ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Programme Design and Impact Assessments: 'Success' of Microfinance in Perspective

The enviable success of microfinance programmes in terms of high repayment rates, vis-à-vis state-led subsidised credit programmes of the past, has mostly been attributed to the group lending aspect with the much celebrated "joint liability" feature. But in practice, such programmes use a host of other dynamic and static incentives features, engage in meticulous designing of financial products and use different models of service delivery, going beyond "group lending with joint liability". This paper attempts to take stock of these developments: the ongoing innovations in designing flexible and tailor-made financial products, the ingenuous use of a variety of incentive mechanisms, and the evolution and acceptance of microfinance from a focused credit intervention approach towards a broader paradigm of livelihood finance.

SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200877Programme Design and Impact Assessments: ‘Success’ of Microfinance in PerspectiveSayantan BeraToday, the emergence and success of microfinance is posed vis-à-vis state-led development credit initiatives of the past.1 Formal banks have always been reluctant to serve the poor, especially the asset poor, who cannot offer any collat-eral. In an effort to reverse this trend, subsidised credit programmes were attempted, notable examples being the integrated rural development programme (IRDP) in India and similar programmes in the Philippines. Banks were given heavy subsidies, not only to keep interest rates low, but to compensate for the high transaction costs and inherent risks of such opera-tions. In case of the Philippines, low interest rates due to subsi-dies created an excess demand for credit and only few well-off farmers received the bulk of cheap credit rather than the target groups [David 1984]. For IRDP, repayment rates fell below 60 per cent in 1989 and came down to just 31 per cent by 2001. One striking finding was that only 11 per cent ofIRDP clients borrowed more than once [Pulley 1989; Meyer 2002]. Since most borrowers took only one loan and there was no guarantee that with repay-ment they will have access to larger loans in future, the incentive structure was absent, and more borrowers chose to default bring-ing down overall repayment rates to dismally low levels. Despite the failure and consequent withdrawal of the programme,2IRDP did register positive impacts on the livelihoods of the poor in terms of non-farm growth, employment and wages [Burgess and Pande 2004; Binswanger and Khandker 1995].The enviable success of microfinance programmes in terms of high repayment rates has mostly been attributed to the group lending aspect with the much-celebrated “joint liability” feature. But in practice, such programmes use a host of other dynamic and static incentives features, engage in meticulous designing of financial products and use different models of service delivery, going beyond just “group lending with joint liability”. This paper attempts to take stock of these features, the ongoing innovations in designing flexible and tailor-made financial products, the ingenuous use of a variety of incentive mechanisms, and the evolution and acceptance of microfinance from a focused credit intervention approach towards a broader paradigm of livelihood finance. Numerous anecdotal evidence and relatively fewer statistical evaluations also point to the success of microfinance in generating “positive” impacts among participants and spillover benefits among non-participants. Advocates argue that micro-finance access, apart from improving household level develop-ment indicators, raises incomes and positively affects growth, rather than only redistributing income within the local economy. The only way to verify these claims is by carrying out rigorous impact assessments. For a better understanding of impact assess-ments, both at the academic and practitioner levels, we go on to The enviable success of microfinance programmes in terms of high repayment rates, vis-à-vis state-led subsidised credit programmes of the past, has mostly been attributed to the group lending aspect with the much celebrated “joint liability” feature. But in practice, such programmes use a host of other dynamic and static incentives features, engage in meticulous designing of financial products and use different models of service delivery, going beyond “group lending with joint liability”. This paper attempts to take stock of these developments: the ongoing innovations in designing flexible and tailor-made financial products, the ingenuous use of a variety of incentive mechanisms, and the evolution and acceptance of microfinance from a focused credit intervention approach towards a broader paradigm of livelihood finance. Sayantan Bera ( is a research scholar at the Centre for Economic Studies and Planning, School of Social Sciences, Jawaharlal Nehru University, New Delhi.
  • 2. Emerging Economies
  • 3. Market Linkages
  • 5. International Capital Flows

  • 7. Poverty and Human Development
  • 8. Inclusive Growth and Microfinance
  • 9. Globalization and Economic Growth
  • SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200879discuss widely used impact evaluation methodologies, the problems in carrying out rigorous and clean statistical investiga-tions and document findings from an array of impact assess-ments. In brief, we try to place the success of microfinance – in terms of high repayment rates and impacts generated – with a wider understanding of practice and the underlying theory, by drawing on diverse experiences from around the world.1 Microfinance Delivery Models Until recent past, the received wisdom was that lending to poor households is doomed to fail: costs are high, as are the risks involved, saving propensities are low, and few households have anything to put up as collateral. The experience of Bangladesh’s Grameen Bank challenged this idea, and now a broad range of financial institutions offer alternative microfinance models with varying philosophies and target groups [Morduch 1999]. The basic precepts of the Grameen Classic System (GCS) are as follows: groups form voluntarily, and while loans are made individually, all in the groups are held responsible for repayment – the celebrated “joint liability” clause. A group consists of five borrowers, with lending first to two, then to the next two, and finally the fifth (sequential financing). This group of five meets together weekly with seven other groups, so that the bank staff meets 40 clients at a time. According to rules, if one member defaults all in the group are denied subsequent loans (contingent renewal). The contracts take advantage of local information, and rely on informal insurance relationships (like bailing out a fellow group member in times of distress) and threats like social isola-tion or sanctions (in case of voluntary non-repayment of loans or strategic default). In contrast to the Grameen, BancoSol of Urban Bolivia lends to groups and to individuals, with a sharp focus on banking rather than social service. Repayment schedules and loan durations are flexible and typical clients are among the richest among poor clustered just above the poverty line. Bank Rakyat Indonesia (BRI), like BancoSol, is financially self-sufficient, and lends to better-off among poor and non-poor households, but does not use the group lending mechanism. Loan officers of BRI get to know clients over time, starting borrowers off with small loans and increasing loan size conditional on repayment performance (progressive lending). Badan Bank Kredit Desa System (BKDs) (village credit boards) in rural Indonesia is a sister organisation of BRI lending to individuals in poorest households without any collateral support. Another approach of microfinance delivery, the village-banking model, was started in the mid-1980s by the Foundation for International Community Assistance (FINCA) in Latin America. The non-governmental organisations (NGOs) help set up village financial institutions in partnerships with local groups and then facilitate a relationship between the village banks and local commercial banks with an aim to create sustainable institutional structures. Incorporating some features of the village banking model and using a modified form Rotating Savings and Credit Associations (ROSCA),3 India has followed an altogether unique route. The largest share of microfinance operation is carried out through the self-help group (SHG) bank linkage programme under the guidance of National Bank for Agriculture and Rural Development (NABARD) with an active participation of commercial banks (predominantly, public sector ones) and regional rural banks. In the most popular model, anNGO starts the process of group-building by initiating a modified form of ROSCA [Aniket 2005]. An averageSHG would have 15 members – usually poor and mostly women – who pool their savings into a fund from which they can borrow. This is done by opening an account in a commer-cial or a regional rural bank. The group continues with its savings for an initial period of six months after which they can borrow from the accumulated internal funds sequentially. As impatient non-borrowers wait for a chance to borrow, they monitor and if need be, audit the current borrowers intensively. Lending sequen-tially thus plays an important role in binding individuals within a group. This is more so as the non-borrowers’ savings are under threat, if not used properly and repaid in time by current borro-wers. If the group is able to manage internal savings and group loans during the initial period of one year, the NGO links the groups to external sources of credit from formal banks and/or government-run programmes with a credit component. The most beneficial aspect of this approach is that (a) the programme builds upon an already existing rural banking network, and (b) it disentangles the role of institution-building (taken up byNGOs) and financial intermediation (by banks). Parallel to the SHG-bank linkage programme,NGO/microfinance institu-tions (MFIs) also provide microfinance services using a variety of approaches: SHG, cooperative and the Grameen methodology. 1.1 Incentive Structures and Innovative DesignsInformation asymmetry between the lender and the borrower is an accepted feature of all credit markets. The formal financial institutions deal with the problem by means of collateral. In contrast, microfinance mechanisms serving the asset and the income poor rely on informal relations like social ties and cultural proximity to avoid information and enforcement problems. Until recently, the theoretical literature on microfinance focused on two features used by Grameen Bank and similar lending institu-tions to explain their excellent repayment record: (a) self-selection of group members that bypasses the adverse selection problem [Ghatak 1999, 2000]; and (b) peer monitoring tackling the moral hazard problem [Stiglitz 1990; Varian 1990; Ghatak and Guinnane 1999]. Both these features, the series of papers establish, arise due to the joint liability clause in group lending schemes. Moreover, the enforcement of credit contracts (or avoid-ing cases of strategic default) was explained by harnessing of social collateral through sanctions imposed on delinquent group members [Banerjee et al 1994; Besley and Coate 1995; Ghatak and Guinnane 1999].However, in practice MFIs employ a host of other dynamic and static incentive mechanisms to ensure repayment and the conse-quent sustainability of operations [Morduch 1999; Armendáriz de Aghion and Morduch 2000, 2005]. Some recent theoretical works have incorporated these design features of microfinance programmes [Jain and Mansuri 2003; Roy Chowdhury 2005, 2006; Tedeschi 2006].4 Prominent among these dynamic incentives are “sequential financing”, “progressive lending”, “contingent renewal” and ”regular repayment schedules”.
    SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly80Another static feature relatively ignored in the literature is “lender monitoring” – referring to the intensive monitoring of borrower activities by lenders – a practice religiously followed by microfinance programmes like Grameen, despite the transaction costs it entails.One useful feature of incentive mechanisms like progressive lending and contingent renewal is that they can be employed beyond group lending schemes. This holds particular importance as microfinance makes its way forward from serving closely-knit rural communities towards a more diverse and non-homogeneous urban populace. Serving an urban clientele would necessitate weakening of group based incentive mechanisms like “joint responsibility” and the effectiveness of “social sanctions”. Instead, individual incentives have to be built into while designing products. This is supported by the successful experience of BancoSol of Urban Bolivia, BRI andBKD in Indonesia and ASA in Bangladesh, practising “individual liability” lending. Roy Chowdhury (2005) explores these much used features of Grameen type lending institutions, namely, sequential financing and lender monitoring in conjunction with joint liability. The focus of the study is to look into the relative importance of these features in group lending schemes vis-à-vis the joint liability feature. The most striking result was that in absence of lender monitoring and/or sequential financing group lending schemes may involve under-monitoring by borrowers, even in presence of joint liability. Moreover, it was shown that lenders, by incurring costly monitoring itself, induces relatively less costly “peer monitoring” among borrowers. Most schemes including the classic Grameen model practice weekly or monthly repayment schedules starting right after the loan is disbursed. Microlenders claim that it inculcates some sort of “fiscal discipline” in the borrowers. One of the serious disad-vantages of the scheme is that borrowers cannot invest in projects with long gestation periods or in highly seasonal occupations like agricultural cultivation. But the rationale for such strict and regular repayment schedules – at least from the lenders point of view – throws up crucial factors. Lenders can get an early warning signal if a borrower fails to repay initial instalments and, more importantly, the lender can select less risky clients. This is because the household must have some other sources of income in order to repay the early instalments, and cannot depend only on the project returns for repayment. In that case a borrower might even choose to borrow from other informal sources like moneylenders to repay her loan [Jain and Mansuri 2003]. Progressive LendingProgressive lending is another widely used incentive mechanism: the promise of larger loan sizes on successful repayment of outstanding loans. But this feature comes with a punishment aspect on the other side of the coin: the threat to exclude default-ing borrowers, and in some cases the entire group from future access to loans (contingent renewal). Recent theoretical work suggests that this, in fact, is a harsh measure in view of the non-systemic (like medical emergency, death, etc, uncorrelated among borrowers) and systemic economic shocks (like natural disasters) that the poor face. Rather than discarding borrowers forever, the design of the contracts can be changed, whereby defaulting borrowers can be made to enter into a “finite punish-ment phase” [Tedeschi 2006]. Further, the results lends support to MFI attempts in improving and expanding upon the existing savings and insurance products as a measure to shelter poor borrowers from negative economic shocks.The devastating floods of Bangladesh in 1998 affected Grameen borrowers and many others alike. Material possessions were lost including houses. Grameen Bank initiated a rehabilitation programme for its borrowers by issuing fresh loans for restarting income-generating activities and to repair or rebuild houses. But borrowers started to feel the burden of accumulated loans in the face of crisis, gradually started to stay away from weekly meetings and repayment rates showed quick declines. As a response, Grameen Bank, in 2001, moved away from the Grameen Classic System (GCS) to Grameen Generalised System (GGS, named GrameenII) doing away with the joint liability feature of group loans and the group fund itself [Yunus 2002]. Loans are now made directly to an individual in a group and other group members are no longer liable for repaying her peers’ dues. As a result the defaulting members are fully responsible for their own failures and other group members carry on operations of repay-ing and borrowing as usual. Moreover, the defaulting member is not penalised and removed from the system but takes a “flexible loan” option, rescheduling loan instalments and repayment period. To add to this, loans are no longer required to be repaid in fixed weekly instalments over a period of a year, but can be tailor-made according to borrower’s needs: instalment sizes can vary with larger sizes in peak periods and smaller sizes during lean seasons, and total repayment period may vary according to a client’s preference. The promise of increasing loan sizes in future, on timely repayment of ongoing loans, is the most prominent incentive mechanism for borrowers under Grameen II. Once a borrower starts defaulting and takes the “flexible loan detour”, her entire credit ceiling built over the years gets erased (the moderate “punishment aspect”, as against the earlier practice of debarring borrowers and/or groups forever from any future loans). On successful repayment of the flexible loan she starts afresh with the entry-level loan ceiling.Other innovative features used by microfinance programmes include a flexible attitude towards collateral, public repayments and targeting of women [Armendáriz de Aghion and Morduch 2005, pp 134-41]. The potential advantage from serving a predominantly female clientele refers to both financial aspects of and social objectives behind microfinance operations. The financial advantage is that women tend to be more conserva-tive in their investment strategies and are easily swayed by peer pressure and interventions by loan officers. This makes themreliablebets as regards repayment. Moreover, women are less mobile and are more likely to work in or near their home, making it easier both for peers and lenders to monitor them. Lower mobility also reduces the incidence of strategic default under the fear of social sanctions. Banks likeBRI take a non-traditional view of collateral. House-hold items can be taken as collateral if they have sufficient personal value for borrowers even when they are worth little in
    SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200881the hands of BRI. This is a break from the traditional banker’s view that collateral should be valuable enough so that banks can sell them to recover problem loans. This flexible approach towards collateral acts as an incentive for poor borrowers without any conventional collateral support to repay. Although microlenders like Grameen andASA in Bangladesh have weakened its insistence on group lending mechanisms, still, customers meet as group and make public repayments. Public repayment schemes have several advantages for the lender, primarily that of using the avoidance of social stigma as an inducement for individual borrowers to repay. Public repayments also serve as a mechanism through which the bank can directly elicit information regarding errant borrowers through cross-reporting and create pressure on them if they are engaging in strategic default.1.2 From Microcredit to Livelihood FinanceIn an impact assessment study, six years after its inception in 1996, Basix India, anNGO involved in microcredit operations, found that only 52 per cent of its three year plus microcredit clients reported an increase in income, 23 per cent reported no change, while another 25 per cent actually reported a decline in income. The reasons found out by the programme operators were unmanaged risk, low productivity in crop cultivation and livestock rearing, and inability to get good prices from the input and output markets. This led Basix to revise its strategy: Basix now offers microcredit along with a range of insurance products covering life, health, crop and livestock. Moreover, to enhance productivity, agricultural and business development services are now being offered to customers. For ensuring better prices, alter-nate market linkages are facilitated in both input and output markets, and producers are encouraged to form cooperatives which are given institutional development services to become more effective [Mahajan 2005].This reflects one of the most significant departures – both in literature and practice – the shift from microcredit (referring to small loans) to microfinance – a broader term encompassing, in addition to lending, efforts in savings mobilisation, insurance provisioning, and in some cases assisting in distributing and marketing clients’ output. The transition reflects a change of outlook resulting from a growing realisation that low-income households can benefit more through a broader set of financial and non-financial services than just credit.An example of assisting the chronically poor through directed and subsidised intervention is given by the Income Generation for Vulnerable Group Development (IGVGD) programme of Bangladesh Rural Advancement Committee (BRAC) in Bangla-desh. BRAC structures this programme around a food aid compo-nent sponsored by the World Food Programme. The programme provides 18 months of food subsidies and half a year of skills training with the aim of developing new livelihoods for the chronically poor. Participants are also made to save regularly in order to build financial discipline and an initial capital base. Once the training programme is over, successful households graduate to BRAC’s regular microfinance programmes. By 2000, the IGVGD had served 1.2 million households with two-thirds of the participants graduating successfully intoBRAC’s regular credit programmes [Hashemi 2001]. Programmes like IGVGD suggest ways in which links can be established between subsidised safety net programmes for the poorest and microfinance programmes. Starting with grants to meet immediate consumption needs and build “micro-assets”, these programmes can provide skills and business management trainings, savings services and small amounts of credit to prepare potential clients for running micro-enterprises [Hashemi and Rosenberg 2006]. 1.3 Flexible and Tailor-made Financial Services The potential benefits of introducing flexible savings options into existing microfinance programmes are large. With savings, not only can households build up assets, but can also smooth seasonal consumption needs, finance major expenditures, self-insure against major shocks and self-finance investments. Hence, some like Robinson (2001) argue that deposit services are more valuable to the poor than credit services.New initiatives are under way to create financial products for microfinance clients with terms and features that will appeal to low-income customers [Armendáriz de Aghion and Morduch 2005, pp 147-74]. SafeSave, a cooperative working in the slums of Dhaka sends its staff members on daily rounds, during which customers are visited in their homes or businesses. The custom-ers can choose to make deposits or pay loans, and there are no limits to how big or small the daily transactions must be. On a larger scale,BRI has built a customer base of over 25 million depositors by reducing minimum opening amounts and required balances, and by creating a network of over 3,900 sub-offices. Thailand’s Bank of Agriculture and Credit Cooperatives (BAAC) have also followed this example.From its inception Grameen Bank created savings accounts for all clients, but they were obligatory in nature. A fixed portion of the loans disbursed had to be deposited in the accounts, and these funds could only be withdrawn upon leaving the programme. Accepting the idea that low-income households deserve more convenient and flexible ways to save Grameen has introduced new flexible saving products in 2001 under GGS. Now each borrower has three accounts – (a) personal savings account, (b) special savings account, and (c) pension deposit account. At the time of loan disbursal 5 per cent is deducted fromtheamount, half of which goes to the personal savings account, while the remaining half is deposited in the special savings account. A borrower is allowed unlimited withdrawal from her personal savings account any time she desires. The mandatory weekly savings go to this personal account. But the special savings account is non-withdrawable for the first three years. The Grameen Pension fund is a long-term savings option which has become hugely popular among borrowers [Yunus 2002].Saving options can either be rigid (with minimum balances, restrictions on number of withdrawals, compulsory deposits, etc) or flexible in nature.BRI has introduced flexible saving options for its borrowers through ‘Simpanam Pedasaan’ (SIMPEDES): a popular scheme in villages despite paying no interest rates on small deposits as compared to BRI’s competing savings product,
    SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly82TABANAS. But TABANAS had the disadvantage of restricting withdrawals to two times per month, while SIMPEDES offered unlimited withdrawals. It was found that although very few TABANAS savers actually withdrew twice a month, the limita-tion is an important psychological barrier to people in rural areas who fear they might not have access to their savings in times of need. In contrast to these developments, NABARD’s flagshipSHG-Bank Linkage Programme in India offers very limited savings choices for group members. Even after 15 years of initiation of the programme, the savings options are limited to mandatory weekly deposits of fixed amount in the group account. Voluntary savings is not encouraged, primarily on grounds of keeping the account-ing procedure simple.Insurance is typically valuable to poor households who are vulnerable to both systemic and non-systemic economic shocks. A growing challenge for microfinance programmes is that of providing insurance on top of credit and savings services [Morduch 2006]. Life insurance has been most successful in this respect. Life insurance or so-called “credit-life” contracts pay off outstanding loans and provide the family with a fixed payout in the event of client’s death. The programme run byFINCA Uganda provides about $ 700 to the dependents of the client, should the client die an accidental death, with all outstanding loans being repaid. In return for this coverage, clients pay an extra 1 per cent over the interest that is charged for loans. Clients have been pleased with the arrangement since it ensures that their own death does not impose an undue burden on their families.5 Grameen Bank also provides this service through the loan insur-ance savings account. Compared to life insurance, health insurance programmes have been far less successful. This is because adverse selection problem is rampant in voluntary programmes. When programmes are voluntary, less healthy households are more eager to avail insurance, and insurers bogged down by imperfect information are unable to set prices. In order to control costs, most programmes have imposed restrictions on the diseases they are willing to cover. The MicroCare Health plan ofFINCA, Uganda covers only a range of outpatient and inpatient services – there is no coverage for chronic AIDS related illnesses. The health insur-ance programme of Self-Employed Women’s Association (SEWA) (Ahmedabad) has controlled costs by limiting coverage and relying on public hospital care. In addition to health and life insurance, attempts have been made to provide property and crop insurance. Rainfall insurance is now being provided byBASIX under pilot programmes in Morocco and India. But for a localised programme operator, these risks are generally correlated with very little scope for diversifi-cation. Natural disasters and crop failures due to draught or flood would affect all borrowers alike, and a microfinance organisa-tion might go bankrupt in absence of adequate reserves. The possible solution is having access to reinsurance policies, whereby some of the risk can be transferred to a large insurance company with a diversified portfolio across geographical boundaries. The strategy is to abandon trying to insure against bad crop yields and instead to insure against bad weather directly.2 Impact Assessment MethodologiesEvidence from microfinance clients around the world indicates that access to financial services help the poor to increase their household incomes, purchase assets and make them less vulner-able to emergencies. Moreover, such access is claimed to trans-late into better nutrition and improved health and literacy status at the household level. Access to finance is also claimed to have made women clients more confident and assertive in the house-hold decision-making process, thus confronting gender inequali-ties. However, the existing evidence – the hypes and hopes behind the success of microfinance – abounds with anecdotes that can never be a substitute for rigorous statistical evidence. In fact, more pertinent are the problems encountered while carry-ing out careful impact assessments [Armendáriz de Aghion and Morduch 2005, pp 199-229]. The question that all careful evaluations seek to answer is how borrowers would have done without the programmes. The normal way to handle this is by selecting a “control group” of non-participants and comparing their status with programme participants or the “treatment group”. But such a method faces several problems. It might well be the case that microfinance clients had initial advantages over non-participants in terms of wealth, income, health and education status and entrepreneurial capabilities. The rigours of frequent repayment structures demand that households must have other sources of income apart from the project they are investing in with the borrowed fund [Jain and Mansuri 2003]. Hence, households which are compara-tively well-off might “self-select” themselves in microfinance programmes. Coleman (2002), for example, finds that households that will later become microfinance clients tend to be already significantly wealthier than their non-participating neighbours. In a small sample from Bangladesh, Hashemi (1997) finds that more than half among the non-participants chose not to partici-pate because they felt they could not generate adequate profits to repay the loans. Similarly, a more able and enthusiastic person can self-select herself into a programme. Due to this self-selection biases, comparing borrowers and non-borrowers would lead to an overestimation in the impact of microfinance on borrower well-being. While observable characteristics like income, assets, age and locational advantages can be controlled for while making comparisons, something like entrepreneurial capability ishardto capture. In those cases, someone’s high business profits might be wrongly attributed to microfinance rather than her entrepreneurial abilities. Moreover, no microfinance programme chooses its area of op-erations independently: they target a specific section of the popu-lation with eligibility requirements in place. Suppose a pro-gramme serves the poorest in accordance with its goals and methodology of operations. Then a simple comparison between borrowers (as treatment group) and non-borrowers (as control group) will have a downward bias in terms of impact since the participants started from a considerably lower level of income and assets than the non-participants. On the other hand, if “pro-gramme placements” favour the relatively well-off, a comparison between participants and non-participants will have an upward bias in the impact. If sufficient measures are not taken to “control”
    SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200883for these selection biases, estimated impacts on income and empowerment would be misleading and microfinance interven-tions will seem to be more “positive” than what is actual. The results of impact evaluation studies are mixed. There is an evidence of modest positive impacts on income, expenditure and other household indicators, but some studies [Coleman 2002] find that positive impacts disappear once selection biases are addressed.Another widely used methodology of evaluating impacts is to compare “old borrowers” to “new borrowers” in a given pro-gramme. But this approach requires a strong assumption that all borrowers share similar characteristics at the time of joining the programme. Then only the remaining difference in welfare levels of mature clients (say, those in the programme for more than three years) and younger clients (those who joined recently) can be attributed to participating and staying in the microfinance programme for a period of time. This approach again faces severe problems [Karlan 2001]. Why is it that the new borrowers did not choose to participate earlier and why were the older borrowers first in line? If the timing of mature borrowers’ entry was due to unobservable attributes like ability, motivation and entrepre-neurship then such comparisons can do little to address the selec-tion biases. Another problem in this kind of cross sectional analy-sis comes with respect to programme dropouts. Borrowers may leave because they are so well-off now that they no longer need microfinance, but more often it is the borrowers in trouble who leave. It is likely that older borrowers have the positive qualities of survivors, while new borrowers are yet to be tested. If “failures” are more likely to dropout then comparing old borrowers to new borrowers will overestimate impacts. Hence, this methodology of impact evaluation requires a careful study of programme dropouts to find out about the direction of bias.2.1 DocumentingImpactsMost impact assessments studies are undertaken at the institu-tional level. In 1995 the United States Agency for International Development (USAID) launched the Assessing the Impact of Microenterprise Services(AIMS) project which designed five tools (two quantitative and three qualitative) to provide practitioners with a low cost and fast way of measuring impacts [AIMS 2000]. This project used cross sectional data to compare old borrowers and new borrowers. But the drawbacks of such evaluations made AIMS impact assessments to use longitudinal data (with base line and end line surveys) and non-client control groups. Another se-ries of studies follow from the qualitative assessments of Imp-Act programme at the Institute of Development Studies at Sussex. These studies focus primarily on tracking performance indica-tors for borrowers only and are more concerned with improve-ments within specific programmes. This set of studies, despite presenting us with helpful data, should be distinguished from impact assessments which also track control groups. Studies are also commissioned byMFIs with support from respective donors to assess programme impact.6 The most reliable, accepted and rigorous statistical evaluations in the literature come from the World Bank, also known asThe World Bank-BIDS Studies [Pitt and Khandker 1998; Khandker 2005]. We also present an array of impact related findings from major microfinance programmes around the world representing all types of studies described above.7 But the interpretations and conclusions following from such studies have to be read with the necessary caveat that the problems in estimating impacts are manifold and “there is not yet a widely acclaimed study that robustly shows strong impacts... (although) many suggest the possibility” [Armendáriz de Aghion and Morduch 2005, p 4].Khandker (2005) did an in-depth study of two Bangladesh MFIs: BRAC and Grameen Bank, and another government pro-grammeRD-12, drawn on research done in 1991-92 and again in 1998-99.8 This study is the most widely cited evaluation of a mi-crofinance programme and is the first serious attempt to generate accurate assessments by dealing with selection bias and non-ran-dom programme placement. The paper assesses the long-term poverty impact of microfinance to answer the crucial policy ques-tion whether accrued benefits at the borrower level are due to sus-tained income impact or simple income redistribution. The study has found that moderate poverty declined by 18.2 per cent in pro-gramme villages and 12.6 per cent in non-programme ones. Poverty declined by more than 20 per cent for participants who were inthe programme since 1991-92 – a decline of 3 percentage points per year. More than half of this reduction was directly attributed to microfi-nance. The impact was found to be greater on moderate poverty than on extreme poverty and spillover effects among non-partici-pants were attributed to growing economic activity. Based on his data, Khandker concludes that microfinance accounted for one-third to one half of the overall reduction of poverty in Bangladesh. Coleman’s (2002) study of microfinance programmes in north-east Thailand has used a novel research design to address bias due to “endogenous programme selection” and “programme place-ment”. This unique approach uses information on the borrowers even before the programme arrives. Data was collected on 445 households from 14 villages of which eight had village banks operating at the start of 1995. The remaining six villages had no programmes running, but village banks were to be set up a year later. At the beginning of 1995 staff from the village bank programmes organised households in these six villages into banks, allowing villagers to self-select according to standard procedures of eligibility requirements. But then these households were forced to wait for one year before getting their first loans. The results following from the analysis indicated that even prior to programme intervention, participants tended to be signifi-cantly wealthier than non-participants and the wealthiest villag-ers were twice as likely to participate in the programme as the poorer villagers. The results from the study demonstrated that microfinance loans positively affected many measures of house-hold welfare for the wealthy committee members, but the impact was largely insignificant for poorer “rank and file” members. Policy recommendations from the study were increased vigilance in targeting the poor, and introduction and enforcement of eligi-bility criteria based on wealth while continuing to allow villagers to self-select. The striking finding was that the averageprogramme impact was not significantly different from zero after controlling for endogenous member selection and programme placement.USAID in 1995 launched theAIMS project designed to provide practitioners a low-cost way of measuring impact and improve
    SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly84institutional performance. The first of these assessments in Honduras [Edgecomb and Garber 1998] and Mali [MkNelly and Lippod 1998] demonstrated the possibility for practitioners to use theAIMS methodology to assess their performance. But the small sample sizes of these surveys, cross sectional nature of the data and the corresponding analysis procedure of comparing new and old borrowers raised severe criticism [Karlan 2001]. The AIMS Core Impact Assessments avoid this problem through the use of longitudinal data, non-client comparison groups and larger sample sizes. TheMFIs chosen for the core studies wereSEWA (India), Zambuko Trust (Zimbabwe) and Mibanco (Peru), all serving a predominantly urban clientele [Snodgrass and Sebstad 2002]. The study found that on average borrowers had net income gains only in India and Peru; in Zimbabwe there were no measur-able increases in average incomes for borrowers, relative to those in control groups. Increased diversification of income was reported in Zimbabwe and Peru, but no impact on diversification was found in India, either from borrowing or saving. Some positive impacts were found on food expenditure in Zimbabwe and Peru with only slight impact for repeat borrowers in India. One severe drawback was the skewed nature of the programmes being selected for the study. Since all three programmes covered by the study primarily serve an urban clientele, the AIMS studies may be more representative of a “typical” microfinance programme. To add to this, neither Zambuko nor Mibanco focus on very poor clients. Another unfortunate feature is that while clients were surveyed in 1997 and again in 1999, the round one data was not baseline since clients were already with the programmes for some time. This diminishes the ability to test for selection bias [Goldberg 2005]. In contrast to the rigorous quantitative techniques employed by the assessment methodologies discussed above, the Imp-Act programme favours the “social performance management” approach emphasising the importance of designing services according to clients’ needs. The Imp-Act programme has built up a body of research focused on improving the performance of their memberMFIs. This set of studies is less concerned with validating the impact of microfinance programmes to the development community, and few of them use comparison groups to gauge “impact”. Simanowitz and Walter (2002) undertake two case studies,SHARE (Andhra Pradesh, India) andCRECER (Bolivia), to look at the achievements of programmes in terms of financial and social objectives.9 BothSHARE andCRECER, working with the poorest clients by international, national and local standards, were found to achieve significant positive impacts on the lives of majority of their clients. Seventy six per cent of SHARE’s mature clients experienced significant reductions in poverty, and one-third were no longer poor. Sixty six per cent of CRECER clients were found to have experienced income increases. The perform-ance of two programmes was further credited since both were financially self-sufficient.Several studies [Khandker 2005; Skoufias 2001; Kabeer and Noponen 2005] suggest that lending to women yields greater social and economic impacts in terms of poverty reduction, school enrolment, increases in food expenditure, and improved child and adult health status. In Bangladesh, positive impacts have also been found on contraceptive use and reduction in domestic violence attributable to women’s participation in micro-finance programmes [Hashemi et al 1996; Kabeer 2001]. However, several evaluations of the empowerment potential of micro-finance programmes for rural women have arrived at conflicting conclusions: some studies claim extremely positive results while others suggest moderate to no impacts [Kabeer 2001].3 ConcludingRemarksMicrofinance started as an ingenious method of financial service delivery using an incentive based approach that exploits social ties in closely-knit and culturally homogeneous societies. However, the experience of Grameen Bank, despite the pheno-menal success, pointed out the tensions that the joint liability aspect might give rise to while lending in a group [Woolcock 1999]. Use of social sanctions against defaulting members, even when such defaults are due to negative economic shocks, leads to animosity among members and subsequent corrosion in social ties. Moreover, as microfinance began to serve a diverse urban populace, its reliance on group incentives weakened. Individual based incentives were attempted and successfully structured into ongoing programmes. However, group-based mechanisms might also promote aspirations when viewed as a coordination device to save (as exemplified by the popularity of ROSCAs), rather than just imposing peer pressure to maintain a degree of financial discipline. On a more practical basis, programmes with social and empowerment objectives find it easier and more fruitful to reach people in groups than approaching them individually. The success of microfinance experience – the flair for continu-ous innovation in service design and delivery – owes much to the dedicated practitioners working close to grassroot realities. But going beyond the “positive” impacts of programmes on income levels and a wide range of development indicators at the house-hold level, it is sensible to recognise that microfinance cannot be expected to work everywhere and for everyone. Sub- Saharan Africa, for example, is ravaged by AIDS, malaria and malnutrition. Extending microfinance services in such areas without complementary investments in human and physical capital is unlikely to reverse the conditions of acute poverty and deprivation. Notes 1 See Chavan and Ramakumar (2002) for a compar-ison of state-led poverty alleviation schemes with NGO-led microcredit initiatives on several fronts like targeting, impact, repayment rates, adminis-trative costs and subsidy dependence. 2 IRDP has been revamped into Swarn Jayanti Gram Swarojgar Yojana (SGSY), a back-ended subsidised credit scheme for self-employment purposes, launched in 1999. Credit is disbursed through microfinance SHGs with active partici-pation of local bodies, district rural development agencies (DRDAs), banks and NGOs [Sriram 2005]. 3 ROSCAs, in its pure form, are self-selected, volun-tarily formed group of individuals who agree to save and contribute a pre-committed sum of money every period towards the creation of a fund [Besley et al 1993]. Generally, purchase of indivisibles by the credit constrained is the reason behind ROSCA formation, but in some cases formation of ROSCA is motivated more as methods of savings rather than as means to borrow [Rutherford 1997].
    SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200885 4 The recent but growing literature on “experimen-tal games”, carried out in laboratory set-ups and sometimes with potential customers of micro-finance, is making a modest attempt in disentan-gling the relative importance of commonly used design features. These experimental games also point to the relative importance of features other than the “joint liability” to explain high repay-ment rates [Abbink et al 2002; Karlan 2005; Cassar et al 2005; Giné et al 2005; Giné and Karlan 2006].5 The actual insurance coverage for the FINCA programme is provided by AIG, one of the world’s largest insurers. AIG gets 45 per cent of the premium collected, and the rest is kept by FINCA to settle administrative costs and supplement revenues. In addition, loans are reinsured by AIG, implying that debts are paid off to FINCA by AIG in event of a client’s death. 6 Notable studies are Husain (1998) for BRAC, Bangla-desh; Todd (2001) for SHARE, Andhra Pradesh (India); and Noponen (2003) for ASA, India. 7 For an extensive survey of impact assessments till date see Goldberg (2005). 8 Pitt and Khandker’s (1998) earlier paper based on the cross sectional data from the baseline survey of 1991-92 was severely criticised by Morduch (1998) in relation to the selection of control groups and the econometric model employed. Khandker (2005) acknowledged the criticisms, and armed with panel data from the 1998-99 end-line survey, employed a simpler model relying on fewer assumptions [Goldberg 2005].9 The study was originally commissioned by the Microcredit Summit Campaign to “refute the myth” that programmes targeting the very poor or low-end clientele cannot be financially sustain-able. See Microcredit Summit Campaign Report [Daley-Harris 2005].ReferencesAbbink, K, B Irlenbusch and E Renner (2002): ‘Group Size and Social Ties in Microfinance Institutions’, Working Paper, University of Nottingham and University Erfurt.AIMS (2000): ‘Learning from Clients: Assessment Tools for Microfinance Practitioners’, Draft Manual, Assesing the Impact of Microenterprise Services, Washington DC.Aniket, Kumar (2005): ‘Self-Help Groups: Use of Modified ROSCAs in Microfinance’,Preliminary Draft.Armendáriz de Aghion, Beatriz and Jonathan Morduch (2000): ‘Microfinance beyond Group Lending’,The Economics of Transition, Vol 8, No 2, pp 401-20. – (2005): The Economics of Microfinance, The MIT Press, MA, US.Banerjee, Abhijit V, Timothy Besley and Timothy W Guinnane (1994): ‘Thy Neighbour’s Keeper: The Design of a Credit Cooperative with Theory and a Test’,The Quarterly Journal of Economics, Vol 109, No 2, pp 491-515.Besley, Timothy and Stephen Coate (1995): ‘Group Lending, Repayment Incentives and Social Colla-teral’,Journal of Development Economics, Vol 46, pp 1-18.Besley, Timothy, Stephen Coate and Glenn Loury (1993): ‘The Economics of Rotating Savings and Credit Associations’,American Economic Review, Vol 83(4), pp 792-810.Binswanger, Hans and Shaidur Khandker (1995): ‘The Impact of Formal Finance on the Rural Economy in India’,Journal of Development Studies, 32(2), December, pp 234-62.Burgess, Robin and Rohini Pande (2004): ‘Do Rural Banks Matter? Evidence from the Indian Social Banking Experiment’, C E P R, Discussion Papers.Cassar, A, L Crowley and B Wydick (2005): ‘The Effect of Social Capital on Group Loan Repayment: Evidence from Field Experiments’, Working Paper for Conference on Microfinance and Economic Journal Symposium on Joint Liability Lending.Chavan, Pallavi and R Ramakumar (2002): ‘Micro-credit and Rural Poverty: An Analysis of Empiri-cal Evidence’,Economic & Political Weekly, March9, pp 955-65.Coleman, Brett (2002): ‘Microfinance in Northeast Thailand: Who Benefits and How Much?’, ERD Working Paper Series No 9, Asian Development Bank, April.Daley-Harris, Sam (2005): State of the Microcredit Summit Campaign Report 2005, Microcredit Summit Campaign, available at, Cristina (1984): ‘Credit and Price Policies in Philippine Agriculture’ in Dale W Adams, Douglas H Graham and J D von Pischke (eds),Undermining Rural Development with Cheap Credit, Boulder, Westview Press.Edgecomb, Elanie and Garber Carter (1998): ‘Practitioner-led Impact Assessment: A Test in Honduras’, AIMS, Washington DC.Ghatak, M (1999): ‘Group Lending, Local Information and Peer Monitoring’,Journal of Development Economics, Vol 60, pp 27-50. – (2000): ‘Screening by the Company You Keep: Joint Liability Lending and the Peer Selection Effect’,The Economic Journal, Vol 110, No 465, pp 601-31.Ghatak, M and Timothy W Guinnane (1999): ‘The Economics of Lending with Joint Liability: Theory and Practice’, Journal of Development Economics, Vol 60, pp 195-228.Giné, X, P Jakelia, D Karlan and J Morduch (2005): ‘Microfinance Games’, Working Paper.Giné, Xavier and Dean S Karlan (2006): ‘Group versus Individual Liability: A Field Experiment in the Philippines’, World Bank Policy Research Working Paper, 4008, September.Goldberg, Nathaneal (2005): ‘Measuring the Impact of Microfinance: Taking Stock of What We Know’, Publication Series, Grameen Foundation, US.Hashemi, Syed (1997): ‘Those Left Behind: A Note on Targeting the Hardcore Poor’ in Geoffrey Wood and Iffath Sharif (eds),Who Needs Credit: Poverty and Finance in Bangladesh, Dhaka University Press, Dhaka.– (2001): ‘Linking Microfinance and Safety Net Programmes to Include the Poorest: The Case Study of IGVGD in Bangladesh’,CGAP Focus Note, No 21, May.Hashemi, Syed M, Sidney Ruth Schuler and Ann P Riley (1996): ‘Rural Credit Programmes and Women’s Empowerment in Bangladesh’,World Development, 24 (4), pp 635-53.Hashemi, Syed and Richard Rosenberg (2006): ‘Grad-uating the Poorest into Microfinance: Linking Safety Nets and Financial Services’,CGAP Focus Note, No 34, February.Husain, A M Muazzam (ed) (1998): Poverty Alleviation and Empowerment: The Second Impact Assessment Study of BRAC’s Rural Development Programme’, BRAC, Dhaka, Bangladesh.Jain, Sanjay and Ghazala Mansuri (2003): ‘A Little at a Time: The Use of Regularly Scheduled Repay-ments in Microfinance Programmes’, Journal of Development Economics, Vol 72, pp 253-79.Kabeer, Naila (2001): ‘Conflicts Over Credit: Revalua-ting the Empowerment Potential of Loans to Women in Rural Bangladesh’, World Development, 29(1), pp 63-84.Kabeer, Naila and Helzi Noponen (2005): ‘Social and Economic Impacts of PRADAN’s Self-Help Group Microfinance and Livelihoods Promotion Programme: Analysis from Jharkhand’, India, Imp-Act Working Paper No 11.Karlan, D S (2005): ‘Using Experimental Economics to Measure Social Capital’,American Economic Review, 95(5), pp 1688-99.Karlan, Dean (2001): ‘Microfinance Impact Assessments: The Perils of Using New Members as a Control Group’,Journal of Microfinance, 3(2), pp 76-85.Khandker, Shahidur (2005): ‘Microfinance and Rural Poverty: Evidence Using Panel Data from Bangladesh’,World Bank Economic Review, Vol 19, pp 263-86.Mahajan, Vijay (2005): ‘From Microcredit to Livelihood Finance’,Economic & Political Weekly, October 8.Meyer, Richard (2002): ‘Microfinance, Poverty Allevi-ation and Improving Food Security: Implications for India’ in Rattan Lal (ed),Food Security and Environ-mental Quality, CRC Press, Boca Raton, FL.MkNelly, Barbara and Karen Lippod (1998): ‘Practi-tioner Led Impact Assessment: A Test in Mali’, AIMS, Washington DC.Morduch, Jonathan (1998): ‘Does Microfinance Really Help the Poor? New Evidence on Flagship Programmes in Bangladesh’, Draft, McArthur Foundation Project on Inequality, Working Paper, Princeton University, Princeton.– (1999): ‘The Microfinance Promise’,Journal of Economic Literature, Vol 37, No 4, pp 1569-1614. – (2006): ‘Microinsurance: The Next Revolution?’ in Abhijit Vinayak Banerjee, Roland Bénabou and Dilip Mookherjee (eds),Understanding Poverty, pp 337-55, Oxford University Press.Noponen, Helzi (2003): ‘ASA Gram Vidiyal, Micro-finance Impact Report’,ASA, Tiruchirapally.Pitt, Mark and Shahidur Khandker (1998): ‘The Impact of Group-based Credit Programmes on Poor Households in Bangladesh: Does the Gender of Participants Matter?’,Journal of Political Economy 106(5), pp 958-96.Pulley, Robert (1989): ‘Making Poor Creditworthy: A Case Study of the Integrated Rural Development Programme in India’, World Bank Discussion Paper58, The World Bank, Washington DC.Robinson, Marguerite (2001):The Microfinance Revolution: Sustainable Banking for the Poor, The World Bank, Washington DC.Roy Chowdhury, Prabal (2005): ‘Group Lending: Sequential Financing, Lender Monitoring and Joint Liability’,Journal of Development Economics, Vol 77, pp 415-39. – (2006): ‘Group Lending with Sequential Financ-ing, Contingent Renewal and Social Capital’, Journal of Development Economics.Rutherfrod, Stuart (1997): ‘Informal Financial Services in Dhaka’s Slums’ in Jeoffrey Wood and Iffath Sharif (eds), Who Needs Credit? Poverty and Finance in Bangladesh, Dhaka University Press,Dhaka.Simanowitz, Anton and Alice Walter (2002): ‘Ensur-ing Impacts’, Imp-Act, Institute of Development Studies, UK.Skoufias, Emmanuel (2001): ‘Is PROGRESA Working? Summary of the Results of an Evaluation by International Food Policy Research Institute (IFPRI)’, Food Consumption and Nutrition Divi-sion, Discussion Paper No 118, Washington DC.Snodgrass, Donald R and Jennefer Sebstad (2002): ‘Clients in Context: The Impacts of Microfinance in Three Countries’, Synthesis Report, AIMS, January.Sriram, M S (2005): ‘Microfinance and the State: Exploring Areas and Structures of Collabo-ration’, Economic & Political Weekly, April 23, pp 1699-1703.Stiglitz, Joseph E (1990): ‘Peer Monitoring and Credit Markets’, World Bank Economic Review, No 4, pp 351-66.Tedeschi, G A (2006): ‘Here Today, Gone Tomorrow: Can Dynamic Incentives Make Microfinance More Flexible?’, Journal of Development Economics, Vol 80, Issue 1, pp 84-105.Todd, Helen (2001): ‘Paths Out of Poverty: The Impact of SHARE Microfinance Limited in Andhra Pradesh, India’, Imp-Act Report.Varian, H (1990): ‘Monitoring Agents with Other Agents’, Journal of Institutional and Theoretical Economics, Vol 146, pp 153-74.Woolcock, Michael J V (1999): ‘Learning from Failures in Microfinance: What Unsuccessful Cases Tell Us About How Group Based Programmes Work’, American Journal of Economics and Sociology, January.Yunus, Muhammad (2002): ‘Grameen Bank II: Designed to Open New Possibilities’, available at

    Dear reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here


    (-) Hide

    EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

    Back to Top