35823 FocusNote NO. 34 FEBRUARY 2006 GRADUATING THE POOREST INTO MICROFINANCE: LINKING SAFETY NETS AND FINANCIAL SERVICES Does microfinance reach the poorest? Microfinance--or formal financial services for the poor--helps people fight poverty on their own terms, in a sustainable way. Poor people use loans, deposits, and other The authors of this Focus Note financial services to reduce their vulnerability, seize opportunities, and increase their are Syed Hashemi, CGAP senior microfinance specialist, earnings. Indirectly, microfinance improves schooling, health, and women's empow- and Richard Rosenberg, erment.1 In most settings, however, microfinance does not reach the people at the CGAP senior advisor. The very bottom of the socioeconomic scale--the "poorest." authors are indebted to Today there is much debate about whether microfinance is for the poorest. Many Elizabeth Littlefield for the initial idea for this paper. They millions of people living on less than a dollar a day (the very poor) are already being are grateful to Alexia Latortue, served by microfinance institutions (MFIs). And yet few MFIs reach the "poorest" Jeanette Thomas, and Samer customers at the bottom of the poverty scale in their own countries.2 Even in the case Badawi for major contributions of MFIs that focus on reaching very poor clients, there are substantial numbers of to the paper. The authors are also grateful to Brigit Helms people who are too poor to participate. For example, in Bangladesh, where MFIs are and Ousa Sananikone for strongly focused on serving the very poor, MFI concentration is highest among the many helpful comments. second poorest quintile group; it is lowest among the poorest quintile. Microfinance services are not aimed at the poorest communities.3 Why is this? © 2006, Consultative Group to One reason extremely poor people may prefer not to borrow is because they think Assist the Poor debt is more likely to hurt rather than help them. If a woman has no reliable income source, she may feel that obligating herself to make a regular weekly or monthly pay- ment will make her more vulnerable rather than less. Although her investment of the The Consultative Group to loan proceeds in a new microbusiness may raise and stabilize her income, this invest- Assist the Poor (CGAP) is a ment is a risky proposition given that a large percentage of microbusiness start ups fail. consortium of 33 development Realizing this, many extremely poor people decide that their life is already risky enough agencies that support microfi- without taking on debt. Arguably, some of these fears may be more about confidence nance. More information is available on the CGAP Web than reality, but the poor are usually the best judges of their own situation. site (www.cgap.org). On the other hand, suppose a very poor woman with no reliable existing cash flow is willing to borrow money. In an MFI that uses a group loan methodology, the 1Littlefield, Elizabeth, Jonathan Murdoch, and Syed Hashemi, "Is Microfinance an Effective Strategy to Reach the Mil- lennium Development Goals?" Focus Note 24, Washington, D.C.: CGAP, 2003. 2In this Focus Note the very poor are defined as those living on less than a dollar a day. The poorest is the bottom sub- set of this group--the destitute who do not have any reliable sources of earnings. 3Zaman, Hassan, ed., "The Economics and Governance of Non Government Organizations (NGOs) in Bangladesh," Consultation Draft, Washington, D.C.: World Bank, 2005. Building financial services for the poor other women in the group may be unwilling to Microfinance is not the only way to help people. increase their own risk by guaranteeing repayment There are other services and institutions, such as of her loan. "safety net programs," that are usually better In addition to self-exclusion and exclusion by suited to the circumstances and needs of the poor- group members, some of the exclusion is driven by est. One approach to helping the poorest gain MFI policy. Because most microcredit is uncollater- access to appropriate financial services may be to alized, MFIs have found that loan default quickly start with safety net programs that will eventually turns into an uncontrollable epidemic unless they help the poorest gain access to financial services. keep it at very low levels. Precisely because new This Focus Note explores a few cases (perhaps the microbusiness start ups are so risky, MFIs have harbingers of an emerging trend?) where the poor- found--with few exceptions--that they cannot est participate in grant-funded safety net programs, keep default within controllable bounds if they lend where they receive nonfinancial support, such as to borrowers who are new to microenterprise and employment, food aid, training, etc., as well as who don't have other income sources to repay the support to graduate from their existing levels of loan if the new business isn't successful. That policy poverty to a level where they can make good use of is entirely rational if the MFI wants to stay in exis- access to appropriate financial services. These tence and serve growing numbers of poor cus- examples raise the questions: Can microfinance tomers in the future, but adhering to this policy help the poorest? If so, how? And can people means that some of the poorest people are excluded "graduate" from being recipients of grants to from financial services. becoming full-fledged microfinance clients? Another factor is that many of the poorest des- perately need nonfinancial support, such as food, Can microfinance be linked to safety net grants, or guaranteed employment, before they are programs? in a position to make good use of loans or deposit services. Over the years, most MFIs have con- Social protection programs promote the economic cluded that they can deliver financial services more and social security of the poor through a range of efficiently and sustainably if they focus exclusively interventions--from safety nets (food aid or guar- on their financial business and either avoid nonfi- anteed employment) for those in immediate des- nancial services like nutrition, health, and training perate need, to social insurance to buttress those at altogether, or at least isolate those services from risk of slipping down to the ranks of the destitute.5 their microlending operation by keeping them in a Safety nets, as part of a social protection strategy, separate department with separate staff. In addi- have been successful in reaching and helping peo- tion to issues of efficiency and focus, clients may ple at the bottom of the economic ladder--people become confused if the same unit is donating who are too poor for conventional microfinance. social support to them with one hand while insist- Unlike microfinance, safety net programs need ing on repayment of the loans that it is giving with to be highly subsidized. Most of the people they the other hand.4 serve cannot bear the costs of the support pro- These choices by clients, borrower groups, and vided. In addition, the skills needed to deliver MFIs are generally sound, and this paper does not 4 A minority of MFIs--e.g., those who use Freedom from Hunger's recommend a reversal of those directions. But still, "credit with education" model--continue to integrate nonfinancial serv- the result is that MFIs are limited in their ability to ices into their microfinance operations. serve the poorest--those who need more help than 5 The World Bank defines social protection as "public interventions to any other group and who also should have priority (i) assist individuals, households and communities better manage risk and (ii) provide support to the critically poor." Holzmann, Robert, and over others for that help. With rare exceptions, Steen Jorgensen, "Social Risk Management: A New Conceptual Frame- even MFIs dedicated to reaching the very poor fall work for Social Protection and Beyond," Social Protection Discussion short of reaching those at the very bottom. Paper # 6, Washington, D.C.: World Bank, February 2000. 2 safety net services are quite different from the are with the safety net program. This win­win sit- skills needed to deliver credit and other financial uation creates little extra cost or risk for either the services. For these and other reasons, MFIs should program or the MFI. Even a weak MFI can use this not try to deepen their outreach by offering safety strategy to pursue its social mission to sign up net activities along with financial services. promising clients without jeopardizing its ability to Without abandoning this general rule, some achieve sustainability. MFIs are finding ways to team up with existing The second model involves a more intense col- safety net programs in hopes of making themselves laboration between an MFI and a safety net pro- at least indirectly useful to the poorest. Some gram. In this model, the MFI establishes a separate safety net and grant programs are deliberately pro- subsidiary or affiliate that works directly with safety viding financial training and information to their net participants. In cooperation with the safety net clients so that their clients can subsequently link program, the MFI subsidiary provides nonfinancial with MFIs. This collaboration is based on the services and, perhaps, some subsidized savings or premise that many of the destitute can save, start credit. Successful graduates gain access to the building assets, and eventually gain the resources MFI's regular programs. The MFI subsidiary will and confidence to engage in sustainable economic need access to soft money to be able to offer its activities, at which point they can make good use services to participants, until participants are able of loans and other financial services offered by to join the mainstream microfinance program. MFIs. In other words, people who benefit from This second model entails high costs and risks safety net programs may "graduate" to become for the MFI, including the risk that handing out full-fledged microfinance clients. "grants" as part of the safety net program could This Focus Note discusses two basic models of undermine the culture of strict repayment disci- linkages between MFIs and safety net programs. pline that is an essential part of the MFI's micro- In the first model, safety net programs themselves credit operation. There needs to be a clear develop basic financial services for their clients to distinction between the safety net and MFI compo- help them better manage their livelihoods. The nents. This is typically accomplished by using separate MFI's engagement with the safety net program is staff working in a separate subsidiary. This direct limited: the MFI simply coordinates with the engagement model would work well only for a safety net program to recruit successful "gradu- mature, exceptionally strong MFI whose core busi- ates" as customers. The advantage for the MFI is ness is operating so solidly and sustainably that it that the safety net program generates information can afford to have its management and staff about participants' behavior that can later help the resources diluted. MFI make better decisions about the likelihood these participants will repay loans. An MFI would Safety Net Programs Providing Training consider a safety net participant with a track and Delivering Clients to MFIs record of showing up for work, saving regularly, or even repaying a loan offered by the safety net pro- CARE/Bangladesh:The Rural Maintenance Program gram to be a less risky borrower when she eventu- The Rural Maintenance Program (RMP) of CARE ally approaches the MFI for regular microcredit. Bangladesh began in 1982 as a public works pro- An MFI with access to such information can make gram that provides employment for destitute rural safer loans to poorer clients than an MFI without women--women who are heads of households or this information. married to disabled men and who have no other The relationship also benefits safety net partici- income source. Women are recruited to the RMP pants, because it gives them a long-term path for- for a fixed four-year period. They receive cash wages ward and motivates good performance while they for maintaining earthen village roads. Women who 3 are selected for the program must be 18­35 years tenance. The program employed 1,600 women. old and physically able to do the job. Every woman Participants were trained in group solidarity, confi- in the program is required to participate in a com- dence building, and basic business skills. They also pulsory savings plan that captures a fifth of her received help selecting appropriate economic activ- earnings.6 The participants are trained in numeracy, ities. A third of their earnings was deducted as human rights, gender equity, and health and nutri- compulsory savings, which they received at the end tion, as well as income-generating skills and of the employment program. Participants also microenterprise management. CARE continues to saved for a voluntary group fund and used it to dis- provide business management advice for a year tribute loans to other women in the group. The after the end of the program cycle. project ended in early 2002. Some elements of the RMP is active in 90 percent of rural districts in project were subsequently incorporated in the Bangladesh. Program crews maintain 84,000 kilo- national safety net program.9 meters of roads. More than 40,000 women partic- CRIMP successfully targeted the poorest women, ipate in the program at any one time, with 10,000 built an effective savings program, and provided use- completing the program each year.7 ful skills training. Program assessments indicate that RMP aims to move its participants beyond need- earnings from the road maintenance work plus the ing continuous external assistance. The strategy is to training helped most of the participants start create new microentrepreneurs with adequate skills income-generating activities (generally in petty trade training and seed capital from the forced savings. and farming). Savings proved extremely useful in Although not all women succeed as microentrepre- meeting emergency consumption needs, especially neurs, RMP has an impressive track record. in a "maize crisis" period when most households Seventy-nine percent of graduates continue to be faced long periods of hunger.10 About half of the self-employed in microenterprise activities three CRIMP participants maintained a culture of savings years after the end of the program cycle. Women in even three years after the end of the program, and the program receive information on local MFIs and said they used their savings for emergencies and to are encouraged to approach the MFIs for working operate their small businesses.11 However credit con- capital and expansion needs after they graduate. A straints remain and long-term economic recovery is CARE Bangladesh "Household Security Survey," not forthcoming. CRIMP started with the assumption that the conducted in early 2005, indicates that 63 percent training and savings participants received would set of RMP graduates remain members of NGOs three them up as microentrepreneurs who would then years after graduation.8 borrow from MFIs for working capital, business expansion, and other financial services. But CRIMP DFID­CARE/Malawi: CRIMP--A Not So Successful Replication 6 The current daily wage is about US$ 0.85; $0.17 is withheld as com- RMP's success should not suggest that graduating pulsory savings. 7 10,000 new women are inducted each year to replace the 10,000 that participants from an employment program to leave at the end of four years. Ahmed, Shaikh S., Delivery Mechanisms microfinance is easy. A replication of the program of Cash Transfer Programs to the Poor in Bangladesh, Washington, D.C.: in Malawi failed to provide long-term sustainable World Bank, May 2005. solutions to its participants. 8 Email from Dr. Phillip Tanner, Program Coordinator, RMP. 9 Potter, Harry, et al., "Malawi Central Region Infrastructure Mainte- The Central Region Infrastructure Maintenance nance Programme: Final Output to Purpose Review," Lilongwe, Malawi: Program (CRIMP), a DFID­CARE program in CARE Malawi, 2002. www.caremalawi.org. Malawi, was started in late 1999 in two districts of 10 Pinder, Caroline, "Economic Pathways for Malawi's Rural House- the Central Region of Malawi. CRIMP was holds," Lilongwe, Malawi: CARE Malawi, 2003, and Potter (2002). 11 Scharff, Xanthe, "Ex-Post Evaluation of CARE International's Cen- designed as a two-year pilot project to provide tral Region Infrastructure Maintenance Program (CRIMP) in Malawi," employment to poor women in rural roads main- December 2005, Mimeo. 4 failed to bring MFIs into partnerships. The districts grants. By the time the cycle of free grain ends, par- were too poor to attract MFI operations. So while ticipants have received training, managed credit, the program created a mechanism for savings, tried some kind of entrepreneurial activity, and accu- encouraged financial discipline, and helped start mulated savings that can be used as investment capi- income-generating activities, it failed to support tal. They have also gained confidence through group transition to MFI services. participation. At this stage, most participants are ready to engage in income-generating activities and Direct Engagement of MFIs in Safety become clients of regular microfinance programs. Net Programs IGVGD's results are impressive. The program has reached 1.6 million destitute women since its incep- BRAC/Bangladesh: The IGVGD Program tion. Nearly two-thirds of these participants have The Income Generation for Vulnerable Groups "graduated" from absolute poverty to become Development (IGVGD) program in Bangladesh is microfinance clients who have not slipped back into a collaboration among the government, the World requiring further relief assistance. Surveys of IGVGD clients show increases in client incomes and Food Program (WFP), and BRAC, a leading MFI. material assets (e.g., homestead plots, land, beds, Program participants are destitute rural women and blankets), as well as decreases in begging. Studies who have little or no income-earning opportunity. of client self-perceptions indicate that IGVGD par- BRAC discovered early on that it is difficult to ticipants feel more confident after being in the pro- include the very poorest in its conventional micro- gram and believe their lives have improved.12 finance operations because they need immediate The IGVGD model is being replicated in grant assistance for basic survival, rather than Bangladesh. The government and WFP collaborated credit. BRAC also knows that government assis- with ten other MFIs to deliver a similar package of tance does little to solve the long-term problems of grain and financial services to about 44,000 women limited and unpredictable access to food among the in the 2003­04 cycle. destitute and that there are not enough govern- ment funds to serve all the destitute over long peri- Alexandria Business Association/Egypt: TSEP ods. IGVGD's goal is to build a bridge that helps Alexandria Business Association (ABA) in Egypt was participants move from a highly subsidized survival established in 1988 as a nonprofit organization. program into a sustainable microcredit program. It runs two microfinance programs--the Small and IGVGD is built on a government safety net pro- Micro Enterprise Project, with an average loan size of gram that provides free grain for 18 months to des- US$500, and the Blossoms of Micro Enterprise titute, female-headed households that are at the Program, which exclusively targets the very poor, espe- highest risk of hunger. A BRAC unit that is com- cially women, with loans ranging from $25 to $125. pletely separate from the regular microfinance In March 2000, ABA launched its Towards Self- operation organizes the women into groups, col- Employment Project (TSEP) for people unwilling or lects savings, and provides skills training, such as unable to become members of the Blossoms pro- vegetable gardening or raising poultry and other gram because they are too poor. Funded by charita- livestock. After the skills training, participants ble gifts the business community makes as part of its receive tiny loans ($50) to use in funding small- religious obligations, TSEP gives grants of $50 to scale income-generating activities. The payments unemployed people. The first installment of $25 is on these loans are so small that they can be financed out of the grain the women receive. BRAC makes 12Hashemi, Syed, "Linking Microfinance and Safety Net Programs to Include the Poorest," Focus Note 21, Washington, D.C.: CGAP, 2001. no effort to recover its finance and administrative Matin, Imran, and Rabeya Yasmin, "Managing Scaling Up Challenges of costs on these loans, so these costs, along with the a Program for the Poorest," in Scaling Up Poverty Reduction, Washing- rest of the services, have to be subsidized with ton, D.C.: CGAP, 2004. 5 given to clients when they demonstrate to program noted that, even in the best cases, a fifth to a third of staff that they are serious about engaging full time the women who complete these programs fall back in an economically viable enterprise. The second into destitution and need further safety net support. $25 installment is given to clients when they suc- The safety net and microfinance partners in cessfully complete three months of business activity these linked programs have their own comparative and make further commitments to business expan- advantage. Mixing safety net and microfinance sion. TSEP is designed so that successful partici- functions can compromise the effectiveness of the pants graduate first to the Blossoms program and linkage model. The skills needed to administer then to the Small and Micro Enterprise project. grants are different from the skills required to TSEP provided 2300 grants until late 2005. deliver sustainable financial services. Viable micro- Seventy percent of these clients have continued credit depends on strict repayment discipline, with their business activities but, as yet, only 5 which could be compromised if the lending institu- percent have joined the Blossoms program. tion is also providing grants. According to TSEP, participants are slow to grad- Safety net or grant programs need to engage in uate to the Blossoms program because using effective targeting first, then they need to ensure credit staff to provide grants sends mixed mes- clients are provided appropriate support. If the sages to clients. As a result, TSEP is developing safety net program includes tiny subsidized "train- new systems that use separate staff and provide ing" loans, it is probably best that these loans be better communication. TSEP staff will work with managed by microcredit specialists, but separately clients to make sure they understand that grants from the regular lending operations of the partici- are meant to be a one-off activity and that future pating MFI. Financial institutions, for their part, business operations and expansions are contingent need to ensure they have a process to recruit on access to reliable credit sources. "graduates" so they eventually join the ranks of their regular client group. Successful programs Lessons tend to do the following: Rely on a separate professional agency to imple- The case studies presented in this Focus Note sug- ment subsidized grant programs or social pro- gest ways in which links can be established tection programs. These types of programs are between existing subsidized safety net programs not part of an MFI's core competency. and microfinance programs. They show how appropriate sequencing of support can produce Do not send mixed messages on grants and good results for the poorest. Starting with grants loans. MFIs must ensure graduates into their programs understand the need to adhere to to meet immediate consumption needs and build strict financial discipline. "micro-assets," these programs then provide skills training, business management training, savings Start financial services with savings, even when participants are just enrolling into social pro- services, and sometimes small credit to prepare tection programs. This builds micro-assets, clients for running microenterprises. Those who provides a cushion against shocks, and initiates successfully move forward in this sequencing are participants into a culture of regular payments. likely to be ready to graduate to become conven- Provide skills training, business advice, and tional microfinance clients. information on financial institutions as part of Although there are no guarantees that everyone the social protection program. who successfully moves through these support pro- Start with simple loans for easy economic grams will graduate to microfinance, properly struc- activities; once enterprises take off, par- tured support programs hold great potential as a ticipants are ready to become clients of pathway to microfinance. However, it should be conventional MFIs. 6 Identify enterprising participants for gradua- Conclusion: Toward Dialog and tion to MFIs services. Experimentation The cases discussed highlight both the potential for success in graduating to microfinance services The poorest people, whom microfinance has diffi- and some of the inherent difficulties in this. Even culty reaching, are precisely the focus of a fundamen- in the best cases (BRAC­IGVGD and CARE­RMP, tally important development initiative--safety nets as for example), a fifth to a third of participants fall part of a social protection strategy. But the microfi- back into destitution and need continuing support nance industry has generally ignored social protection from safety net programs. or deliberately distanced itself from it. Microfinance In the first model, safety net and grant programs practitioners often equate social protection (specifi- provide training and information to prepare par- cally, safety nets) with grants and subsidies that distort markets and hamper efforts to become sustainable. In ticipants for entry into microfinance programs. turn, social protection experts commonly associate MFIs offer financial services to safety net gradu- microfinance with indebtedness that increases poor ates who have performed well. This is a win­win people's vulnerability. model that benefits both the MFI and the safety These beliefs impede creative exploration of some net program, while creating little additional bur- important potential synergies. The microfinance sec- den or risk if the MFI already has systems suitable tor needs to explore new approaches if it is to extend for very poor clients. Even a young and relatively less solid MFI can link to a social protection program the impact of its services to the poorest. Safety net in this way. programs need to help their participants create a The second model puts much heavier demands viable long-range plan--a graduation scenario--that on the MFI, especially where the collaboration includes access to financial services. Practitioners of involves delivery of subsidized loans to destitute microfinance and social protection may be able to social protection participants to finance start-up serve both these needs through carefully structured activities. An MFI should consider such a collab- collaborations. This potential can be exploited if pro- oration only if it is a mature institution whose fessionals in both fields leave preconceptions behind, core microlending business is operationally and start talking to each other more, and engage in financially stable and sustainable. mutually beneficial, collaborative efforts. Sequencing for Graduation to MFIs Guaranteed Skills Employment Training Graduation to Increased Asset Creation Food Aid Savings Conventional Food Security Microenterprise MFIs Experience Small Cash Subsidized Grants Loans Stage 1 Stage 2 Stage 3 7 Focus Note Bibliography No. 34 Ahmed, Shaikh S. Delivery Mechanisms of Cash Transfer Programs to the Poor in Bangladesh. Washing- ton, D.C.: World Bank, May 2005. Hashemi, Syed. "Linking Microfinance and Safety Net Programs to Include the Poorest." Focus Note 21. Washington, D.C.: CGAP, 2001. Holzmann, Robert, and Steen Jorgensen. "Social Risk Management: A New Conceptual Framework for Social Protection and Beyond." Social Protection Discussion Paper # 6. Washington, D.C.: World Bank, February 2000. Littlefield, Elizabeth, Jonathan Murdoch, and Syed Hashami. "Is Microfinance an Effective Strategy to Please feel free to share this Reach the Millennium Development Goals?" Focus Note 24. Washington, D.C.: CGAP, 2003. Focus Note with your colleagues or request extra Matin, Imran, and Rabeya Yasmin. "Managing Scaling Up Challenges of a Program for the Poorest," in Scaling Up Poverty Reduction. Washington, D.C.: CGAP, 2004. copies of this paper or others in this series. Pinder, Caroline. "Economic Pathways for Malawi's Rural Households." Lilongwe, Malawi: CARE Malawi, 2003. CGAP welcomes Potter, Harry, et al. "Malawi Central Region Infrastructure Maintenance Programme: Final Output to Pur- your comments on this paper. pose Review." Lilongwe, Malawi: CARE Malawi, 2002. www.caremalawi.org. Scharff, Xanthe. "Ex-Post Evaluation of CARE International's Central Region Infrastructure Maintenance All CGAP publications are Program (CRIMP) in Malawi." Mimeo. December 2005. available on the CGAP Web site Zaman, Hassan, ed. The Economics and Governance of Non Government Organizations (NGOs) in at www.cgap.org. Bangladesh, Consultation Draft. Washington, D.C.: World Bank, 2005. CGAP 1818 H Street, NW MSN P3 - 300 Washington, DC 20433 USA Tel: 202-473-9594 Fax: 202-522-3744 Email: cgap@worldbank.org 8