared as a background to t h s report, indicates that rural banks serve primarily the needs of richer rural borrowers: some 66 per cent of large farmers have a deposit account; 44 per cent have access to credit. Meanwhile, the rural poor face severe difficulties in accessing savings and credit from the formal sector. Some 87 per cent of the poorest households surveyed (marginal farmers) do not have access to credit, and 71 per cent do not Note: Marginal farming households=landholding4acres; Commercial households= with or w/o land but with source. Access to formal credit i s income from non-fann sources exceeding half o f total household income; Others=mixed households with land and non-farm commercial incomes but the particularly a problem for meeting latter being less than half o f their total household income. Source: RFAS. 2003 'The focus on povertyand finance was articulatedmost famously in the 1954ReserveBank of India (RBI) report on the All-India RuralCredit Survey of 1951-52(RBI, 1954). Of the estimated 260 million Indians (or 26% of the population)who live in poverty, some 193 million (or 74%) live in ruralareas. Another 180d o n ruralpeopleare "near poor". A majorityof these householdsare marginal or small farmers, and the pooresthouseholds are landless. The total number of bank branches (rural and urban) is 66,500. The bulk of these branches are of public sector commercialbanks. i unforeseen expenditures and poor access to formal finance has resulted in a heavy reliance among poorer rural households on informal finance, mostly moneylenders and shopkeepers. Around 44 per cent of the households surveyed by RFAS, 2003 report having borrowed informally at least once in the preceding 12 months; the interest charged on informal loans averages 48 per cent per annum. Access to other financial services, such as insurance, also remains h t e d for the ruralpoor, although many would k e to have access to insurance - over 82 per cent of households surveyed in WAS, 2003, did not have any insurance, and practically none of the poorest households surveyed had insurance. Recent years have witnessed the growth of new microfinance approaches designed to serve poorer households, and often involving partnerships between Government, Non-Govemmental Organizations (NGOs) and banks. But these approaches have not yet been able to make a dent. Informal sector lenders thus retain a strongpresence in rural India. I1 What Explainsthe Inadequate Access to Finance for India's RuralPoor? T h e lack of access to adequate finance on reasonable terms for India's rural poor may be attributed to a combination of factors affecting both banks and their clients. Why banks don't want to serve the ruralpoor: Serving the rural poor is a high-risk, hgh-cost proposition for banks: First)is the problem of unce~ainty- about the repayment capacity of poor rural borrowers, and their irregular/volat.de income streams and expenditure patterns - which, in the absence $credit information, drive up default risk. These problems are exacerbated by the borrower's lack of collateral, and/or difficulties incontract design and enforcement. Second, the transactions costs of rural lending in India are high, mainly due to m a l l loan sizes, high frequency of transactions, large geographical spread and heterogeneity of borrowers, and widespread illiteracy. For private sector banks lack of a rural branch network i s an additional problem. Third,the Govemmentkpolicieshave made thngs worse from the banks' perspective, creating a "jkancial climate" that i s not conducive to lending in general, and rural banking inparticular: HighJiscal d$cits and statutory preemptions imposed on banks crowd out credit to the private sector; another persisting distortion that has failed to generate the desired results i s the `priority sector' lending target; Persisting interest rate restrictions- "floors" on short-term deposit rates and lending rate, "caps"on small loans impose an "implicit tax" on banks; Government's domination of/interference in rural banks, particularly the RRBs and cooperative banks, further distorts bankers' incentives; inefficiencies arising from weak governance, poor management and weak regulatory standards and supervision, have meant that many of these banks are in deep financial distress, and no longer able to perform their task of financial intermediation; Bankers' risk aversion to lending i s exacerbated by a pervasive culture of suspicion towards bankers, whose lending decisions are often subject to stringent scrutiny by parliament, the Central Bureau of Investigation, etc. Why smaU, rural borrowers find rural banks unattractive: First, rural banks do not provide flexible products and services to meet the income and expenditure patterns of small rural borrowers. Second, the transactions costs of d e a h g with formal banks are hgh. Procedures for opening an account or seeking a loan are cumbersome and costly (with high rejection rates) and, as indicated by the RFAS-2003, clients often have to pay hefty bribes (ranging from 10-20per cent of the loan amount) .. 11 to access loans, so that the ultimate cost to borrowers is very high (despite interest "caps"). I t takes, on average, 33 weeks for a loan to be approved by a commercial bank.4 Third, banks demand collateral, which poor rural borrowers lack. L a n d remains by far the most predominant form of collateral. But, this collateral i s seldom executed, so it i s just another cost, with little benefit in practice. I11 N e w Approaches and Products to Improve RuralAccess to Finance inIndia Inrecent years, severalnewapproaches andproducts have evolved to provide finance to India's rural poor. SHG-Dank linkzipprogram Most notableamong recent approaches to improve access tojnancefor the ruralpoor is the `SeghelpGroups (SHGs) Bank Lznkage' model, the growth of which -jo m just 500 SHGs linked to banks in the ear& 1990s, to over 700,000 2003 - has been tmb remarkable. Data from the RFAS-2003 shows that SHG Bank ltnkage appears to have targeted poorer segments of the rural population in an effective manner, reducing the vulnerabllity of clients. But outreach, volume of lendmg and average loan size remain h t e d . With an outreach of about 12 d o n women and their households ltnked to banks through group savings accounts and an estimated 2-4 d o n w o m e n having outstandmg credt from banks, SHG Bank Linkage has a long way to go before it can really make a dent. K e y challenges facing the initiative are: (1) Inadequate attention to group quality could jeopardize longer term credibhty and sustainabllity. (2) Capacity constraints and the cost of group formation5. (3) State-owned banks have been lendmg to SHGs at interest rates of between 9 per cent to 12 per cent per annum., even though recent studies indicate that the all inclusive costs are, in fact, m u c h higher, and could range anywhere from 15 per centb to 28 per cent per year. Unless banks charge interest rates to recover costs, the model's financial viabdity and longer term sustainability may be jeopardrzed. Microfinance Institiitions I n recent years, other institutional structures for mimjnance have emerged, notabb, independent, specialized Mimjnance Institutions WFIs).T h e Small Industries Development Bank of India (SIDBI) has been the largest lender to the emerging MFIs, though Friends of Women's World Bankmg India (FWWB) as well as the National Women's Fund(Rashtriya Mahila Kosh) have also played an important role. T h e outreach of Indian MFIs, however, i s modest in comparison to SHG Bank Linkage, and also in comparison to MFIs elsewhere in the world. InMarch 2003, the Indian MFIs sector as a whole had a total outreach of less than one d o n borrowers. T h e h t e d outreach and scale of Indian M F I s , reflects, at least in part, the absence of an enabling policy, legal and regulatory framework which impact the abdity of MFIs to mobdue member deposits, equity and raise debt from external sources. MFIs are also constrained by the lack of adequate capacity and skills in financial control and management, MIS, new product design, etc. Further, since most M F I s in India lend to SHGs, this means that MFIs in India are constrained by many of the same factors that have held back the High transactions costs of dealingwith formal banks translate into a low frequency of transactions. According to the RFAS-2003,more than 60% of households with bank accounts report accessingtheir accounts at a frequency of less than once a month (62.3%of households with accounts in banks access their accounts less than once a month, while the same percentageamongst households with accountsin RRBs is 70.8). The estimated cost of creating and sustaining new and high quality SHGs is controversial, with National Bank of AgricultureandRuralDevelopment (NABARD) claimingthat it is as low as Rs 1,000 (US022) per group and NGOs saying it takes as much Rs 12,000 (US$267). The Ministryof RuralDevelopment has established a norm of Rs 10,000 (USF222) per group,which experts claim is realistic. Whichis what privatebanks like ICICIBank charge when they lendto SHGs. ... 111 outreach and scale of SHG Bank Linkage. In particular, capacity, time and cost issues related to group formation have posed constraints. Pmtiierships Between Private Bmtks, 2l.licrofinanciersand Senice Providers Encouraged ear4 results, the new private sector banks, most notab4 ICICI Bank, but also UTI Bank, HDFC Bank and others, are active4 seeking exposure in the microjnance sector. These banks have been lendmg to MFIs. While their current exposure to microfinanceis too small to make a difference to their overall portfolio, some of these newer banks are pursuing innovative approaches to microfinance - as a potential business and not merely as a social or priority sector lendmg obligation. K e y innovations include a pilot scheme by ICICI Bank that uses NGOs or M F I s , traders, or local brokers (who are close to the farmer by the nature of their business), as intermediaries/ `service providers' for origmating, managmg and collecting loans to groups of small and marginal farmers. Banks are also experimenting with an approach now termed the "Integrated Agricultural Service Provider" (IASP) approach, where the bank identifies an IASP that has a good relationshlp with farmers and provides genuine and timely information through extension services, and enters into a tripartite agreement with the IASP and the output buyer. This reduces transactions costs and the risk exposure of all parties and, therefore, presents a potentially low-cost way of serving the rural poor engaged in marginal or smaller farming. TIxeXsan Credit Card Another recent approach toproviding agrikultural credit, including to smallfarmers, is the Kzsan Credit Card (KCC), launched in 1998-99. By March 31, 2003, some 31.6 d o n KCCs had been issued by commercial banks, R R B s and cooperative banks. Though these are not credit cards, the I4acres; Commercial households= with or wio land but with per year. income k o m non-farm sources exceeding half o f total household income; Others=mixed households with land and non-farm commercial incomes but the ,lCc:L"J:c to.1.111!117~s./&po.ritar1ntll:tf latter being less than halfo f their total household income. Source : WAS-2003 Over one-half of rural households (59 per cent) do not have an account with a formal financial institution. (Table 7). What's more, among those with no accounts, 97.5 per cent of households report never having applied. Almost two-hds of households without an account do not perceive the need for deposit account services, whde another fifth report not being aware that they could open an account. T h e biggest reason for not having an account appears to b e a perceived "lack of need". 26 Higher incidenceof debt can be interpretedin differentways; it could s i g n i f y greater access and ability to borrow but it could also denote greater distress leadingto higher demand for debt. 27 Of the estimated 260 million Indians (or 26% of the population)who live in poverty,some 193 million (or 74Yo) live in ruralareas. Another 180million ruralpeopleare "near poor". A majorityof these householdsare marginalor smallfarmers, andthe pooresthouseholdsare landless. 28 The RFAS, 2003 covers 6000 ruralhouseholds.The last such survey, the AIDIS, was undertakenby NSSOin 1991. 12 Marginal Small Large Commercial Others Total With Account 29.6 55.3 66.05 41.96 60.88 41.16 Without Account 70.4 44.7 33.95 58.04 39.12 58.84 households; with or w/o land but with income from non-farm sources exceeding half of total household income; Otherszmixed households with land and non-farm commercial incomes but the latter being less than half of their total householdincome. Source: RFA4S-2003 While around 41 per cent of rural households have an account with a formal financial institution, wide differences exist between the access of large and small farmers, and between those who had other sources of income, and `pure' farmers. By far the most deprived segment in terms of access to an account are the "marginal" farming households (withless than 2 hectares of land), followed by the "commercial" category (households that rely on non-farm sources for more that one-half of their income). Over 70 per cent of marginal farmers have no bank account. (Table 7). T h e bulk of rural accounts are in commercial banks, which Figure 10: Distribution of Accounts have one and half times the numbers of rural accounts compared to the dedicated RRBs. (Figure 10). T h e cooperative sector only has a small fraction of account Others 1% 1 holders (12 per cent). T h e post office system, despite a very wide network of branches and close proximity to Caoperative7 i3., 4% rural clients (refer below to the discussion on transaction costs at the end of Section 119, surprisingly accounted for an even smaller percentage of account holders (2 per cent), perhaps indicating the latent potential for tapping Bank the large and widespread postal network for provision of 51% deposit serviceszg. While multiple accounts with formal financial institutions are relatively rare, the survey indicates that rural households do not expect to have all their financial service needs m e t by just one provider, or by just Source : WAS 2003 - one type of provider, and that they are used to multiple informal providers. Households' preference for commercial banks (over other types of formal financial institutions) does not appear to be related to physical distance (the mean distance to R R B s and cooperatives i s lower than to commercial banks). In general, frequency of visits to formal financial institutions i s low, with the main reason for infrequent visits being the high costs related to travel time/transport. T h e most important reason for having an account i s for safekeeping of monetary assets: over 72 per cent of with an account report safekeeping as the primary use of the account. 29The low share of the post office system in WAS-2003, could however, be attributed partly to a lack of association amongst respondents of post offices as `fmancialservice providers', whichis the terminology that the survey used. 13 1 ' ~ ! ~ ' / / r e i,.SUY&W - Lipd/rd LIJ.L)> h&h ~ i c b i l ~ ' ht ~ ~ ~ ~ t / ~ i t ~ j ' Even those households who do report having bank accounts appear to have little use for payments and checktng services. Moreover, there i s h t e d transactions demand as most transactions are cash- based. Some 48 per cent of households in our survey report receiving their incomes in the form of cash and 93 per cent report cash at home as the main mode of financing usual household expenses. And even in the case of payment services such as remittances, cash dominates: 82 per cent of households report cash as the main mode of remittance, as opposed to 15 per cent who report checks/drafts and 1.7 per cent reportingpostal money orders as modes of remittance. It follows that a substantial segment (over one-half) of rural households may have no use for the type of depository or payment services offered by formal financial institutions, which may b e explained by (1) their low levels of financial assets and flows, (2) an environment characterized by the preponderance of cash transactions. Households with accounts tend to use them in a h t e d way, accessing them with low frequency and with little use of checks (the primary use of deposit accounts i s for safekeeping, with little role for payment services). Households without accounts are presumably too poor to have any excess savings to safeguard. There i s some indication of an increase in `indebtedness' among rural households over the past decade - however this can also mean increased access to credit, and as such m a y be a step forward. According to the AIDIS of 1991, just 16 per cent of rural households had a formal loan outstandmg. Based on the World Bank WAS-2003, the correspondmg number was 21 per cent. Some 79 per cent of rural households do not have access to a formal loan, and once again, access i s particularly a problem for the marginal farmers and commercial households. (Table 8). Furthermore, some 97 per cent of households without a formal loan report not having applied for a loan in the past 3 years; the main reason reported for not applying i s a "lack of need", and the second most important reason i s "complicated procedures--an indication, perhaps, of the fact that based on the experience of others, they know they don't stand a chance of receiving a loan. Table 8: Access to credit from financialinstitutions, by household category (YO) Marginal Small Large Commercial Others Total With formalloan outstanding 12.97 30.79 44.36 16.78 29.47 21.01 Without formal loan 87.03 69.21 55.64 83.22 70.53 78.99 Access to formal credit i s particularly a problem for meeting unforeseen expendtures - family sources and the local moneylender dominate. Over 90 per cent of households report financing unusual expenses from cash at home, whde the second most important source of financing such expenses was reportedly informal loans from family, friends, or moneylenders. Newer sources, such as the recently introduced IGsan Credit Cards (ICC), are s t d l statistically insignificant. Use of bank credit for investments depends on the type of investment -bank loans are most frequently used for land acquisition/improvement and perhaps machinery and vehcles, whde the purchase of livestock ( w h c h cannot b e collateralized easily) i s typically financed from 14 informal sources. About 4-6 per cent of loans from banks and R R B s respectively are used for purposes of working capital.30 Among households who have a formal loan outstandmg, commercial banks are the main source, at interest rates of 12.5 per cent p.a. (Figure 11). There i s h t e d evidence of c r e l t rationing for formal Figure 11:Credit outstanding by source - - c r e l t on the basis of amounts applied for, approved and Olhrrs received, with commercial banks having the best record (loan amount received as a percentage of amount applied i s 92 per cent), followed by R R B s (88 per cent) and schams government schemes (87 per cent). However, on the l% basis of applications rejected, rationing appears to be Bank hgh. 51% Longer processing times for loans, together with bribes, could result in higher effective costs to borrowers and consequent c r e l t rationing. I t takes, on average, 33 weeks for a loan to be approved by a commercial bank. Source : WAS-2003 On average, some 27 per cent (and 48 per cent inUP) of sample households who borrowed from an RRl3report having to pay a bribe to get the loan (and thts figure is 48 per cent in the case of UP), a little under 27 per cent of households who borrowed from a commercial banks paid a bribe, and 10 per cent of households who borrowed from a credit cooperative paid a bribe. T h e bribe amounts vary from between 10 per cent of the loan amount (inthe case of banks) to 20 per cent (inthe case of cooperatives). In all cases the performance of UP is much than the average, indicating deeper problems in the integrity and internal controls of formal financial institutions in UP. (Table 9). Source :WAS-2003 The majority of loans extended by commercial banks, R R B s and cooperatives are collateralued, with 89 per cent of households who borrowed from RRBs, and 87 per cent of households who borrowed from commercial banks, reporting that they had to provide collateral. Land remains by far the most predominant form of collateral. T h e value of collateral required as a proportion of the loan, however, remains relatively low in the case of banks and RRBs, at under 10 per cent, but high in the case of government schemes. (Table 10). Bank RRB coops Schemes Others Collateral required (Yo of loans) 87.0 89.3 72.9 58.3 83.1 Value of collateral as O/o of loan 9.1 9.5 11.0 26.8 24.3 Source: RFAS-2003 30 Loans for w o r h g capital are relatively fewer since the sample consists primarily of cultivating households with only a small fraction of small and microenterprises. Even within the latter, the majority are engaged in trading and other service activities. 15 Cre&t i s mostly used as stated; however, &version i s more likely in the case of short term (working capital) loans. It may follow that the same households who report the lack of need for depository/payment services, may also comprise a substantial proportion of households without repayment capacity for any kind of cre&t from banks or other formal financial institutions. These households may comprehensively lie outside the ambit of the formal financial sector. Access to insurance remains h t e d over 82 per cent of households surveyed report not having any insurance, 3 per Table 11:DistributionofInsurance. bv%ofhouseholds cent report not knowing what insurance Life Insurance 13.21 n2.86 is, and 15 per cent report having 0.371 5.48 insurance. L i f e insurance i s by far the HealthInsurance most predominant insurance product Fire Insurance 0.031 0.26 available and the demand for h s product Theft Insurance 1.01 0.74 appears to far exceed its supply: while 13 GopInsUrance 0.2 18.98 per cent of households have access to life OtherInsUrance 0.13 1.68 insurance, some 73 per cent report that thj wouldbe their preferred insurance product. Demand for crop insurance i s also relatively high:while only 0.2 per cent of households report having crop insurance, some 18 per cent express a preference for this product over other types of insurance. (Table 11). Around 44 per cent of the households surveyed report having borrowed informally at least once in the 12 months; the interest charged on informal loans averages 48 per cent per annum. Not surprisingly, informal borrowing is very important for the poorest (margmal and commercial categories), who are the most deprived of formal finance, but it i s also important for small farmers. (Table 12). Table 12: Incidenceandsize ofi n f dbomraingbyhouseholdcategory (ohofhouseholdsunless othenvises t a d I I I Source: WAS-2003 T h e main source of non-formal borrowing i s Table33: Sourceoflastnon-formalloan, moneylenders (some 56 per cent of households b y % o f h o ~ l & who report having borrowed informally in the S H G W 289 past 12 months used moneylenders); hdlord 7.76 microfinance (through SHGs or NGOs) plays a Mane$gder 55.87 modest role at present. (Table 13). T h e strong Friends/R$atives 31.36 growth of microfinance in AP in recent years is ctkrs 212 16 households inAP are SHG members compared to 8 per cent for UP; in terms of credit from sources other than the formal sector institutions, 11.5 per cent of AI?households borrowed from SHGs over the last year compared to 3 per cent for UP. T h e largest uses of informal loans are for Table14Usesoflasti n f dloan, byO/o ofhouseholds meeting "famdy emergencies" and "social expenditures" arising from events such as births, marriages, deaths. Some 13 per cent of borrowers report using the informal loan for investment-related purposes. (Table 14). Putheinputs 3.92 E u r d y T " q 28.76 Flexible repayment of informal loans, their shorter duration (as compared to formal loans), and the frequency with which they can b e I 11.61) accessed, appear to be significant features that make such loans more attractive to the poor. Sources: WAS-2003 Some perceived advantages of informal loans over formal loans are as follows: (1) contractual tlexibdity, flexiblltty in repayment terms/schedules; (2) lower discrepancies between loan amounts sought and received; (3) less reliance on collateral: only 16.5 per cent of households report providingcollateral against the loan. But among those who report providing collateral for informal loans, the overwhelming collateral cited i s "self labor" - evidence of interlinked credit contracts spanning credit and labor. Almost all cases reporting self labor as collateral are landless/marginal farmers, who are too poor to offer any assets as collateral, and hence most vulnerable to harsh contracts linkingtheir labor to loans. 17 111.WHAT CONSTRAINSACCESSTO FINANCE FORINDIA'S RURALPOOR? A combination of factors - affecting both banks and their clients - contribute to reducing the supply of finance for the poor, drivingup costs and hampering access. \Yhy BanksAre Reluctant To Lend to RuralClients Fromthe perspective of the ruralbanks, serving the ruralpoor is a high-risk,hgh-cost proposition. Lending to some segments, especially to the very poor, i s surrounded by uncertainty about repayment. T h e rural poor tend to have irregular/volatde income streams and expenditure patterns (Box l), they also tend to be highly exposed to systemic risks such as crop failures or a fall in and commodity prices, and therefore, may face real difficulties servicing loans. So banks have legitimate concerns while d e a h g with the rural poor, and tend to perceive such loans as risky. Box 1:Income and Expenditure Patterns of India's Rural Poor Rural households have highly irregular and volatile income streams. Irregular wage labor and the sale o f agricultural products are the two main sources o f income for rural households. The poorest rural households (landless and marginal farmers) are particularly dependent on irregular wage employment: About two-thirds o f marginal farming households surveyed rely on wage labor as their primary source o f income (incontrast, only 9% o f large farming households rely on wage labor for primary income and none these households rely on irregular wage labor.). while 29% r e h on s e h e farm Droduce. Rural households also have irregular expenditure patterns. Over one-half o f the households surveyed report the bulk o f their expenditures as being either daily or irregular. The typical expenditure profile o f rural households surveyed i s that o f small, daily or irregular expenses, incurred all through the month. Furthermore, some 99% o f households report having incurred at least one unusual expenditure over the past six months, with the most frequent reasons for the latest unusual expenditure reported as medical or social purposes (related to births, marriaees. etc.). Source.RFAS-2003 18 O n e of the main sources of risk faced by rural financiers in their dealings with prospective rural clients i s a fall in commod~typrices below production cost levels, and t h i s risk becomes all the more significant in the absence of efficient price discovery and credible national commodity prices. Market-based tools to insure against commodity price volaultty (e.g. futures and options) already exist and are widely used in high-income countries, but they are not prevalent in India. Why? First, the minimumsize of contracts traded onorganized exchanges far exceeds the annual value of production of individual small- and medium-sized producers. Second, small producers, as well as many market intermediaries, lack knowledge of such market-based price insurance instruments and an understanding of how to use them. Third, the sellers of such instruments, generally international trading firms, are often unwihg to engage with a n e w and u n f a d a r customer base of small-scale producers, characterized by high transaction costs, credit issues, and performance risk. Problems caused by uncertainty are exacerbated by the lack of reliable information on the past credit history of borrowers. There are a number of sources of credit information in India, but none of these focuses on small, rural borrowers. Credit information on such borrowers i s difficult to obtain because the majority of the rural poor rely on moneylenders and other informal lenders, and it i s not in the interest of such lenders to pass on a borrower's good credit repayment record to other providers of finance. The unavailabhty of credit information has reduced the volume of lending in IndIa; because performance risk measures are unavailable, the current risk management practice of banks i s to control loan amounts. Better credit information could dxectly increase the amount of financing available to rural borrowers. O n e way in which a financier can reduce the risk of losing his money to uncertainty i s by requiring collateral-valuable assets that the financier can keep in case the borrower defaults. Collateral reduces the problem of uncertainty, since the lender can theoretically recover some, or all, of h s loan in the event of non-payment. It also reduces information asymmetries-it i s often easier to value physical assets than to value character. Moreover, the borrower will find it costly to put valuable collateral if she intends to decamp with the proceeds of the loan, because she will lose the collateral. Thus, the collateral requirement can also help weed out the rogues from the potential honest borrowers, leaving only those bonajde applicants who fully intend to repay the loan. T h e potential loss of her collateral also makes the borrower think twice before investing in risky ventures. Collateral's twin effects, of keeping rogues from applying for loans, and reducing the borrower's incentive to take undue risks, make it a valuable device in encouraging lending. T h e potential financier sees lower risk in collateralized lending, while the borrower benefits from the consequent lower interest rate the financier charges. O n e problem in all t h s , of course, i s that the rural poor typically do not have collateral, so they lose out once again. Most of India's rural poor, for instance, have no futed collateral or only small plots of land that most often cannot b e mortgaged. Identification of alternative collateral i s costly and cumbersome. So only those with assets can borrow! Another problem i s that collateral can only provide security to lenders in an environment where households have proper titles to their assets, and where the legal system makes it relatively straightforward for lenders to enforce contracts and repossess collateral; the legal system in India makes collateral registration, and i t s repossession by the financier, a long and arduous process. 19 T h e transactions costs of rural lending in India are high, mainly due to the m a l l loan size, high frequency of transactions inrural finance, large geographical spread and heterogeneity of borrowers, and widespread illiteracy. For private sector banks - their lack of a rural branch network i s an additional problem. Given the extent of rural poverty in India, the value of financial services required tends to be small. The small size of rural loans in India resulting in a high due dhgence cost per loan, exacerbated by the heterogeneity of borrowers, making it difficult for formal financiers to cover such costs. T h e geographical spread of customers in rural India, and widespread fiteracy, further drive- up the a h s t r a t i v e costs, after the loan is granted. Borrower supervision costs are high, as are compliance costs for customers. Financiers thus have to achieve a delicate trade-off between minimizing the loan default rate and minimizing a h s t r a t i v e / c o l l e c t i o n costs. For example, micro- financiers are able to offset the high maintenance costs of their loans with low default rates, whereas low supervision costs of rural loans is generally associated with high default rates. Furthermore, h t e d or no price discovery (only local price benchmarks, no national grades and prices for agricultural commodities, very limited dissemination of prices, brokers have information monopoly) makes credit risk assessment of rural borrowers more difficult, and credit information on such borrowers i s non-existent. Government has n o t been able to develop and enforce a legal and regulatory framework conducive to rural finance, so that contract design, contract renegotiation, and contract enforcement remain weak, making it even more difficult for financiers to provide borrowers with the right incentives for repayment. While the recent enactment of the securitization and asset reconstruction law (2002) has helped improve the legal framework for recovering bad loans, by facditating out-of-court settlements on non-performing loans and instituting alternate methods of dispute resolution between creditors and debtors, the law does not cover small loans. L a n d tithng and registration systems are weak, and the use and transfer of land i s difficult under the current framework. M a n y states do not permit leasing of land, while in some states, the lease creates long-term irrevocable rights for the lessee to the disadvantage of the lessor. All this encourages unrecorded and unofficial year-to-year oral leases, which prevent the lessor from getting a good lease value, and the lessee from puttingthe land to best use. Government policy has created a "financial clunate" that i s not conducive to lendmg in general, and to rural banking, in particular. High fiscal deficits, the government's domination of rural finance institutions, persisting weaknesses in the regulatory and legal framework, and a set of policies towards the sector that have been designed to gain politicalpatronage, have resulted inthe distortion of risk/retum signals and inefficiencies inthe delivery of rural finance services. An outcome of these realities has been a dilution of the credit creating role of rural banks. First,higbjscal deJzcitshaves led to Government's appropriation of a large share of financial savings for itself, preempting credit to the private sector. At the same time, Government's deficit financing policies have provided bankers with opportunities to deploy bank resources in Govemment securities, which are not only safe, but also, have yielded high profits for banks in a declining interest rate environment. 31 W e it is true that statutory preemptions (Statutory LiquidityRatio (SLR) and 3' Indian bank managers have found that they can in fact make large profits from the trading of Government securities in an environment of declining interest rates (and bankers inIndia do not appear to have a full appreciation of the interest rate riskinvolved insuch investment allocation decisions, viewing G-Secs as risk-free assets). Income on sale of investments has 20 Cash Reserve Ratio (CRR) imposed on banks have been gradually reduced in recent years, they remain high by international standards32 That banks prefer to deploy resources in Government securities rather than lendmg to the private sector i s evident from the data: At end-September 2003, investments in Government securities accounted for some 45 per cent of net demand and time liabhties of commercial banks, much higher than the mandated SLR of 25 per cent. Bank credit to the private sector i s thus modest, at best, and the bulk of it takes the form of relatively safe, mortgage-backed lendmg, other types of consumer loans to high-income individuals, and loans to top-end corporates. Directed lending norms that require commercial banks to allocate 40 per cent of their lending to the "priority sector" (including agriculture) have not generated the intended results, as most banks get around this requirement by subscribing to other eligible instruments including bonds issued by NABARDand SIDBI. Second, interest rate policies reduce the attractiveness of lending to small, rural clients. Banks' borrowing costs are kept high by " f l o ~ r ~on short-term deposit rates33, while Government policy dictates that " lending rates on small loans (Rs 200,000 or below) in the rural sector be "cupped' at the Prime Lending Rate (PLR), which banks are free to set. These restrictions impose an "implicit tax" on banks: Assuming a differential of 200 and 400 bps between PLR and market interest rates, the loss inthe income ofruralbanks fromlowerlendingrates is estimatedat $550 millionto $1.1bilhon; ths implies that the net profit of the banking sector could b e as m u c h as 15-30 per cent lower than what it could have been. Naturally, t h s reduces the attractiveness of rural lendmg, and particularly to smaller clients, with the unintended consequence of "rationing" credit to the poor, rural households. 34 Thus, the access of poor borrowers to formal loans is effectively cut off, and they end up borrowingat much higher rates fromthe informal sector. Third, RBI's creditplanning policy, whereby each rural branch i s given a set of defined villages (typically 15-20 vdlages) within w h c h it can operate, (`service area' approach) restricts competition in rural banhng. Further bank branches are unable to optimize branch infrastructure and entry of new, non- service area bank branches, including that of private sector bank branches, in the "service area" i s impeded since t h ~ srequires a `no-objection certificate' from the service area branch w h c h i s often not easily forthcoming. been risingsteadily over the past 5 years accounting for around33% of operating profits in FY 2002-03. Simultaneously the share in total income of interest income on advances has fallen to levels below 40%. RBI, Trend and Progress of Bankingin ''India, 2003-03. SLR has declined from 38.5% in 1991 to 25% at present. Similarly the CRR has declined from 15% to 4.5% over the same period. However, the overall statutory pre-emption levels are higher than levels in other countries - South Africa (So/,), Malaysia (15%), Singapore (18Yo); Sri Lanka (20%), U S A (none). 33 A floor/fmed rate of 3.5% is set for savings deposits. 34 Such difficulties can imply that sometimes, entire communities, and particularly the rural poor, may face limits on credit. Such theories of `credit rationing' have been discussed for example b y Stiglitz and Weiss (1981), Williamson (1986). Besley (1994) and Murdoch (1999) have discussed this specifically in the case of communities such as small farmers and microfmance providers. 21 Fourth, Government's domination of/interference in rural banks (Box 2) has further distorted bankers' incentives and generated deep inefficiencies in the rural finance institutions. Of the 196 R R B s in the country, some 15 per cent were loss-making and 58 per cent were estimated to be undercapitalized35 for the financial year endmg March 2002 (Figures 12-13). RRBs' operating costs relative to average assets are high at about 3 per cent, compared to less than 2 per cent for leading commercial banks. And staff productivity i s low - business/staff i s Rs9mwhich i s less than half that of the larger public sector banks and between a sixth to a tenth that of the levels in leading private sector banks. Asset quality i s poor, with Non Performing Loans (NPLs) at 16.5 per cent of total loans (refer to Annex 3 and Figure 14). And while the profitabiltty of these banks, reported at 1.2 per cent of average assets, appears good, the earnings forecast for these banks i s particularly susceptible to market andinterest rate risks. Box 2. Ruralbanking government's omni-presence - T h e public sector banks @e., majority government-owned) account for 80% of commercial bank assets in India. Some 95% of commercial bank branches in rural areas are of public sector banks. And these account for 93% of total rural credit outstandmg. In terms of governance, the boards of all public sector banks have majority representation from the government/RBI. However, with banlung reforms and increasing number of public sector banks being listed in the capital markets (12 of the larger public sector banks are listed), governance standards have improved significantly in conformity with the requirements of the capital markets regulator, Securities and Exchange Board of India (SEBI). RRBs are also majority government-owned. T h e ownership structure of these banks i s as follows: Go1owns 50% of the capital of an RRB, the `sponsoring' public sector commercial bank owns 35% (except for two R R B s that have a private sector bank `sponsoring' the 35% commercial bank stake), and the state govemment accounts for the remaining 15% of the capital. Four of nine board members are govemment appointees; in addition RBI and NABARD have one nominee each and the sponsor bank (which are mostly public sector banks) have three representatives. Rural cooperatives - For the State-Level Cooperative Banks (StCBs) the government ownership i s around lo%, whde it i s around 15% for the District Central Cooperative Banks (DCCBs); nonetheless, through the powers assigned to the Regstrar of Cooperatives, control of the state tends to b e pervasive. T h e state government, through the registrar of cooperatives has considerable powers and can supercede elected boards; As of end-March 2001, the elected boards of 75 of the 367 district cooperative banks had been superceded. T h e state government also retains the rights to appoint senior management - the CEOs of StCBs are often government officials (rather than professionals) and those for the DCCBs are either government or StCB officials. 35 By regulation, capitaladequacy standards have not been extended toR R B s or cooperative banks. However, even by the lenient requirement of Rs 100,000 as minimum capital for cooperative banks, as the RBI Annual Policy Statement 2004 reports, at end-March 2004, as many as 143 out of 366 DCCBs and 7 out of 30 StCBs had not complied with even this requirement.Similarly61 out of 196RRBs hada networthlowerthan Rs 10d o n (March 2002). 22 Figure 12: Status of Rural Banks inIndia Figure 13: RFIs profitability (Return on - Assets, RoA) _ . 70.0% i I 1.5% 4l 60.0% CI x E 2 P -.OX < 50.0% s 40.0% ,.5% I 0 30.0% x0 20.0% 0.0% 10.0% -0.5% 0.0% -1.0% UnprofRable Under-capitalized Note:RRBs (196) DCCBs(367) Source:RRB/Coop.,NABARD(RoAvgAssets) ; StCBs:(29) RBI : PSBdata (RoTotalassets) Figuresmbracketsrepresenttotalnumbersof ruralbanks Source: RBIandNABARD Figure 14: NPLs inRFIs Cause for ... ......."concern..."...""..""...... - Figure 15: Deposits of RFIs costs and - significance 35.00%7 -1 I"""" I H 800% 80x 9 I 700% 70% ; 600% 60% 'P 15 00% -z2: 500% 50% d * 400% 40A 10 00% c 300% 30% 5 00% 11 -:: H a 20 0% 20% .... .. 100% 10% u 1337.38 1338'33 1333'00 2000.0i 2001.02 0 0% 0 0% PSBs-FYO2 RRBs-FY02 StCBsFYOl OCCBsFYO1 I assets DepositslTotdl +Cost Source:RBI reportsandNABARDstatistics ofdepositsrelative t0aVgaSSetS Source:NABARDandRBI T h e rural cooperative bankmg sector i s in an even deeper state of distress.36 Some 32 per cent of DCCBs and 22 per cent of the StCBs are unprofitable and over two-thirds of the DCCBs and just under one-half of the StCBs are undercapitalized. (Annex 3 provides more details; Figures 12, 13 37 38). Asset quality i s quite poor (Figure 1439) with the StCBs reporting an NPL ratio of 13.4 per cent, and the DCCBs reporting an NPL ratio of close to 20 per cent. While the DCCBs' have a reasonable operating cost ratio of a little over 2 per cent, staff productivity at Rs 12m i s weak (the StCBs' 36 Commercial banks' rural operations are not analyzed separately as these are subsumed in their overall operations. However,as some indicators on asset qualityindicate, their overallperformancedisaggregated for the rural operations, may not be too differentfrom that of RRBs and cooperativebanks. 37 Given the politically sensitive nature of rural credit, with falling interest rates in recent years, announcements by successive governments, typically suggesting interest rate caps on agricultural credit, have not been uncommon. The last major announcement was made around a year ago and suggested a cap at 9% on farm loans less than Rs50,OOO (around $1,200). While these announcements do not get mandatedby RBI policy, such moral suasion provides the wrong signals for ruralbanks and increases the chances that couldlead to ruralbanks being forced to adopt such pricing- if they do so, the sugestedrateswouldprobablyleavean inadequate, or even negative financial margin. 38 While many of the public sector commercialbanks are now listed companies raising capital from the public and having independentand professionalboards and are relatively free of the adverse impacts of government ownership, this is not the case with the R R B s and ruralcooperatives. 39 For commercial banks, a proxy for ruralasset quality has been taken as the NPL ratio on the priority sector advances of public sector banks; the public sector banks account for more than 90% of the outstanding advances in rural areas of commercialbanks. 23 performance on these indicators i s better but is really not comparable to retail banks since the former are structured as apex institutions). Further, the weighted average cost of funds for the cooperative banks - 8.3 per cent and 8.6 per cent for StCBs and DCCBs respectively (Figure 15) - i s very hgh and leads to compressed financial margins. Overall the combined effects of moral suasion from the government to keep lending interest rates low, poor asset quality, low abhty to enforce recoveries and high cost of funds has led to the present tenuous position in the balance sheets of rural cooperative banks. Government ownershp/control also means that R R B s and rural cooperatives have to operate inline with government diktat and are not able to take decisions independently. Some examples of this include: forced rescheduhg of agmultural loans and `coerced' lendmg to unviable government owned/controlled commodity cooperatives (especially, the sugar and textile cooperative sectors), state governments' use of cooperative banks as channels for the delivery of various govemment schemes and debt or interest waiver schemesa, and elected boards being superceded. Rural debt or interest waiver schemes announced by state governments also create adverse incentives for farmers, who do not service the debt in a timely manner on the expectation that, at some point, a waiver will be granted. All these factors hinder proper management and operations of rural banks and reduce the attractiveness of rural lending for these banks. F$h, lax regulatory standards for R R B s and rural cooperative banks, the poor enforcement of prudential regulations, and regulatory forbearance, have undermined market discipline and added to the financial fraghty of these banks, thereby impeding their abhty to perform the task of efficient rural financial intermediation. W e the RBI i s the overall regulator of the rural finance sector, supervision of R R B s and rural cooperatives i s delegated to NABARD.However, with regard to the regulation of cooperatives, there i s dual control - the state through the Registrar of Cooperative Societies also plays a large role. T h e state government retains significant powers inrural cooperative banks, controlling all matters relating to registration, membershp, election, financial assistance, loaning powers, business operations, loan recovery and audlt. This leads to cross-directives, inadequate levels of control of the central bank with respect to banking functions and consequent weakening of the overall qualtty of regulation of the cooperative banks. And whde regulatory standards have been progressively tightened for commercial banks and brought in line with best practices, this has not been the case for the R R B s and rural cooperative banks. Prudential regulation standards related to capital adequacy have not been applied for R R B s and cooperative banks (there are no minimum capital adequacy ratios prescribed) - given the large percentage of rural banks that are inadequately capitalized (Figure 12 above) this means that a significant proportion of deposits of rural banks i s potentially at risk. Also, asset classification standards, and related to h s regulation on income recognltion and provisioning, need to b e upgraded to match up with that of the commercial bankmg system w h c h bases itself on NPLs on a 90-day level. Given these weak regulatory standards, it i s perhaps not surprising that eligibllity norms from NABARD for a v a h g refinance assistance remain quite lenient -cooperative banks that have eroded deposits less than 50 per cent are eligible for refinance, as are R R B s which have a deposit erosion level of less than 30 per cent. NABARD, in its role as the apex development bank for rural finance has, in the past, provided subsidued refinancing to commercial banks, R R B s and cooperative banks. This could have led to some crowding out of market based credit to rural areas and importantly, reduced the incentive and pressure on rural banks to improve financial performance and move towards increased reliance on market based funding. However, with declining interest rates inthe economy, the incremental cost of 40 M o s t significantly the Agriculture and Rural D e b t Relief Scheme of 1989-90 but several other subsequent schemes, including, m o s t recently, the kharifinterest waiver in 2002-03. 24 funds for rural banks has declined faster than for NABARD.The result of this is that NABARD's typical refinancing rates of between 6.25-6.75 per cent pea.are no longer subsidrzed and demand for refinancing has deched, especially from the more efficient rural banks. W h y D o Small, Rural BorrowersFind Rural BanksUnattractive? From the perspective of small, rural borrowers (the users), rural banks are unattractive for the followingreasons: Small, ruralborrowers findrural banks unattractive for the following reasons: Rural banks do not provide flexible products and services to meet the income and expenditure patterns of small rural borrowers. As noted above, small rural borrowers have irregular/volatde income streams and expenditure needs, and therefore, prefer to borrow frequently, and repay in small installments, but most banks do not offer such products. Also, while small rural borrowers seek savings and lending products, they also seek insurance (llfe, health, crop), w h c h banks do not generally offer. `Iji i ~ ? . i z r i ~ i ~~~~i ~.. ~ / . ~ T h e transactions costs of dealing with formal banks are hgh. In part, high transactions costs stem from distance to the nearest financial institutions. Accordtng to the RFAS-2003, the median distance to the nearest financial institution ranges from 2 k m s @ost office branches) to 5 k m s (commercial banks, cooperative banks); the median time taken to travel to the nearest commercial bank, cooperative or RRB i s 30 minutes @ost offices are available at closer proximity). Procedures for opening an account or seeking a loan are cumbersome and costly (with high rejection rates). Government policy dictates that rural borrowers currently need to acquire a "no dues certificate" from every other lender in the village (as defined by local lenders) that they do not have a loan outstanding. Furthermore, clients have to pay hefty bribes (ranging from 10 per cent-20 per cent of the loan amount) to access loans, so that the ultimate cost to borrowers i s very high (despite interest "caps"). On average, some 27 per cent (and 48 per cent in UP) of sample households in our survey who borrowed from an RRB report having to pay a bribe to get the loan, a little under 27 per cent of households who borrowed from a commercial banks paid a bribe, and 10 per cent of households who borrowed from a credit cooperative paid a bribe. The bribe amounts appear to vary from anywhere between 10 per cent of the loan amount (inthe case of banks) to 20 per cent (in the case of cooperatives). Moreover, longer processing times for loans, together with bribes, could result in higher effective costs to borrowers and consequent credit rationing. I t takes, on average, 33 weeks for a loan to be approved by a commercial bank.4' Fable 15). 41 High transactions costs of dealing with formal banks translate into a low frequency of transactions. According to the WAS, 2003, more than 60% of households with bank accounts report accessingtheir accounts at a frequency of less than once a month (62.3% of households with accounts in banks access their accounts less than once a month, while the same percentageamongst householdswith accountsin R R B s is 70.8). 25 Bank RRB coops Schemes Others Interest rate (median) O/o p.a. 12.5 11 11 14 14 Loanamount receivedas Yo of amount 91.8 88.2 83.5 86.6 93.9 applied Percentagehouseholds reportingbribes 26.8 27.0 9.7 27.27 23.21 Bribeas Y o of amount approved 10.1 18.2 19.9 42.3 8.3 Time taken to process a loan application 33 28.5 24 8.9 14.3 Source : WAS-2003 A third factor that makes formal banks unattractive for rural borrowers i s that banks demand collateral, which poor rural borrowers lack. Indeed, the majority of loans extended by commercial banks, R R B s and cooperatives are collateralized, with 89 per cent of households who borrowed from RRBs, and 87 per cent of households who borrowed from commercial banks, reporting that they had to provide collateral (RFAS-2003). Landremains by far the most predominant form of collateral. But since this collateral i s seldom executed, it i s just another cost, with little benefit in practice. 26 IV.RECENT EFFORTS ININDIATO IMPROVE RURALACCESS TO FINANCE: THE ROLE OF FORMAL-INFORMAL LINKAGESAND NEWPRODUCTS $2 In light of the inefficiencies that characterize India's rural finance markets and the relative lack of success of the formal rural finance institutions in delivering finance to the poor, NGOs, financial institutions, and government have made efforts, in partnership, to develop new fmancial delivery approaches to service the financial needs of the rural poor. These approaches--or "micro finance" programs-have been designed to overcome some of the risks and costs associated with formal financing, and also to overcome the tyranny of collateral. They involve providing thrift, credit and other financial services and products of very small amounts to the poor, with the aim to raise income levels and improve living standards.43 They attempt to combine the safety and reliability of formal finance with the convenience and flexibillty that are typically associated with informal finance. W e some of these programs have been more successful than others, their h t e d outreach, scalabdity and financial sustainabihty remain matters of concern. SHG-bankLinkage Approach: Linking CommercialBanks to GrassrootsBorrowers O n e approach to microfinance that has gained prominence in recent years is the self help group (SHG)-bank linkage program, pioneered by a few NGOs such as MYRADA in Karnataka and Professional Assistance for Development Action (PRADAN) in Rajasthan (and later inT a d N a d u andJharkhand), with strong support from NABARD, which has been instrumental inpromotingt h i s groWth.4 SHG-bank h k a g e involves organizing the poor, usually 15-20 women, into self-help groups (SHGs), inculcating inthe group the habit of saving, linking the group to a bank (usually the rural branch of a commercial bank, but also RRBs, cooperative banks, etc.), and rotating the saved and borrowed funds through lendmg with the group. T h e SHGs thus save, borrow and repay collectively. T h e lenders (banks) are often refinanced by NABARD at slightly subsidized rates although, in recent years, highrecovery rates have encouraged some banks to lend to SHGs without NABARD refinancing. The SHGs are not formally registered. T h e funds may be distributed either to one or more members of the group - who are personally responsible for repayment (generally the group borrows at about 12 per cent per annum) - or spent collectively by the group. T h e group i s free to decide the interest rate charged to its members, but typically, a member borrows from the group at about 24 per cent per annum.45 T h e groups make their internal credit decisions, decide on the repayment period, etc. Money i s used for both consumption (health, marriages, etc.) and, over time, for individual and group investment products. After a loan i s fully repaid, the group may borrow again, often a larger amount. Banks typically provide a loan equivalent to four times the group's savings but, as the group matures, and based on the group's track record, banks are ready to lend more; some of the R R B s visited reported makmg loans amounting to 10 times the SHG's savings. 42 This sectiondraws largelyon the backgroundpapers by MorduchandRutherford(2003), Mahajan andRamola (2003) and Hess andMapper (2003). 43 While microfinancetypically covers the poor in rural, semi-urbanandurban areas, the focus here i s on the ruralpoor. 44 Guidelines from the RBIwere issued in 1992, to experimentwith a pilot of 500 SHGs to link with banks.This pilot programas well the work of a number of NGOs was reviewedby a WorkingGroup on BankLendingto the Poor through NGOs andSHGs (1995) and detailedguidelines were drawn to encourage banks to use this method. NABARD was given the task of leadingthis effort andit took to task with exemplary diligence. It involvedNGOs, commercialbanks, regional ruralbanks and even cooperativebanksin forming SHGs and then linking those up with nearby bank branches. 45 In contrast,money lenderswould chargeannualinterest rates ranging from 36 -120% p.a. The average interest rate chargedby money lendersis 48%. 27 T h e success of the SHG-bank linkage model depends critically on the tasks of promoting, nurturing, strengthening and monitoring SHGs - tasks that are performed by the Self H e l p Promoting Institutions (SHPIs). Traditionally, grass-roots level NGOs have performed the tasks of promoting and monitoring SHGs. M o r e recently, rural branches of commercial banks, cooperative banks, RRFh, NBFCs, etc. have all begun to play the role of SHPIs. But,recent evaluation studies reveal the comparatively better performance of SHGs promoted by NGOs (as opposed to the other SHPIs). SHGs require a large amount of pre- and post-lending monitoring. Most lenders use "facihtators" (who may or may not earn a commission and/or bonus for group repayment and recruiting new groups). Facilitators may be local business people, health workers, government employees, teachers, etc. Before a loan i s granted, the group must prove their abdity to save over time, learn bookkeeping skills, and show their commitment to continue as a cohesive group. It often takes over a year before an established group can borrow. T h e facilitator i s responsible for introducing the group to the bank, implementing savings patterns and teaching basic accounting practices. (Although the facditator may not accept money on behalf of the lender). After the loan i s made, the facilitator attends monthly meetings (attendance i s a good indicator of the continued success of the group) and enforces repayment. Over the last 10 years, the SHG-hkage model has become the dominant m o d e of micro finance in India, and the model has been successful in encouraging sigmficant savings and high repayment rates. T h e number of SHGs h k e d to banks has increased from just 500 in the early 199Os, to over 700,000 by 2003. The SHG-bank linkage program today reaches some 12 d o n women and their households in terms of deposit services, cumulatively providing over Rs 2,049 Crore (US$ 445 d o n ) as credt between 1992to March 2003. (Table 16). Table 16 :Growth involumes of SHG-bank linkage ny iviar I I ~ 1 1 1 ~ Number of SHGs linked to banks, Cumulative bank loans (Rs. JI ~~ cumulative nos. million) 1999 32.995 571 2000 114,775 1,930 2001 263,825 4,809 2002 461,478 10,263 2003 717,306 20,487 There appears to be widespread enthusiasm in India about the benefits of the S H G - h k a g e approach, notably because: (1) I t helps reduce transactions costs for the banks (their costs related to credit evaluation, loan monitoring and decisions are reduced, since banks can rely on the SHPI to identify and promote groups and pass-on the loan appropriation decisions to the group) as well as the borrowers (as the group itself provides constant watch and follow-up); (2) By using "peer- pressure", the approach increases the likelihood that individual group members drepay as well as that the group as a whole dnot default (other group members effectively proxy for collateral); (3) L o a n default rates are very low, on average less than 1per cent; (4) It empowers ruralwomen. Although data available to u s cannot answer this question du-ectly, we can look at some indirect indicators on the relationshp between SHGs and the poor using results of the WAS-2003 for the states of AP and UP.4 First, the majority of the beneficiaries of SHGBank Linkage are from among 46 The RFAS-2003 targeted roughly 60 villages in each state, though the actual number of villages sampled was higher in UP. Data from WAS-2003, which was a household survey, include responses from those who were members of SHGs, as 28 the poorer groups. WAS-2003 indicates that nearly 54 per cent of SHG members are from the poorest groups-the landless and marginal farmers. However, significant differences exist across states. While 73 percent of SHG members inUP are from the poorest households - the landless and marginal farmers, the c o r r e s p o n h g proportioninAP i s lower (43 percent). InAP, though, almost a quarter of the poorest households belong to SHGs, but the proportion of households from richer categories that belong to SHGs i s much hgher. F a b l e 17). Table 17. SHG membership by type of households Marginal Medium Large Others Total AndhraPradesh SHGmember (Yo households) 29 29.6 YOof SHG members 11.4 Uttar Pradesh SHGmember (Yo households) 10.3 4.1 O/Oof SHGmembers 72.8 8.8 7.4 4.4 5.7 100 I Marginal = landholding less than 1acre, Small = 1-2acres, b :&m 2-4 acres, Large = more than 4 acres. = lthers includes household with or without landbut involved incommercial activities. Source: RFAS-2003 To explore further the success with w h c h SHGs are targeting the poorest, we used the WAS-2003 data to examine the attributes of households that are SHG members. Our results indicate that that SHG Bank Linkage is certainly quite effective intargeting poorer households (inthe particularly the two quintdes above the poorest households). The relationshp between the poorest households and SHG membership is positive too, though not significant statistically. This only further underhes the challenges ahead in expanding the outreach of microfmance to the poorest. (Box 3). well as vdlage-leveldata. Basedon the latter, a total of 736 SHGswere inoperationin the villages coveredby RFAS-2003, with the overwhelmingmajority in AP as would be expected. Not only does AP have the preponderanceof SHGs, these groups are overwhelmingly comprisedof women only. Women's groups accounted for 95 percent of all SHGs inthe state. By contrast,inUP groups comprisingonly menand mixed groups were alsowellrepresented. 29 Box 3. How successfully has SHG bank Linkage targeted the poor? The estimated coefficients are reported in Table 18. Probit estimates of likelihood of a household terms of the incremental impact on the being a member of an SHG probability o f the household being an dF/dx s.e. 2 S H G member due to a small change in Shgnum 0.002 0.001 2.27 the explanatory variable. Thus, an Coastal -0.045 0.017 -2.58 increase in the number o f SHGs in the Poorest 0.039 0.030 1.32 d a g e by 1increases the probability o f a Quint2 0.072 0.028 2.67 household being in an S H G by 0.2 Quint3 0.102 0.034 3.13 percent. Similarly, households in the second quintile have a 7.8 percent higher Quint4 0.039 0.028 1.41 hkelihood of being in SHGs compared to Bankac -0.042 0.019 -2.23 households in the richest quintile. -0.007 0.002 -3.46 Although the overall fit o f the regression N= 2910 PseudoR2= 0.011 W e can also use the WAS-2003 data to answer a related question using villages and SHGs in AP, where SHGs were found in all villages: Arepoorer villagesmore Like4 to have a lager number ofSHGs? What explains inter-village variations in SHGs? One approach i s to use the standard Poisson model for count data to analyze the number o f SHGs in a given village. Specifically, let y, denote the number o f SHGs in village i and let yl be distributed as a Poisson with mean y,where PI= exp @l P) and X, i s a vector o f explanatory variables. However, a preluninary look at the data shows an over dispersion in number o f SHGs relative to the Poisson, with the variance substantially higher than the mean. To account for the over dispersion, we use a negative binomial distribution that can be viewed as modifylng the equation above to pl= exp (X,P + u,) where u1denotes some omitted variable(s) such that eu, follows a Gamma distribution with mean 1 and variance a. Larger values of a imply greater dispersion in the data. 47Friends and family are a major source of informal borrowings. 30 Results of the negative binomial regression are Table 19. Negative Binomial Regression Estimates for provided below using a number of explanatory SHGs ina village, AP variables related to poverty in the village - village Coef. z P>Z average per capita income (avpcy), inequality in Constant 5.96 3.98 0 land holdings (giniland), and connectivity o f the Size 0.002 6.15 0 village measured by distance to nearest railway Illiterate -0.04 -2.25 0.024 station and nearest metal road (rail & road p-ary . -0.07 -3.88 0 respectively). Poorer villages will have lower per A T Y 0.00 0.07 0.943 capita incomes, while greater inequality for any g v e n per capita income would indicate more poor Giniland -1.47 -0.94 0.348 people in the village. Villages farther away from Ngo -0.33 -1.60 0.11 access to rail and road, and hence markets, may be Rail 0.01 1.43 0.154 relatively poorer. The percentage o f illiterate Road -0.06 -1.95 0.052 households in the village i s also likely to be -0.03 -2.01 0.045 correlated with village poverty and i s included. In COOP 0.02 2.29 0.022 addition, we also include an index o f educational Alpha 0.22 0.06 3.49 attainment for villagers who are not illiterate in the N 59 Pseudo R* 0.12 form Of percentage having primary education' Other are village size in fikelihood ratio test ofa@a=O: &2(1) = 50.79 Prob > chi2 = 0.0000. Robusl estimatesusedjir standarderrors. terms of number of households (size) and whether The results suggest absence o f a strong relationship between village-levelindicators of poverty and the number o f SHGs. The size of the village is significant indicatingthe larger the village, greater the number of SHGs found. K g h e r rates o f illiteracy, presumed correlated to poorer villages, are associated with fewer SHGs in the village.48 Similarly, the negative sign for (distance to) metal road implies villages in the interior, away from good roads, have fewer SHGs. Noticeably, neither village per capita income nor inequality within the d a g e i s significant, again underlining the weak link between poverty and presence of SHGs. The coefficient for NGOs is negative but insignificant, implying no significant relationship between presence of NGOs and number of SHGs in the village. Recent analyses indicate that access to loans under SHG bank Linkage has contributed to the reduction in vulnerability of poor households.49 This reduction in vulnerabdity takes the form of: (1) Improvement in asset position: the program sigmficantly improved the asset position (comprising livestock and consumer durables) of sample households. T h e average increase in assets was about 72 percent, from Rs 6,843 to Rs 11,793 in real terms (in one to three years). About 59 percent of households saw assets increase after groups were formed. Before the groups were formed, one in three households had no assets; after the groups were formed that changed to one in six; (2) Increase in savings: the average savings per member more than tripled, from Rs 460 before the group to Rs 1,444 after; (3) Cbanges in borrowing patterns and activitiesjnanced average borrowing per household increased from Rs 4,282 to Rs 8,341. A shift was observed in the activities of the self-help groups, with a lower share of consumption and cultivation loans after the groups formed and a larger share 48 However, this would also be consistent with the notion that greater illiteracy makes more difficult the formation of SHGs. 49 See, for example, Puhazhendhi and Satyasai (2000) and World Bank (2003). 31 of allied agricultural activities and small businesses; (4) Increase in employment employment per household went from an average of 318 days a year to 375 days. T h e proportion of employment generated through nonfarm and off-farm activities increased; (5) Increase in conszrmption expenditwe: consumption expenditure per household per month increased from Rs 799 to Rs 993. Per capita consumption increased from Rs 197 per month to Rs 249; (6) Impact on income: the average net income per household increased from Rs 20,177 to Rs 26,889. About 43 percent of the incremental income generated was from nonfarm activities; (7) Impact onpove3: about 234 households were below the poverty line before groups were formed, compared with 122 after; (8) Social impact 89 percent of members reported that, as a result of the group's activities, they could meet officials from the government or from banks, whde about 77 percent had never had that opportunity before. Changes were also reported regarding attitudes toward women. Further work should b e conducted on the performance of the program from the point of view of the financial institutions that provide the h k a g e . This would provide lessons on the costs and benefits associated with the self-help group model, including which entity in the lmkage covers what costs, what financial benefits banks get from the self-help group program, potential adjustments and innovations banks have made to the self-help group model, and incentives that motivate them to b e involved in the program and in microfinance in general. I t will b e particularly important to analyze private banks to draw on their experiences and assess their potential to play a larger role in microfinance in India. Inlarge part, the success of SHG Bank Linkage may be attributable to the fact that it is well aligned with Indian history and circumstances, and capitakes on the country's vast network of rural bank branches. T h e idea of local savings-and-loan clubs enjoying access to formal financial services by becoming corporate customers of banks i s a good one and i s practiced in a small way in many countries. A well-run club can keep its reserves at the bank and take bulk loans w h c h it can on-lend to its members at a premium, coveringits costs and rewarding its savers in the process. InIndia, thls practice seems particularly appropriate on two counts: First, the country has active NGOs that have been zealous in their efforts towards group formation; NGOs view SHGs as having many benefits (such as women's empowerment) beyond microfinance. Second, and perhaps even more important, SHG Bank Linkage seems particularly suited to India because the model capitalizes on the country's vast (and unique) network of rural banks that are otherwise unable to reach out to the poor. However, as the analysis above indicates, the S H G Bank Linkage program faces important challenges inexpanding outreach and reaching the poorest quindes of ruralhouseholds. Equally important, the success of S H G Bank Linkage underscores just how important a role slullful leadership, good policy, and a conducive legal and regulatory framework can play. Indeed, the role of government in establishing the necessary policy and legal framework, and the leadershp role assumed by NABARD in championing the movement, cannot be underestimated. Government recognized the potential of SHG banking very early on in the movement's history; NABARD was given the task of leading this effort and it took to task with exemplary diligence. It involved NGOs, commercial banks, R R B s and even cooperative banks in forming SHGs and then hktng those up with nearby bank branches. Legalobstacles were removed and the RBIissued Guidelines in 1992, to experiment with a pilot of 500 SHGs to link with banks. This pilot program as well the work of a number of NGOs was reviewed by a Working Group on Bank Lendmg to the Poor through NGOs and SHGs (1995) and detaded guidelines were drawn to encourage banks to use this method. To encourage banks to lend to SHGs, NABARDmade available subsidized refinancing to banks for their lendmg to SHGs, so that the groups could take bulk loans from banks that could be on-lent to group members who could use them to take up or expand microbusinesses. (More recently, such 32 subsidrzed refinancing to banks has been phased out, as banks have begun to see SHG lendtng as a profitable and commercially viable business). Further inducements for banks came in the form of allowing banks to count SHG lending towards their legal obligation to direct a fraction of their loans to the poor (`priority sector lending' obligations). It seemed an ideal way to realize an old Indian dream - to make the vast network of rural banks key suppliers of loans to the poor. However, outreach remains limited in terms of the number of households served, and the scale has been modest, in terms of the volume of credit outstandtng and average size of loans. T h e program has reached only about 12 d o n women and their households (in a country where 460 d o n people live on less than $/day) in terms of savings accounts and an estimated 2-4 million women in terms of outstanding credit accounts. The outstandmgs of the SHG program in March 2003 were around Rs 10 billion ($ 217 d o n ) , thus catering to 2.2 per cent to 6.6 per cent of the estimated demand and amounting to around 1.9 per cent of rural outstandmg bank credit. L o a n amounts remain small. In 2003, the SHG member households got an average of R s 1,766 as credit, after being ina group and meeting monthly for anywhere between 9 to 24 months.In2002-03 only 22 per cent of SHGs existing at the beginning of the year, received loans during the course of the year - even assuming a two year loan period, this means that more than half of the existing SHGs did not receive a loan during the year. And the program remains concentrated in the South of India, with nearly 75 per cent of funds flowing to SHGs in the four southem states. T h e following factors could explain the constraints to scaling up the SHG-Bank linkage program: A key constraint is the lack ~capaitytopromoteand maintaingroups to enszire quality. Inthe early phase of the SHG movement, the tasks of promoting, nurturingand strengthening groups50, were performed by grass roots NGOs. However, in NABARD's zeal to link a cumulative total of one d o n SHGs to banks by 2008, and as SHG Bank Linkage has caught the attention of politicians who view the program as an easy vote winner, quantitative targets on the number of groups to be promoted each year are overriding concerns about the quality of the groups promoted. M a n y of the recent groups have been promoted by institutions that either lack the required skills and local knowledge or ones that are driven by short term monetary incentives. M a n y groups have come together on an ad hoc basis, only because they want a loan. Inadequate attention to group quality could threaten the longer term credibihty and viabihty of the entire program. Indeed, recent evidence suggests that the quality of groups is already b e p n i n g to suffer. A recent Mahlla Abhviniddh Society, Andhra Pradesh Survey (APMAS) in 2002 indicated that only 17 percent of all groups were of adequate quality for bank linkage and this was in a state which i s considered the leader in the movement. Thus, lack of good quality SHGpromoters -withincreasing use of non-traditional promoters of SHGs, affects the quality of group formation by such entities. Thus, while scale objectives may b e attained in the short term through using non-traditional SHPIs, long term sustainabdity and quality remain as issues to b e tackled. A second problem relates to the cost Ofgroupformation, and the h e taken. Promotion of good quality groups requires an investment of both time and money. Someone has to incur the cost of promoting groups (organizing meetings, training the members). The estimate of this cost i s controversial, with NABARD claiming it to be as low as Rs 1000 per group and NGOs saying it takes as much Rs 12,000. The Muustry of Rural Development has established a norm of Rs 10,000 per group, w h c h experts claim i s realistic. Thus, reaching NABARD's internal target of forming an additional one 50 These tasks include inculcating in the groups a culture of savings and repayment, teaching them bookkeeping skills, strengthening their internal capacity to undertake administrative tasks (accounting, meeting minutes, correspondence, and negotiations with bankers) and commercial activities (business start-ups, marketing, and reinvestment), ensuring the groups remain financially sustainable and have the ability to weather personallosses (accidents, sickness, death), and natural disasters, etc. 33 million groups by 2008, wouldrequire an estimated Rs 10 billion. Where these funds will be sourced from remains unclear. Moreover, even after a group has been promoted, continuous efforts are needed to monitor these groups and strengthen their internal capacity to undertake adrmnistrative tasks (accounting, meeting minutes, correspondence, and negotiations with bankers) and commercial activities (business start-ups, marketing, and re-investment). Efforts are also needed to ensure the groups remain financially sustainable and have the ability to weather personal losses (accidents, sickness, death), and natural disasters. LeadlngNGOs that have been engaged inthis activity indicate that it takes a minimum of three years of nurturingbefore a group i s ready to b e ltnked to a bank. Third, issues concerning the model'sJl;nuncials~stu~nubil~~ face of pressures on banks to lend to inthe SHGs at subsidized interest rates could constrain the further growth of SHG Bank Linkage. Banks have been lendmg to SHGs at interest rates of between 12 per cent and 12.5 per cent. Recently, two state-owned banks, the State Bank of India and Andhra Bank, have announced their intention to lend at 9 per cent per annum, viewing SHG lending as a highly profitable business. Recent studies, however, indlcate that the all inclusive costs of lendmg to SHGs are in fact m u c h higher than what state-owned banks seem to think, and could range anywhere between 15 per cent (which i s what private banks like ICICI Bank charge when they lend to SHGs) to 28 per cent. In a study of five RRB branches, Sinha (2003) shows that the all inclusive costs of lendmg to SHGs ( t a h g into account the relatively high transactions costs of dealtng with SHGs as well the costs of group formation, which bank are increasing beginning to bear) would translate into interest rates of anywhere between 22 per cent and 28 per cent per year, and in one case, where the RRB was located ina low density, forested district, the costs translated into interest rates as high as 48 per cent per annum. The SHG portfolio i s a small part of the total bank lending, portfolio quality is good, and it may b e possible to cross subsidize this, but unless banks charge interest rates to recover costs, the model's financial viabdtty and longer term sustainabdtty may be jeopardized. Microfinance Jtistitutions (MFls) A second approach to microfinance involves delivery of finance to the poor through the creation of specialized MFIs. This effort has been led by the SIDBI Foundation for Micro-Credit (SFMC) and other apex lending institutions, including the RMI< and FWWB. M u c h of the credit of the growth of the sector has been pioneered by these institutions. SFMC, as the largest player, led a number of innovations includmg mainstreaming the microfinance sector, facdttating ltnks to commercial banks and lenders, promoting better practices amongst MFIs through its capacity bddlng program, pioneering support to rating initiatives and n e w instruments includmg the recently launched transformation loadquasi-equity product. As a result of such initiatives the MFI sector has grown at a fast rate over the last six years or so. Some of these M F I s are based on the Grameen model. Others promote and establish financial hks with SHGs. T h e M F I s offer some of the features of the informal sector such as flexible products, customer friendly practices but at a higher interest rate51 than formal sector, while bringingin some features of the formal institutions - such as documented loan contracts, detailed books of accounts, MIS, staff, and some degree of supervision by a regulatory authority. 51 M-CRIL's 2003 Micro Finance Review presents the Annual Percentage Ratio (APR) of rated Indian MFIs as 24.3%. There are signs that with competition and growing efficiency in operations, interest rates of MFIs can reduce over time; Spandana, a leading MFIinAP, amongst the largest inthe country, reduced its lending rate from 18-20% p a flat to 12-15% pa flat and has plans to reduce it further on account of operational efficiency gains with growing scale of operations; ASA, a leading MFIinTamil Nadu also reduced its interest rate from a high 24% pa flat to 15-16% pa flat that i s close to the rates that SHGs themselves charge each other for internal lending. Other examples can be found particularly from these two states, where the degree of competition between MFIs i s strongest. 34 Over the past decade, the number of MFIs in India has grown. However, with a few exceptions, most M F I s in India are very region specific, small in size-the largest being S H A R E Microfin Ltd with loans outstanding of about Rs 50 crore andBASIXof Rs 35 crore inMarch, 2003-and their collective outreach has been h t e d . In March 2003, the M F I s sector as a whole had outstandmgs of Rs 240 crore ($52d o n ) reaching less than one d o n borrowers. O n e estimate indicates that the average loans disbursed by the top 10 M F I s amounted to just Rs 16 crore per MFI.Another estimate, based on 69 rated MFIs (which are among India's top 100MFIs), shows that these M F I s had about 6,500 borrowers and Rs. 2.3 crore loan outstanding, per MFI. (Sinha, 2003). In comparison, MFIs in Bangladesh are estimated to reach m o r e than 60 per cent of the poor in the country with the larger programs such as Grameen Bank, BRAC, P r o s M a and ASA all reaching well over one d o n clients each. Grameen Bank`s loan portfolio alone exceeds that of the entire microfinance sector in India by a factor of five whereas BRAC's portfolio i s more than three times that of all MFIs in India; both ASA's and Proshlka's portfolio i s also greater than that of the entire microfinance sector inIndia. In addition to the relatively small scale of their operations, Indian MFIs also tend to have a h t e d scope. D u e to regulatory reasons, only a handful of MFIs, such as Vivekananda Seva Kendra 0 Sishu Uddyan, VSSU y e s t Bengal) offer savings as a service. Apart from promoting mutual savings among groups (SHG or Grameen type), a few NGO M F I s offer savings services by taking deposits from their members. Others have had to use Mutual Benefit Trusts or Mutually Aided Cooperative Societies (MACS). Only the SEWA Bank, Ahmedabad and the BASIX Local Area Bank ICBSLAB(inthree districts ofAP and Karnataka) offer savings as RBIregulated entities. T h e h t e d outreach and scale of Indian M F I s , relative to the MFI giants in Indonesia and Bangladesh, reflects, at least in part, the absence of an enabling policy, legal and regulatory framework. M F I s suffer from the fact that their regulatory oversight i s fragmented across many government agencies. MFIs are not allowed to mobilize deposits (even from their own members) unless they convert themselves into an NBFC. And even as NBFCs, an `investment grade' rating from corporate rating agencies is required for mobilizingdeposits. T h ~ iss difficult for most MFI- NBFCs; based on past examples, on account of the typically geographically concentrated and non- collaterahed portfolios that MFIs have, rating agencies, in almost all cases, have not assigned the required credit rating. T h e minimumstart-up capital requirement for registering as an NBFC (Rs 20 d o n or US$450,000) is typically beyond the reach of most MFIs. Similarly, the minimumcapital requirements for insurance companies (Rs 1billion, or US$23 million) are high. MFIs have problems raising equity: NGOs are not allowed to invest in MFI equity, because of the charitable status of NGOs under the Section 11and 12of the Income Tax Act. Regulation on foreign dtrect investment (FDI) inM F I s dictates rather highminimum levels; foreign equity must be a minimumof $500,000 for FDIupto 51%, $5 million for FDIbetween 51-75'/0, and $50 million for FDI 75-100%. What's more, since 2002, MFIs are no longer allowed to raise debt from foreign donors and development finance institutions through the `External Commercial Borrowing' (ECB) route. Second, the cost of funds for Indian MFIs i s relatively high, and unlike in Bangladesh and a number of other countries, the Indian MFI sector has not benefited from grants/subsi&ed fundmg. Unhke in, say, Bangladesh, where Palli Karma-Sahayak Foundation (PICSF) lends to M F I s at 4-6 per cent p.a. (less than half the market interest rate), Indian M F I s , right from inception, tend raise debt (from SIDBI, FWWB or commercial banks) at market rates (between 11-13.5per cent p.a.). W e , inmany ways, this i s a more sustainable way to grow, in practice, the hgh cost of funds combined with 35 problems in accessing equity, has meant that achieving profitabhty and growth has been m o r e difficult for Indian MFIs than their counterparts in countries like Bangladesh. Third, the Indian MFI sector suffers from capacity and skds constraints, and inadequate support systems. As microfinance i s a specialized activity and given that many MFIs have evolved from NGOs that have otherwise been focusing on grant based activities, staff tend to have stronger inclination towards social development issues and tend to possess limited s k d s in finance, accounting and business management. Thus, sensitization to issues hke internal controls, importance of credit dtscipline amongst groups/members, MIS, financial control and management, financial analysis, business planning, systems development, n e w product design, etc tend to be of relatively low quality. MFIs need considerable technical assistance to scale up skills inthese aspects. Fourth, most M F I s in India lend to SHGs.This means that MFIs in India are constrained by many of the same factors that have held back the outreach and scale of SHG Bank Linkage. In particular, capacity, time and cost issues related to group formation have posed constraints. The "Service Provider" Model of MicroWnance Piloted by Privatc Banks InJanuary 2000, the RBIallowed banks to lend to MFIsand treat this as part of their priority sector lending. Since then, a number of banks have used this opportunity to lend to MFIs, mainly NGOs, and all banks now offer lines ofcredit to MFh in addition to term loans. This enables MFIs to drawn down the loan at the pace they build their portfolio, thereby reducing the effective interest payment. Banks appear to have had a positive experience of lending to MFIs, where transaction costs for banks are lower as compared to lending to SHGs, and the repayment rates are 98 percent and above. Encouraged by early results, the n e w private sector banks, most notably ICICI Bank, but also UTI Bank and HDFC Bank, are actively seeking exposure in the microfmance sector. While their current exposure to microfinance i s too small to make a difference to their overall portfolio, even their priority sector lendmg portfolio, these new banks are pursuing new and innovative approaches to microfinance - as a potential business and not merely as a social or priority sector lending obligation. T h e various approaches to microfinance launched in recent years by ICICI Bank to reach rural borrowers are noteworthy. One approach involves linkingICICI Bank`s network of about 100 rural branches to SHGs; through this approach, ICICI Bank funds about 6,000 groups. To overcome the constraints faced by the lack of ICICI Bank`s rural branch network, ICICI Bank uses local "promoters" to help organize the groups. T h e interest rate on loans under this approach are about 18 per cent and promoters are paida salary that depends on recovery rates, size of loans, etc. Another approach piloted inareas where ICICI Bank does not have a physical presence involves the use of NGOs or MFIs52, traders, or local brokers (who are close to the farmer by the nature of their business), as intermediaries/"service providers" for loans to small and marginal farmers. T h e tasks of loan appraisal, processing, management, collection, etc. are delegated to the NGO/MFI but the loans are always on the books of the bank (ICICI Bank funds the borrower directly and the loan does not pass through the NGO/MFIs). ICICI Bank provides an initial loan to the NGO/MFI to develop SHGs, but then requires that the NGO/MFI repay the loan in a few years and become a "viable unit" through charging service fees to the groups directly. Ingeneral, ICICIBank charges the group 12 per cent plus the service provider charges 6 per cent, equal to 18 per cent. Another recent initiative taken by some private sector banks and insurance agencies to overcome the lack of rural branch presence has been the use, for a fee, of the vast postal branch office network as a means to 52 T h e I C I C I Bank haslaunched a pilot effort for this jointly with Cashpor Micro Credit, a section 25 company specially set up for this purpose by Cashpor Financial and Technical Services Ltd, in the Chandauli district of Uttar Pradesh. 36 provide financial services. T h i s includes channeling of insurance products and mutual funds and use of postal branch office space for settingupAutomated Teller Machines (ATMs). ICICI Bank, as well as other banks such as OrientalBank of Commerce, are also experimenting with an approach now termed as the "integrated agricultural service provider" approach. O n e version of t h s approach that could perhaps b e replicated on a wider basis, i s the ICICI Bank Farmer Seruice Center operating model (Mahindra Shubhlabh model). Under this model, ICICI has identified an integrated agricultural services provider, or LASP (Shubhlabh), that has a good relationship with the farmer and provides genuine and timely information through extension services. ICICI Bank enters into a tripartite agreement with the IASP and the output buyer. ICICI Bank provides credit to the farmers on the recommendation of the IASP, the farmer pledges his produce to the output buyer at a market- based price, the LASP provides inputs to the farmer. L o a n processing, disbursement and collection are effectively done by the IASP, while the credit decision remains nominally with ICICI Bank. At the end of the season, the farmer supplies the crop to the output buyer and the output buyer deducts the loan amount from the sale proceeds and remits the loan to ICICI Bank in full settlement of the loan amount. The IASP receives a service fee for the loan processing and supervision services (1.5 per cent on recovered loans). T h e model creates a symbiotic relationship between the input supplier, financier and trader. This reduces transactions costs and the risk exposure of all parties and, therefore, presents a relatively low-cost way of serving farmers. It helps improve information collection, reduces credit risk, and increases access to rural financing. However, deepening these relationships to the marginal farmers, s c a h g up the pilots, and replicating them, remain major challenges. Other variations on this model include ICICI? traderfamerjnancing model (Rallis - HLL)in Haryana's Basmati growing area, where Rallis, as an IASP, provides comprehensive field support with fortnightly checks and ensures pest control; and ICICI?farmer finanakg coupled with insurance model, being piloted inTamil Nadu's cotton growing area (Appach), offers tailor-made insurance packages and bulk storage capacity to farmers in order to avoid contamination. Whenever possible, the lender would &e to avoid paying the farmer directly. A model being used by ICICI Bank is to pay the input supplier directly and pre-arrange with the trader to prepay the lender before paying the borrower. T h e borrower contracts to sell his crops at a market-based price - since ifa contract pricewas usedand the market pricewas higher, the farmer wouldnot deliver and default (and sell at the higher rate). The KisanCrcdit Card (KCC) A recent approach to providingcredit to the agriculture sector, including small farmers, is the Iin supporting hundreds of small NGOs all over the eastern region i s useful in this regardand lessons from such experience need to be taken into account. Attention to the demand-side. W e the importance of microfinance in consumption- smoothening should not be underestimated, its success in buildmgup poor peoples' assets, over the medium term, would depend very m u c h on efforts h e c t e d at providmg assistance in sluUs development, technology and marketing- all of which are critical to ensuring that investments made by poor households reap returns and contribute to a sustained increase in incomes and improvements inrural livelihoods. 51 References Anderson, Jock and Verma, Niraj (2003). "Informal Finance in Rural India". Background paper preparedfor the WorldBank.WorldBank,WashmgtonDC. APEC, 2002 `WcrobankingDevelopment, Regulation and Supervision', Chapter 2 of 2002 APEC Economic Outlook, Singapore: Asia Pacific Economic Cooperation Secretariat Basu, Priya and Srivastava, Pradeep (2004). "Scaling-up Access to Finance for India's Rural Poor". Paper prepared for the `Scaling Up Poverty Reduction'conference in Shanghai, China. World Bank, WashingtonDC. CGAP, (2000) Focus Note No 18: Exploring Client Preferences inMicrofinance: Some observations from SafeSave. WashingtonDC: WorldBank. Deshpande, Ramesh and Verma, Niraj (2003). "Review of Rural Finance Institutions in India". Backgroundpaperprepared for the WorldBank.WorldBank,WashingtonDC. Fernando, Nimal and Meyer, Richard (2002). "ASA - the FordMotor Modelof Microfinance," ADB Finance for the Poor 3 (2), M a d a : Asian Development Bank. Fisher, Thomas and Sriram, M. S. (2002) Beyond Micro-Credit: Putting Development Back into Micro-Finance. New Delhi:Vistar. Government of India, Ministry of Finance, (2001). "Report of the Joint Committee on Revitalisation Support to Co-operative Credtt Structure" (Pad Committee). New D e h Government of Indta. Hashemi, Syed, Sidney Ruth Schuler, and Ann P. Rdey (1996). "Rural Credit Programs and Women's Empowerment inBangladesh," World Development 24 (4): 635 - 653. Hess, Ulrich and Iaapper, Leora F. (2003). "The Use of New Products, Processes and Technology for the Delivery of Rural and Microfinance Loans in India". Backgroundpaper prepared for the WorldBank.WorldBank,WashmgtonDC. Hickson, Robert (1999). Reaching extreme poverty: financial services for the very poor. Nairobi, Kenya: MicroSave-Africa. Mahajan, Vijay and Ramola Gupta, Bharti (2003). "Microfinance in India: Banyan Tree and Bonsai". Backgroundpaper preparedfor the WorldBank.World Bank, Washington DC. Meyer, Richard (2002). "Micro finance, Poverty Alleviation, and Improving Food Security: Implications for Indla," chapter in Rattan Lal, ed., Food Security and Environmental Quality. Boca Raton, Florida:CRC Press. Morduch, Jonathan (1999a). "Between the Market and State: Can Informal Insurance Patch the Safety Net?" WorldBank Research Observer 14 (2), August: 187 - 207. Morduch, Jonathan (1999b). "The Microfinance Promise," Journal of Economic Literature 37 (4), December1569 - 1614. 52 Morduch, Jonathan and Stuart Rutherford (2003). "Microfinance: analytical issues for India", Backgroundpaper prepared for the WorldBank.WorldBank, Washngton DC. Mutesasira, Leonard (1999). "Savings and Needs: An InfiniteVariety", Kampala, Microsave-Africa. Patten, Richard and Jay Rosengard (1991). Progress with Profits: The Development of Rural BanhnginIndonesia. San Francisco, CA: International Center for Economic Growth/HIID. Rajan, Raghuram and Luigi Zingales (2003).Saving Capitalism from the Capitalists: How Open Financial Markets Challenge the Establishmentand Spread Prosperityto Rich andPoor Ahke. Reddy, Y. V. (1999). "Future of RuralBanhg." Prof. G. Ram Reddy Third Endowment Lecture, Hyderabad, I d a , December 4,1999. RBI [Reserve Bank of India] (1954). All-India Credit Survey. Bombay:RBI. RBI [Reserve Bank of India] (2004). "Report of the Advisory Committee on Flow of Credit to Agriculture andRelatedActivities from the BanhgSystem". pyas Committee). Mumbai:RBI. RBI [Reserve Bank of India] (2000). "Report of the Task Force to study the Cooperative Credit System and Suggest Measures for its Strengthening". (Capoor Committee). Mumbai:RBI. Robinson, Marguerite (2001). The Microfinance Revolution. Washington, DC: The WorldBank. Rutherford,Stuart (1996). ASA, the Biographyof an NGO.Dhaka, ASA. Rutherford,Stuart (2000). The Poor andtheir Money. Delhi: OxfordUniversityPress. Rutherford,Stuart (2002). "Money Talks: Conversations with Poor HouseholdsinBangladesh about Managing Money," University of Manchester Institute for Development Policy and Management, Finance andDevelopmentResearch Programme Paper 45. Ruthven, Orlanda (2001). "Money Mosaics: Financial Choice and Strategy in a West Delhi Squatter Settlement," University of Manchester Institute for Development Policy and Management, Finance andDevelopment ResearchProgramme Paper 32. Ruthven, Orlanda and Sushd Icumar (2002). "Fine-grain Finance: Financial Choice and Strategy Among the Poor in Rural North India," University of Manchester Institute for DevelopmentPolicy andManagement, Finance andDevelopmentResearchProgrammePaper 57. Srivastava, Pradeep and Shukla, Rajesh (2004). "Rural Financial Access Survey : Summary Findmgs". Backgroundpaper preparedfor the WorldBank.WorldBank,Washngton DC. Srivastava, Pradeep (2004): "Rural Credit and Finance in Uttar Pradesh: A Case study of Three Villages". Backgroundpaper preparedfor the WorldBank.World Bank, Washgton DC. Seibel, Hans Dieter (2001). "SHG Banking: A Financial Technology for Reaching Marginal Areas and the Very Poor." Cologne, University of Cologne. Sinha, Sanjay and Meenal Patole (2002). "Microfinance and the Poverty of Financial Services: How the Poor in India could be Better Served," University of Manchester Institute for Development Policy andManagement, Finance andDevelopment Research Programme Paper 56. 53 Thomas, Susan (2004). "Agriculture Commodty Markets in Inda: Policy Issues for Growth". Backgroundpaper preparedfor the WorldBank.WorldBank, WashingtonDC. Todd, Helen (1996). Women at the Center: Grameen Bank BorrowersAfter One Decade. Dhaka: University Press Ltd. World Bank (2003). "Micro finance in Inda: Issues, Challenges and Policy Options". Washington DC: WorldBank. WorldBank (2003a) "ImprovingAccess to Finance inBrazil". WashmgtonDC:WorldBank. World Bank (2003b). "India: Development Policy Review: Sustaining Reform, ReducingPoverty". Washington DC. WorldBank. World Bank (2004a). "Migrant Labor Remittances in South Asia: Development Impact and Future Prospects". Washgton DC. World Bank. Wright, Graham (2000). Wcrofinance Systems: Designing Quality Financial Services for the Poor. London:ZedPress; Dhaka: University Press Limited. Yunus, Muhammad (2002). Grameen Bank 11: Designed to Open New Possibilities. Dhaka: Grameen Bank. [Available at www.grameen-info.org/bank/bank2.html.] 54 Annex 2: WorldBank/NCAER RuralFinance Access Survey (2003) Summaty ofSurveyMethodology and SampfingFramework T h e survey covered 6,000 households and micro-enterprises (households that rely on non-farm income for m o r e than 50 per cent of their income) intwo states, AP and UP. Wh_yAp and UP?The choice of AP and UP as the two states in which the survey would be conducted was based on the following considerations: it was thought that AP, a leader in the area of delivering finance to the rural poor through microfinance, would present lessons that could be replicated elsewhere in India. Also, AP's formal rural finance sector has undergone reforms inrecent years, notable among which i s the adoption by AP of a n e w Cooperatives Law, and efforts by commercial banks, regional rural banks and credit cooperatives to provide microfmance through establishing links with SHGs. Indeed, AP accounts for 55 percent of the total volume of credit extended under the SHG-bank h k a g e program, India-wide, and a little over 50 percent of the total number of groups benefiting from t h i s program are located in AP. Moreover, a diversity of microfinance models coexist in the state. UP, it was thought, would provide a sharp contrast to AP, accounting for less than 2 percent of the total number of SHGs benefiting from the SHG-linkage program and the total amount of credit extended under the program. Given that UP i s the most populous and amongthe poorest states in India, considerable attention is now being focused on how to improve rural access to finance in that state. Therefore, whde UP'Sexperience would help us understand better the constraints and challenges to improving rural access to finance, AP's experience would help highlight what can be done to overcome constraints and improve access. T h e sampling framework methodology for the survey used random sampling techniques including stratified random sampling. T h e sampling framework was formulated to select a representative sample of 0 households (across landholding and occupation categories with each vdlage), 0 vdlages (geographical proximity to the nearest town and size of village) and 0 districts (per capita agriculture income and per capita formal credit). T h e detailed framework for each level - district, vdlage and household is presented below. Level Distnct and per capita formal credlt for ranking villages Village 0 Census data for listingo f all d a g e names and 0 Census information was used random selection o f dages Secondary data o n distance from the nearest town and number o f households 0 Village questionnaxe to obtam basic data on the 0 Key informant invillage was village used Household 0 Listingproforma 0 1page sheet administered to all households inthe 60 selected villages/state 0 Detailed survey questionnaire Detailed questionnaire administered to selected households o f each selected village. 56 Level District 0 Two primary variables -Per Capita Agriculture Income 12 districts per state and Per Capita Formal Credit -used to rank all districts (UP has around 70 districts and AP has 23) into six strata for each variable. 0 Stratified random sampling then used to select two districts from each stratum. Village 0 Used 1991 Census information on listingo f villages and Five villages per district; 60 randomly select 10 villages per district. d a g e s per state 0 Classified villages into strata according to distance from nearest town (three categories: 4 0 kms, 10-25 kms, >25 kms) and in terms o f number of households (size). 0 Selected five villages from the origmally randomly selected list o f 10 to ensure a degree o f variation on both &stance and size Household Ineach of the selected villages a quick listingo f all Assuming on average around households was undertaken. 200 households per village, Ineach selected village, the households were stratified the quick listingwas for into nine categories that include: -12,000 households per Two landless categories (one who are state; laborers and the other who are micro- entrepreneurs) 50 households per village are Five landholding categories according to administered the detailed land holding sizes survey questionnaire; T h e landholders' category was further subdivided into two sub-categories (one A total of 3,000 (50x60) those with <50 per cent dependence on households were non-farm income and the other with >50 administered the detailed per cent dependence o n non-farm income) questionnaire ineach state Five households were randomly selected from each stratum except the one with landless micro-entrepreneurs where 10 households were selected. Thus, a total of 50 (8*5+10*1) households were selected for detailed survey from each village. 57 Annex 3: Summary Review of the Performance of Retail RuralFinance Institutions (RFIs) '' Governance and Ownersliip A distinctcharactenkticofIndia?banking Figure A.1: International Comparison of govt. owned bankassets f%\ gstem is the signzjkant extent ofstate ownership. Close to 80 per cent of banlung assets are state-owned Philippines (Figure A.l) and more than 90 per cent of the outstanding rural loans Argentina are accounted for by state owned banks - the state owned assets Brazil comprise the assets of all the 27 public sector banks and the 196 Russia I1RBs64, that are all majority government owned banks. This India ownership structure dmorts 0% 20% 40% 60% 80% 100% bankers' incentives, affects prudent management and leaves them Source: The Regulationand Supervision of Banks Around the vulnerable to the pressures of moral World,BarthJ, GCaprio, RLevine, 2001, WorldBank suasion. And w l d e the rural cooperative banks have only minority government shareholdmg, the state continues to exercise considerable control - the state government through the Regstrar of Cooperative Societies, controls all matters relating to registration, membership, election, financial assistance, loaningpowers, business operations, loan recovery and audit of cooperatives. With the banking sector refDms initiated in 199I,governance issues have started receiving increasing attention in commercial banks, especially with regard to the responsibhties of board of hectors, accountabhty to shareholders, criteria for selection of independent members of the board, size and composition of boards, appointments of CEOs and committees of the board including those for audit, nomination, remuneration, investment. Increasingly public sector banks are raisingcapital from the public, which has helped improve their corporate governance through greater shareholder participation. However, cooperative banks and RRBs are lagging behind due to relative4 slower pace of adoption andlor implementation of refoms that are undemy in the rest ofthe bankhg ystem. Additionally, governance issues incooperative banks are exacerbated by duality of control by state governments on the one hand, and, RBI/NABARD on the other. State control continues to mean that the R R B s and rural cooperatives have to operate within the contours of government diktat and may not b e able to take decisions independently - having to reduce interest rates on loans to farmers on account of moral suasion, reschedule agricultural loans, operate as channels 63 Those under the purview of the BankingRegulationAct, 1949 - these includethe commercial banks, the RRBs and the StCBs andDCCBs.Together these account for almost all the lendingto andbranch presenceinruralareas. The 32,000 odd rural branches of the commercialbanks and the RRBs are supplementedby14,OOO cooperative bank branches and 98,000 grass root levelPACS. 64 RRBs' ownership structure i s 50% central government stake, 15% by the state government and 35% belongs to the sponsor commercialbank. Of the 196 RRBs, only two RRBs have private sector banks as their sponsor bank, but are still majorityowned by the government. 58 for the delivery of various government schemes and debt or interest waiver schemes65 and having their elected boards being superceded,"6 are just some examples of this. These factors hinder proper management of rural banks and reduce the attractiveness of rural lendmg for these banks. Management issues in RFIs On-going banking sector rej%rms have accelerated the pmcesses of qualitative transfarmation of management and deplyment of technologp in commercial banks, but their mral branches as well as most RRBs and mral cooperative banks lag behind. M o r e needs to be done for rural banks in terms of trainingof managers and staff for improving credit appraisal, supervision and monitoring skdls, improving the quality of services provided to rural clients as well as in creating the right incentive structures for rural lending. Weak management systems and HR issues are most serious in cooperative banks, which lack in professionalism, and are faced with imbalances in staffmg and poor quality personnel. C E O s for many StCBs are often government officials (rather than professionals) appointed by local state governments and those for the DCCBs are either govemment or StCB officials. On the technology side, while the mainstream/urban commercial banking system has seen important progress in introducing electronic banking, ATMs, electronic funds transfers, electronic clearing services, etc, rural branches of commercial banks, the R R B s and the cooperatives are way behind in this regard. Financial performance of RFIs I n overall terms thejnancialpe$mance of RRBs and cooperative banks is quite weak; on4 the commercial banks' overalljnancialpodion can be consideredsatisfdctoy. However, commercial banks' rural operations cannot b e analyzed separately as these are subsumed in the overall performance of commercial banks. But as some select indcators on portfolio quality and asset composition below reveal, their performance on some critical indcators may not be m u c h better than that of the R R B s and rural cooperative banks. Capital adequacy Figure A.2: Status of Rural Banks in India Commercial banks peform quite well on capital adequay. Capital adequacy for commercial banks was quite comfortable at the end of March 2003 - all 27 public sector banks, which account for more than 90 per cent of total rural portfolio of all commercial banks, have a very comfortable risk weighted capital adequacy ratio in excess of 10 per Y cent; only two banks (from amongst the x0 smaller private banks) out of the ninety three commercial banks, have a ratio below 8 per cent. Unp.oflbb Wsr-wpitalirud Note : R R B s (196) DCCBs (367) StCBs: (29) Figuresmbrackets represent total numbers of rural banks Source : RBI andNABARD 65 Most significantly the Agriculture and Rural Debt Relief Scheme of 1989-90 but several other subsequent schemes, including, most recently, the kharifinterest waiver in2002-03. 66 The position at end-March 2001, was that the elected boards of 75 of the 367 district cooperative banks had been superceded. 59 However, the capital base of RRBs and cooperative banks is vey weak on account of sustained lossesover time. (Figure A.2 67).M a n y of these institutions continue to operate as banks despite having unacceptably low, and even negative, levels of networth. More than a quarter of RFU3s and DCCBs and a little over a fifth of StCBs had negative networth. An addttional and significant proportion of these banks have a positive but extremely weak capital base. T h e average for all R R B s of the networth to total assets ratio (proxy for Capital to Risk Weighted Asset Ratio (CRAR) which i s not computed at present) i s just 2.5 per cent (despite re-capitalization by the government) and 4.9 per cent for all DCCBs and only StCBs with a 7.2 per cent ratio are, in aggregate, at a relatively comfortable position.Since these institutions68 constitute an important part in the flow of formal sector credit in rural areas, their extreme4 weakjnancialposition, especial4 that ofthe cooperative credit institutions,particular4 at the district level, is therej%re,an issue ofgreatconcern. As inadequately capitaliqedRRBs and cooperativebanks continue to operate and mobiliqe deposits, consequent4 a sigmzJicantpropoln ofdeposits are housed in such weak mral banks. For the RRBs, a little more than 40 per cent of total deposits was housed in weak banks (end-March 2002), whde for the rural cooperative banks, the level was over 60 per cent (end-March 2001). While RRB and cooperative bank deposit liabilities are only a small proportion of the total liabllities of the larger b a n h n g system, given that these institutions account for a large percentage (60 per cent) of the physical branch presence of the banking system in rural areas, perceived dtstress could have m u c h larger impacts on the banking system than otherwise expected. While risk to depositors in the RRl3s is mitigated on account of their ownershp structure, deposit insurance and the financial stake of the sponsor commercial banks that themselves have a strong capital base, for the cooperatives such risk-mitigating factors are absent. Asset quali9 and composition Figure A.3: NPLs inRFIs Cause for concern.. - . Asset qual$ of WIs across the board, including commerezal banks, is weak and indicates that whatever lending that does takeplace is genera4 of poor qua@. Figure A.3 shows that NPL ratios 2aoo3 are hgh for all categories of RFIP and even 1500t though the RRBs have shown improvement over the last five years, present NPL levels k m remain at uncomfortably high levels. *m Disturbinglymuch of the NPLs of R F I s are in the doubtful category of classification. Reasons for poor asset quality are many Source RBI reports andNABARD statlshcs including the directed nature of lending, lack of accountability structures, poor management and human resource quality, weak credtt appraisal systems and lack of ability to enforce collateral. 67 Data for R R B s i s for year ending March 2002; for cooperatives for the year ending March 2001. Since risk weighted CRAR is not available for RRBs and rural cooperative banks (both state level, StCBs, and district level, DCCBs), networth to total assets (NW/TA ratio) has been used as a proxy measurefor measure capital adequacy.The NW/TA threshold for capital adequacy for RRBs has been taken at 5% and for cooperative banks at 7.2% - the thresholds are based on the proportions of risk weighted assets to total assets. Since RRBs' asset compositionis majority held in investment portfolio with an expectedlower riskweight, therefore, the threshold for RRBs is lower than that for cooperatives. 68Cooperatives andRRBs accountfor more than 50% of agriculturalcreditflow. Also, they havegreat branch outreach; the R R B s have around 14,000 branches, while the cooperative banks rural branch penetration i s very deep - there are 14,000 cooperativebankbranches + 98,000 PACS. 69 For commercialbanks, a proxy for rural asset quality has beentaken as the NPL ratio on the priority sector advances of public sector banks; the public sector banks account for more than 90% of the outstanding advances in rural areas of commercialbanks. 60 I n tems ofasset composition, the balance sheetsofthe cooperatives indicate clear emphasis on rztral lending; however, commercial banks and RRBs have been stead$ moving awgfrom rural advances and toward investments. FigureA.4: RRBs assetcomposition trends - For the W s , poor portfolio quality 1QQ.QX and lack of incentives to undertake P b . 0 ~ lending, have contributed to increased `Ob' risk aversion and led to a declining 4 ~ 0 7 Credit Deposit (CD) ratio hovering at - ~ - ~ ~ **` ).I a low 40 per cent and f a h g over time zQoy I Ilh...-tlf~-n+rrnt.l,.c o o y with asset composition skewing 1984- 144Q- 1991- 9442-1 9 0 - 1444- 1441- 1446- 1497- 1441- 1949- 2QQQ- ZQQI- towards an investment driven portfolio 40 91 92 93 9 4 45 96 47 98 99 QQ 01 Q2 (Figure A.4). Commercial banks, especially Source: NABARD inrecent years, have faced a much sharper declme inrural-CD ratiorelative to overall CD ratio -h s can b e partly attributed to the poorer portfolio performance on rural loans. At present the rural CD ratio stands at 42 per cent declining from around 60 per cent in the mid-1980s and compared to an overall level of 54 per cent. Liability composition All RFIs in India have a h&h dependence on deposits Figure A.5: Deposits of RFIs costs and - as a funding sowce. Figure A.5 . For R R B s and simificance public sector commercial banks, deposits stand IH at more than 80 per cent of total assets and 90 H m per cent of total debt, and even for the - Ic cooperative banks, the deposits to total assets Ib c ratio exceeds 60 per cent for the state level - H banks and 65 per cent for the district level cooDerative banks. T h i s i s indicative of a stable source of fundmg for banks. As the figure also Wh.FYQ1 RRBsNOP BtCRFlYOl oma2701 indicates, the cost of deposits, relative to that of I ~OSlb!rablSMk d -ChSlQ4 R R B s and commercial banks, i s m u c h higher for Source: NABARDandRBI the cooperative banks and consequently impacts or moral suasion on interest rate caps on aavances are far worse on these institutions. Earnings quality Figure A.6: RFIs profitability - Overall recent earnings qualio of RRBs and (Return onAssets. ROA'L commercial banks has been sati.$actoty; however, cooperative banks'pe$omance is weak. Figure A.6. Only commercial banks have had a steady performance on earnings and FY2003 has seen further improvements with return on assets at around 1 per cent. However, this i s not necessarily reflective of their rural operations and profits have been significantly driven by trading income that has been a major contributor to the operating profit (35 per cent).. For the R R B s too, whde in recent years they have showed improvedprofitabhty, Source, RRB/Coop.. NABARD(RoAvgAssets), RB1:PSBdata m u c h of this i s not on account of lending (RoTotal assets) 61 operations and does not represent afunahmental improvementin their lending business andjnancialpefomance - as m u c h as 64 per cent of total income came from sources other than advances of which 43 per cent was investment income in FY2002. This makes the earning forecast for R R B s particularly susceptible to market and interest rate risks. In fact, the low CD ratio for R R B s (hovering around 40 per cent for the last five years) indicates that lending has really stagnated. While the cooperative banks have had a lending focused approach with high CD ratios, the poor quality of assets and low financial margins have led to weak performance on earnings. The continued weak earningspe@mance ofcooperative banks raises special concernsgiven their alrea4 weak capital base. It i s evident that urgent attention to their financial health and management needs to b e given. T h e particularly poor performance of DCCBs i s of concern since in the three-tier short-term credit structure, DCCBs play a crucial role by m o b h z i n g deposits and channeling these funds to the grass root level PACS and a range of other cooperatives. This i s particularly true since the analysis above wouldappear far worse if it were undertaken with the application of currently used prudential norms for commercial banks, many of which presently do not apply to R R B s and cooperative banks, particularly for provisioning,income recognition and capital adequacy. Regional and other variations Regionaland inter-bank variationsare signzJZcantand indicate the additional complexitiesthat need to be addressed in improving the pe$omance of RFIs. W e on a sector wise aggregate average basis, commercial banks, R R B s and StCBs made profits and DCCBs as a whole were close to break even, a high proportion of the R R B s and cooperatives were loss making - as many as 15 per cent of R R B s (FY2002), 21 per cent of StCBs and 32 per cent of DCCBs (both FY2001). Variations exist between RFIs across a number of financial parameters across regions/states and these are highlighted for profitabhty and capital adequacy in Figures A.7 and A.8, which clearly indicate that R F I s in the north east of the Figure A.7: RFI capitalbase across regions - Figure A.8: RFI profitability across regions x I . I X c Bm% 5 * 1.u = 5az I x B&L I . I X "5or. 3 I . I . I X *mx 4 . I X 7 -%@4 7 1 . L I X Source NABARDStattsucs, FYOI for coops, FY02 for R R B s Source. NABARD Statlstlcs, F Y O l for coops, FY02 for RRBs Figure A.9: RRBs variations by sponsor banks - country are much worse off than those anywhere 25.0% nr;r T 120.0%a else. M o r e generally the north, south and western regions are better performing than the other regions. 20.0% For the RRBs, variation also exists in terms of 15.0% sponsor banks. Syndicate Bank clearly appears to be 10.0% the most successful amongst the commercial banks having a stake in multiple R R B s , while on the whole 5.0% the 30 R R B s sponsored by State Bank of India are 0.ox under performers (Figure A.9). Source: NABARDRRB Statistics, 2002 62 I n conclusion, while commercial banks' overalljnancial position is and /age, sound, thejnancial viabilioofthe mral operations ofthe RRBs and mral cooperative banks, is inherent4 weak.Adequate attention to revamping urgently the operations and financial position of these institutions i s critical, including the rural operations of commercial banks that are a hkely under-performing segment for these banks. Given the large, physical branch presence of R R B s and cooperative banks in rural India, their poor performance on capital adequacy, profitabdity and asset quality indcate that there are serious issues across critical financial parameters - systematic and drastic change in the way R F I s are operating needs to b e made urgently if these institutions are to play an effective role in the future in the provision of rural finance services. 63 A m e x 4: Apex RuralFinance Institutions Three apex institutions are of main relevance to rural finance. These are the RBI, NABARD and SIDBI. A short summary of the roles of each of these institutions i s provided below. lu31 Being the central bank of the country, RBI's primary responsibhty has been the management of the country's monetary and payment systems. However, starting with the era of economic (central) planning in early 1950s, RBI increased its focus on promotional and/or developmental activities in the financial sector, using financial institutions as catalysts for supporting plan objectives and targets. T h e overall policy environment included admmstered resource allocation at the macro level, regulated interest rates, h e c t e d credit and presence of RBI and Go1officials on boards of financial institutions to assure policy compliance and oversight.70 With the gradual implementation of the financial (banking) sector reforms since the mid-l990s, RBI is, by and large, changing toward h t i n g i t s functions to traditional central bankmg, allowing retail financial institutions to increasingly operate in a market-oriented, competitive environment backed by a regime of increasingly deregulated interest rates, supplemented by R B I ' s enhanced prudential supervision. However, in some areas, especially in rural credit, the progress has been slow and RBI, in conjunction with GoI, continues to play a dominant role, directly and/or inhectly, in policy- making, credit planning, priority sector lendtng, and regulating supply of funds to NABARD. RBIretains the responsibhty for the overall nationalrural credit policy and for issuinghectives on rural credit. T h e central bank, thus remains active in the sphere of (a) formulation of rural cre&t policy (as part of the overall monetary and credit policy); @) priority sector lendtng, and (c) interaction with Go1and NABARD on issues concerning rural credit. As part of its central banking functions, RBI i s also responsible for prudential supervision of commercial banks including their rural credit 0perations.7~To fachtate this role, RBI created a department called Rural Planning and Credit Department (RPCD) with a number of staff at i t s central office in Mumbai and regional offices, generally one in each state. RBI staff at the regional level participate in annual district/state level rural credit planning exercises done by the local government agencies and banks, along with NABARD's district and state level staff. However, over time the h e c t role of the RBI in rural finance has been declining in relative terms particularly after the creation of NABARD in 1982. RBI has delegated to NABARD the responsibhty to supervise rural cooperative banks and R R B s . R B I ' s residuary functions include, mainly (a) sanctioning of a general h e of credit to NABARD (though in relative terms the importance of this for NABARD has been d e c h g over time) and @) making annual allocations from profits to the two statutory funds now maintained by NABARD. RBI has three of its own h e c t o r s (members of the RBI board) on NABARD's board, which perhaps provides a n effective link between the RBI (as the central bank) and NABARD (the second tier apex bank) for required deliberations at the policy level. Clearly as the regulator of the rural finance sector, the central bank continues to play a critical role. Strengthening prudential regulations and ensuring improved supervision particularly of the weak rural ~ 70 Established in 1935 as a shareholders' bank, RBIwas nationalizedin 1948.The Reserve Bank of India Act, 1934, which was amended a number of times, providesfor provisions that either enable Go1to give directions to RBI or requireRBI to consultwith Go1on specific matters.Go1plays an importantrole, on its own and throughRBI, inmatters concerningrural creditpoliciesand institutions. 7' RBI Act, 1934andReport of the AgriculturalCredit Review Committee, 1992. 64 cooperative banks and R R B s , facilitating the creation of an enabling environment for rural finance including improvingthe contract enforcement mechanism and fachtating credit information on rural finance, improving policies related to interest rates and directed credit, regulations for MFIs, etc are all areas where the RBI needs to play a leading role going forward. NABARD NABARD was established in 1982 by an Act of Parliament as a public sector institution. Its share capital i s contributed by RBI and GoI, currently at 73 per cent and 27 per cent of the total ( R s 2,000 crores or US$435 d o n equivalent) respectively. NABARD has a nominated Board of Directors comprising a Chairman, a Managing Director (both appointed by Go1in consultation with RBI) and representatives from the RBI, GoI, state governments and other directors nominated by GoI. NABARD has three main functions: (a) institutional development; (b) credit provision; and (c) supervision. T h e institutional development functions include development and implementation of rural credit policy with a focus on integrated rural development and related training, research and consultancy/ - . advocacy. T h e credit function primarily covers refinancing of co-operatives, R R B s -andthe commercial banks credit includes short, medium and long term loans for agriculture; refinancing for SHG-bank linkage, production and marketing credit, etc. NABARD can also make direct loans but currently this window i s h t e d to loans provided to state governments for rural development and rural infrastructure creation which i s financed through commercial banks' deposits with NABARD to the Source : NABAFDdata extent of their shortfall in meeung tne manaatea priority sector lending target of 18 per cent of net bank credit for agriculture (see Table A.l for amounts reflected against the Rural Infrastructure Development Fund, RIDE).Additionally, NABARD provides technical assistance and advice on rural credit policy and institutional issues to GoI, state governments, local govemments, micro- finance institutions and Self-Help Groups (SHGs). As for supervision, by powers delegated by RBI, NABARD does on-site and off-site survedlance of short term cooperative credit banks and RRBs. Besides this, NABARD undertakes supervision of the long term cooperative banks on a voluntary basis in i t s capacity as a major subscriber to their debenture portfolio. In 1999, NABARD established an independent Board of Supervision (BoS) for cooperative banks and R R B s to provide guidance and direction on matters relating to supervision. 65 This is consistent with a s d a r initiative taken by RBIinsetting up a Board of FinancialSupervision (BFS) in 1994 following the recommendation of the Narasimham Committee I. Inpractice, under its delegated authority, NABARD sends all supervision reports to the RBI as t h s continues to be the regulatory authority and regulatory follow-up actions based on these reports are taken with R B I ' s concurrence. Clearly given NABARD's mandate, i t s role in rural finance i s critical and a lot of the credit for the development of this sector, the increase in the scale of the SHG-bank h k a g e program, etc, can b e attributed to NABARD. Nonetheless, going forward in addition to the challenges to sustain the performance of the well performing initiatives (I