|About ADBINews & EventsSpecial ProgramsPartnerships|
|HomePublicationsBrowse ListingMicrofinance and the Millennium Development Goals in Pakistan: Impact Assessment using Propensity Score MatchingLiterature Review and Motivation|
Literature Review and Motivation
A. Review on the Effects of Microfinance on MDGs
The contribution of microfinance toward the achievement of MDGs goes beyond simply financial services for businesses investing in health and education, managing household emergencies, and meeting the wide variety of other cash needs encountered. The following reviews highlight the role of microfinance in the areas of eradicating poverty, promoting children's education, improving health outcomes for women and children, empowering women, and environmental sustainability.
Eradicate Poverty and Hunger (MDG 1)
Target 1: Reduce by half the proportion of people living on less than a dollar a day.
Microfinance services contribute directly to reducing extreme poverty by improving the income of poor people. In a study conducted in Lombok, Indonesia, Panjaitan-Drioadisuryo et al. (1999) find Bank Rakyat Indonesia (BRI) clients' incomes increased by 112%. Moreover, this increase was enough to move 90% of these families above the poverty line. Only 12 out of 121 respondents reported that their income did not increase, because their husbands used the money for other purposes. Simanowitz (2003), with the use of Poverty Assessment Tool (PAT), found out that, in India, three-fourths of the Microfinance Institution (MFI) clients saw significant economic improvements and half the clients got out of poverty. The World Bank found, in 1998, that the poorest 48% of Bangladeshi families with access to microcredit from Grameen Bank rose above the poverty line. In People's Republic of China (PRC), microfinance programs have helped lift 150 million people out of poverty since 1990 (UNHDR, 2005). Moreover, in Ghana, MkNelly and Dunford (1998) found that clients increased their income by $36, compared with $18 for nonclients. Clients of microfinance generally shifted from irregular, low-paid daily jobs to more secured employment in India (Simanowitz, 2003) and Bangladesh (Zaman, 2000). Filipino households increased income, consumption, and capital (Chowdhury, 2004).
Target 2: Reduce by half the proportion of people who suffer from hunger.
Microfinance allows poor people to diversify and increase income sources, the essential path out of hunger. Diversification makes people more resilient to external shocks. The study in Lombok, Indonesia, reported that 93% of microfinance members ate three meals a day, compared with 51% of nonmembers (Panjaitan-Drioadisuryo et al., 1999). A survey by UNICEF in Viet Nam showed that 73 of the nonborrowers faced food shortages of three months or more compared with 12% of borrowing households (UNICEF, 1996). In Bangladesh, a study on MFI clients found that fewer members suffered from severe malnutrition (relative to the control group), and, more importantly, the extent of severe malnutrition declined as the length of membership increased (Chowdhury and Bhuiya, 2001).
Universal Primary Education (MDG 2)
Target 3: Ensure that all boys and girls complete a full course of primary schooling.
Studies show that children of MFI clients are likelier to attend and stay in school longer. Student drop-out rates are also much lower in MFI client households. A study in Bangladesh found that basic competency in reading, writing, and arithmetic among 11- to 14-year-olds in member households increased from 12 to 27% between 1992 and 1995. In nonmember households, only 14% of children could pass the education competency tests in 1995 (Chowdhury and Bhuiya, 2001).
There has been significant improvement in school attendance of children as well. In UNICEF's Viet Nam microfinance program, 97% of borrowers' daughters attended school compared with 73% of nonborrowers' daughters (UNICEF, 1996). Children (ages 6-21) of Zimbabwe repeat borrowers were likelier to stay in school than those of non-clients (Barnes et al., 2001).
Gender Equality and Women's Empowerment (MDG 3)
Target 4: Eliminate gender disparity in primary and secondary education.
A majority of microfinance programs generally target women—often more financially responsible at repaying than men—as clients, providing them with direct control over resources. A survey in Bangladesh showed that credit-program clients' empowerment increased with duration of membership, suggesting strong program influence (Hashemi et al.,1996).
The Women's Empowerment Program of a Nepalese MFI found that 68% of its female members made household decisions like selling property, the children's education, and budgeting—all traditionally male duties (Cheston and Kuhn, 2002). For one MFI, female household-fund managers increased from 33 to 51% (Cheston and Kuhn, 2002). Also, in the Lao PDR, women who ran family-owned economic activities significantly increased household asset value (Sengsourivong, 2006).
Female clients of MFIs in the Philippines, Nepal, Bolivia, and Bangladesh have become elected officials. In Russia, female MFI clients organized a campaign for democracy in the Russian elections. Clients of MFIs in India have organized rallies for better wages, female worker rights, and legal changes. (Littlefield et al, 2003).
Children's Health, Maternal Health, and Diseases (MDGs 4, 5, & 6)
Target 5: Reduce by two-thirds the mortality rate among children under five.
MFI client households appear to have better nutrition, living conditions, and preventive healthcare than comparable nonclient households. UNICEF, in 1995, noted that infant mortality rates in Nepal were lower in areas with a combined credit and basic social services approach than in areas where credit was extended without social services and in those where no credit was provided. Severe malnutrition declined with the increase in length of MFI membership in Bangladesh (Chowdhury and Bhuiya, 2001). Indonesian MFI members ate three meals a day (93%) (Panjaitan-Drioadisuryo et al., 1999).
Many MFIs also provide target clients with useful health information and with healthcare education to improve nutrition and to make them more aware of contagious diseases and preventive care. A growing number of MFIs have forged partnerships with insurance providers to offer health insurance to clients. An impact study showed that clients had better breast-feeding practices, were likelier to give rehydration therapy to children with diarrhea, and had higher rates of diphtheria, tetanus, and poliomyelitis (DPT) immunizations for their children (MkNelly and Dunford, 1999).
Target 6: Reduce by three-quarters the maternal mortality rate.
Target 7: Halt and begin to reverse the spread of HIV/AIDS.
Target 8: Halt and begin to reverse the incidence of malaria and other major diseases.
The awareness of family-planning activities among clients of the MFI appears higher than that of nonclients. A survey in Bangladesh indicated that rates of contraceptive use were significantly higher for Grameen clients (59%) than for nonclients (43%) (Schuler et al., 1994). This is generally due to both greater awareness of contraceptive programs gained by attending group meetings and from increased mobility that allows women to seek out such services. Of Ugandan MFI female clients, 32% tried at least one AIDS prevention technique versus only 16% of nonclients (Barnes et al., 2001).
Environmental Sustainability (MDG 7)
Target 9: Integrate the principles of sustainable development into country policies and programs; reverse loss of environmental resources.
Target 10: Reduce by half the proportion of people without sustainable access to safe drinking water.
Target 11: Achieve a significant improvement in the lives of at least 100 million slim dwellers.
A number of MFIs are currently integrating sustainable development concerns into their credit services. The MFIs' role is especially important since developing countries often lack environmental awareness and management.
There is evidence that increased earnings—stemming from access to financial services— lead to investments for improved housing, water, and sanitation, thus leading to improved health. Nepalese households with latrines were twice as high in areas where credit and basic social services were linked (UNICEF, 1995). Many MFI programs provide loans specifically for tube-wells and toilets. In India, the MFI provides loans to upgrade community infrastructure (including tap water, toilets, drainage, and paved roads). In one notable development, one Thai MFI recognizes organic agriculture certification and contracts with buyers of organic products as loan collateral. Organic agriculture is environmentally friendly and is most often practiced by poor households in marginal areas. For example, 70% of organic tea in the E.U. market is grown by poor households in mountainous areas in Wuyuan County of the PRC. MFI efforts to finance activities such as organic agriculture should add positive notes on environmental sustainability. At the same time these could foster global partnerships between consumers in developed countries and the poor in developing countries.
Develop a Global Partnership for Development (MDG 8)
Target 12: Develop further an open, rule-based, predictable, non-discriminatory trading and financial system.
The providing of financial services by MFIs to the poor is an MDG goal. Since MFIs are the key to development of microenterprise operated by the poor, they allow the poor to produce products for the market. Very often, these products are sold in export markets; thus, it could be said that microfinance enhances global partnership for development.
B. Review on Impact Assessment of Microfinance Institutions
MFIs are mandated to serve the poor. To target the poor, one of the common strategies among MFIs is to limit loan size. Some MFIs also use a poverty checklist to screen prospective borrowers. Since offering loans regardless of the amount requires the same transaction costs, MFIs tend to drift away from serving the poor. The better-off poor tend to self-select themselves into MFIs' programs, making impact assessments challenging.
In one study, Thai borrowers, prior to borrowing, were much wealthier than nonborrowers ( Coleman, 2006). He attributes the difference to either a selection bias or a programplacement bias. A proper impact analysis should control for the initial wealth differences between borrowers and nonborrowers.
Coleman (1999) addresses this problem by collecting data on 445 households in 14 villages. Of these, eight had village banks operating at the start of 1995. The remaining six hadn't started operations, but village banks were set up already, allowing the households to selfselect according to the village banks' procedures. Since the selected households were forced to wait one year before getting their first loans, it was possible to use this group of households as a control group. This “quasi-natural” experiment controlled for self-selection bias. Coleman finds that controlling for selection makes an important difference: The average program impact is not significantly different from zero after controlling for endogenous member selection and program placement.
In studying the impact of the Khushhali Bank on a range of outcome variables, Montgomery (2005) basically followed the methodology of Coleman (1999) with one crucial difference. Since it was not possible to identify a control group in the sense of Coleman's study, Montgomery followed USAID's AIMS project methodology, which compares “old borrowers” to “new borrowers” within the same area. Montgomery (2005) distinguishes three distinct groups: KB borrowers (Treatment Group), soon-would-be borrowers (Control Group 1), and nonborrowers (Control Group 2). Since borrowers self-select into the microfinance program, the Treatment Group and Control Group 1 should have similar characteristics. The only difference between them is that the Treatment Group is already participating, while Control Group 1 is yet to participate. If the two groups share similar characteristics (e.g., initial wealth level), then Control Group 1 can serve as a controlled variable in impact estimations, effectively allowing estimation of the impact of borrowing from Khushhali Bank on numerous outcome variables, such as consumption, income, education, health, and empowerment.
To apply their methodology to the microfinance impact study, Coleman and Montgomery made a crucial assumption: The characteristics of early- and late-entering clients are the same. As Armendariz and Morduch (2005) note, this assumption might be unrealistic. Why did the new borrowers not sign up earlier? Why were the older borrowers first in line? If their timing of entry was due to unobservable attributes like ability, motivation, and entrepreneurship, the comparisons may do little to address selection biases and could, in fact, exacerbate bias.
Karlan (2001) and Alexander-Tedeschi and Karlan (2002) point out additional problems due to MFI program drop-outs and graduates that can be especially severe in cross-section studies, which rely on comparing old borrowers to new ones. Sometimes borrowers graduate MFI programs because they are doing so well that they no longer need assistance. More often, it is the borrowers in trouble that leave. The result of drop-outs is that only successful borrowers remain in the program, resulting in overestimation of program impact. Hulme (1999) reports that dropout rates are 25 to 60% per year in East Africa. Gonzalez-Vega et al. (1997) report that just half of the clients who ever borrowed from BancoSol were still active at the time of analysis. In rural areas, however, the fraction of active borrowers numbered two-thirds of all borrowers, possibly reflecting the fact that there are fewer alternative lending sources in the countryside.
It is likely that these “older borrowers” (i.e., those who remain active) have the positive qualities of survivors, while “new borrowers” have yet to be tested. If borrowers who had difficulties paying the loan are likelier to drop out, comparing old to new borrowers will overestimate impacts. We suspect that the Khushhali Bank survey also suffers from these problems and distinguishing “old” borrowers from “new” ones will not solve them. If this is the case, the findings of Montgomery (2005) likely exaggerate the impact of microfinance.
To cope with program placement and self-selection biases, we need to identify a control group that is identical to the “treatment” population. In this study, we use the Propensity Score Matching (PSM) method, extensively employed in medical studies and sociology, to assess the impact of Khushhali Bank borrowing on a host of outcome variables.
Download this Discussion Paper [ PDF 198.6KB| 31 pages ].
[previous chapter] [next chapter]
There are  comment(s) for this entry. Post a comment.
|Back to Top|
|© 2014 Asian Development Bank Institute.|