Microfinance
The Microcredit Alternative?
Economic and Political WeeklyMarch 31, 20071172production.2Nevertheless, the advocates of microcredit do con-sider it necessary for microcredit institutions to get borrowersto make the transition from consumption loans to production loans(or loans for income-bearing projects) [Rangarajan 1997: 71].The characteristic features of microcredit operations then, aresmall loans to poor households in rural and urban areas forincomegeneration through self-employment. Microcreditinstitutionsmay also provide facilities for savings and otherfinancial services.The following are some of the recurring empirical featuresofmicrocredit: first, microcredit involves loans withoutcollateral.3In the absence of specific policy intervention, landlessand asset-poor households are deemed to be not creditworthyby formal sector lending institutions since they cannot providecollateral that is deemed to be appropriate.Second, NGO-controlled microcredit loans are generallyadvancedto individuals who are members of groups. The group(or SHG) is, in fact, viewed as standing in the place of collateral[Hashemi and Morshed 1997: 217]. The presence of a group hasbeen called a form of “social collateral” [Johnson and Rogaly1997]. The formation of groups, it is argued, has the doubleadvantage of lowering transactions costs and improving repay-ment. The NABARD task force, for instance, identifies threeways of banking with the poor: by means of conventional banklending, linking SHGs with bank lending, and banks lending tomicrocredit and microfinance institutions for on-lending to groupsor individuals. The task force goes on to say that the second andthird methods “are characterised by low transactions costs andhigh repayments” [NABARD 2000].Third, microcredit is viewed as a way of promoting market-led growth or in the words of Mohammed Yunus, of “privatisingthe economy” [Yunus 1997]. This objective was stated in anotherway by World Bank president James Wolfensohn in his speechat the microcredit summit, “Microcredit programmes have broughtthe vibrancy of the market economy to the poorest villages andpeoples of the world” [Microcredit Summit 1997].Fourth, the main target group of microcredit projects constitutea fraction of those in need of credit: this target group is generallyonly at below a line of absolute poverty as determined by nationalestimates.Fifth, while all definitions concur on microcredit as the pro-vision of “small loans”, the scale of “smallness” of loans variesand has to be identified empirically. Loans from the GrameenBank had an upper limit of 5,000 taka or around $100 [Hossain1993]. In a sample survey conducted in 1985, however, Hossainfound that the loans averaged Tk 3,040 (Tk 3,279 for men andTk 2,843 for women) (ibid). The scale is similar in other de-veloping countries; the average loan size was $ 88 in Mexicoand $157 in Pakistan [Johnson and Rogaly 1997: 88-89]. TheNABARD task force estimated the credit requirement per familyas Rs 6,000 in rural areas and Rs 9,000 in urban areas butrecommended that the average loan given to members of SHGsbe around Rs 1,000 [NABARD 2000]. The microcredit cell ofthe RBI, however, has proposed a ceiling of Rs 25,000 formicrofinance, and suggests that the ceiling may be raised, sayto Rs 40,000, for borrowers with a track record of regularrepayment over two to three years [RBI 1999].Finally, while these are the general characteristics of microcredit,a great deal of discussion of the “microcredit alternative” hasbeen on institutional mechanisms for the delivery of microcredit.A very important component of the argument in favour of alarge-scale microcredit effort is that commercial banks cannotand should not be directly responsible for disbursing microcreditloans (because of high transactions costs and poor recovery).Themicrocredit cell of the RBI, for instance, clearly states that“NGOs have widespread appeal as microfinance deliveryvehicles”[RBI 1999]. In the plan of action of the microcreditsummit, the responsibility for achieving the goals of the summitwas placedclearly on “the thousands of existing microcreditNGOs, cooperatives, credit unions, grassroots groups, andpovertybanks that at present comprise the microcredit move-ment” [Microcredit Summit 1997]. Thus, the term “microcredit”commonly used means microcredit mainly by the private sector,including NGOs,where the private sector not only controlsdisbursement but also determines the terms and conditionsattachedto eachloan.4To summarise, microcredit is usually associated with: (a) verysmall loans, (b) no collateral, (c) the formation of borrowergroups, (d) borrowers from among the rural and urban poor,(e)loans for income generation through market-based self-em-ployment, and (f) privatisation, generally through the mechanismof NGO control over disbursement and determination of the termsand conditions attached to each loan.NGO-Controlled Microcredit: EvaluationIn official statements, the move to hand over banking functionsin rural areas to NGOs is motivated by weaknesses in the bankingsystem itself, most notably the “twin problems of non-viabilityand poor recovery performance” of existing rural credit insti-tutions [Rangarajan 1996: 68]. The failure of financial institutionsto deal with income-poor borrowers in an imaginative andsustainable way and the inaccessibility of these institutions tothe poor are stated to be major disadvantages of the existingsystem. Microcredit institutions are seen as being able to rectifythese weaknesses; according to the governor of the RBI, “themain advantage to the banks of their links with the SHGs andNGOs is the externalisation of a part of the work items of thecredit cycle, viz, assessment of credit needs, appraisal, disbursal,supervision and repayment, reduction in the formal paper workinvolved and a consequent reduction in the transaction costs”(ibid: 70).Thus, microcredit is the favoured alternative to the presentsystem because first, it is assumed that the transaction costs ofbanks and other financial institutions can be lowered significantlyif these costs are passed on to NGOs or SHGs, and second, becauseNGOs are expected to perform better than formal-sector creditinstitutions in respect of recovery of loans. Do NGO-controlledmicrocredit institutions incur lower transactions costs than for-mal-sector financial institutions? Is their record with respect tothe repayment of loans superior to that of formal-sector financialinstitutions? We examine the evidence below.Transactions CostsTransactions costs include the costs of information collection,screening of borrowers and projects (by means of project evalu-ation), monitoring and supervision, coordination and finally, theenforcement of contracts and collection of dues.To begin with, it should come as no surprise – despitesuggestions to the contrary – that the administrative costs ofNGOs (and such costs are, of course, the major component of
Economic and Political WeeklyMarch 31, 20071173total transactions costs) are relatively higher than those ofcommercial banks. NGOs cannot match the economies of scale ofa comprehensive system of banking (in the case of India, perhapsthe best network of rural banks in the less developed world).For the period 1988-92, the costs of administration of theGrameen Bank constituted 12.3 per cent of the bank’s totalportfolio, and the costs of administration of the Bangladesh ruraladvancement committee constituted 40 per cent of its totalportfolio[Hulme and Mosley 1996, cited in Chavan andRamakumar 2002]. An important finding of Hulme and Mosleyis that, in a cross-country study of rural credit institutions, thelowest costs of administration, 8.1 per cent of the total portfolio,were incurred by regional rural banks in India (ibid).Second, the costs of administration of NGO-controlledmicrocredit have actually risen when NGO activity is scaled up.As the Grameen Bank expanded its activities, administrativecostsrose from 8.6 per cent of liabilities in 1988 to 18.1 per centof liabilities in 1992 [Hossain 1988, cited in Chavan andRamakumar 2002].Third, repayment rates in NGO-controlled microcredit projectsare directly related to the level of administrative costs andmobilisation efforts [Rahman 1999 and Bhat and Tang 1998, citedin Chavan and Ramakumar 2002]. Organisations such as theGrameen Bank need large numbers of employees for regularmonitoring and assessment, to conduct weekly visits and meet-ings and to collect dues. Mahabub Hossain notes that, in the caseof the Grameen Bank, “the paperwork and the staff time forservicing a given amount of loan are higher than that for a normalrural credit programme”, and that “the benefits of this intensivecredit programme…need to be evaluated against the high costsof operation” [Hossain 1993: 119-20].How do NGO-controlled microcredit projects finance theirhigh-cost operations? The evidence on this seems clear. Theydo so by turning to donors for funds or by raising interest ratesto levels higher than those offered by the banking system or bydoing both. In his review of the performance of the GrameenBank in 1984-86, Mahabub Hossain found that although theannual reports of the bank reported a small profit, his scrutinyof the account books showed that “the credit operations of thebank involve losses that are compensated for by profits fromdeposits in other banks of a substantial amount of low-cost fundsavailable from international donors” (ibid: 120). It is widelyacknowledged that interest rates charged by microcreditorganisations are higher than the corresponding rates charged bycommercial banks or other financial institutions. Real interestrates in 1992 varied from 15 per cent per annum in Bangladeshfor Grameen Bank to 45 per cent in Bolivia for loans advancedby Banco Sol, and 60 per cent in Indonesia for loans advancedby Badan Kredit Kecamatan [Hulme and Mosley 1998, cited inChavan and Ramakumar 2002].Examining interest rates on microcredit projects in India, PallaviChavan and R Ramakumar (2005) find that the costs of microcreditare high. They point out that NABARD has three variants ofdelivery of microcredit through the bank-SHG route. In twovariants, banks lend directly to SHGs who on-lend to membersand in the third variant, NGOs are intermediaries between banksand SHGs. In terms of interest charges, as Chavan and Ramakumarpoint out, the final interest rate includes a margin charged byeach “particular link in the credit chain”. For example, NABARDprovides refinance to commercial banks at 7.5 per cent perannum,banks on-lend to NGOs at 10-15 per cent, NGOs thenlend to SHGs at 12-24 per cent and the groups lend to individualmembers at 24 to 36 per cent (ibid).While annual interest rates in the range of 24 to 36 per centare common, it is not unknown for microcredit SHGs to chargeeven 50 or 60 per cent per annum [Harper 1998, cited in Chavanand Ramakumar 2005]. In fact, the literature notes that in theera of financial liberalisation, NGOs are “free to charge whateverinterest rates they wish in order to cover the (at present veryconsiderable) costs of institution building, supervision, experi-mentation and insurance” [Mosley 1999: 377]. Since proponentsof microcredit do view it as providing credit for productiveactivities, it follows that micro-enterprises funded by high costmicrocredit will have to generate a very high rate of return (say,24 to 36 per cent annually) to be profitable and sustainable forthe borrower.To conclude, the transfer of the task of serving the credit needsof rural borrowers from the banking system to NGO-controlledmicrocredit projects does not reduce transactions costs but ineffect, transfers transactions costs – higher transactions costs –to donors as well as borrowers.Repayments and Overdue LoansA record of near 100 per cent repayment is a major successof NGO-controlled microcredit. Repayment rates are reportedto be over 95 per cent in many microcredit programmes [Hossain1988, Hulme and Mosley 1998, cited in Chavan and Ramakumar2002, Johnson and Rogaly 1997]. This achievement, however,is not costless. A system based on quick repayment of very smallloans does not allow for funds to go into income-bearing activitiesthat have a gestation period of any significance. Only projectswith very quick and high rates of return relative to the tinyinvestment can meet existing repayment schedules. The firstpayment on a microcredit loan is generally to be made in a veryshort time after the loan is given. It has been argued that thiscan put the poorest out of the pail of microcredit, since the abilityto pay the first few instalments depends on the initial resourcebase of the borrower.5The repayment record of NGO-controlled microcredit projectsslackens as the size of the loan increases and as the frequencyof borrowing rises. To take the example of the Grameen Bankonce again, the default rate was 0.4 per cent among first-timeborrowers, 1.2 per cent among second-time borrowers, 6.6 per centamong third-time borrowers and 9.5 per cent among fourth-timeborrowers [Hossain 1988, cited in Chavan and Ramakumar 2002].Further, when the pressure to repay is as overbearing as it oftenis, borrowers have had to borrow from moneylenders in orderto repay NGO-advanced loans [Rahman 1999, cited in Chavanand Ramakumar 2002]. High repayment is dependent on hightransactions costs. As already mentioned, NGOs invest heavilyin supervising, monitoring and enforcing loan repayments. Whenthe activities of NGO-controlled microcredit projects are scaledup, the relative burden of administrative costs tends to increase.One of the criticisms of rural banking, both commercial andcooperative, has been the problem of overdues. While there isno large-scale study of overdues in independent microcreditagencies, two recent analyses published by the RBI [Ghosh 2001,RBI 2001] permit some observations on the issue of overduesin microcredit-oriented organisations.In 1974, the Self-Employed Women’s Association (SEWA),an organisation of working women that has long been involved
EPW
