Peeling Back the Layers of Group Liability in Bolivia

Policy Issue:
Policy Issue:

Since its inception in the 1970s, the group liability model has been hailed as the innovation that made lending to the poor possible.  Under the group liability model, borrowers take out loans with groups of other borrowers and all group members are  jointly liable for loans extended to others within the same group.  Many microfinance institutions continue to work under this model, yet some practitioners and scholars have begun to question the assumption that the benefits of group liability outweigh the costs. Many contend that group liability not only fails to increase repayment rates but that the policy also locks out those who could potentially benefit from credit access but are unwilling to take responsibility for other people’s loans. It is still unknown whether group liability or self-selection produce better outcomes for microfinance clients and lenders.

Context:

Despite abundant natural resources, Bolivia remains one of the poorest countries in South America. El Alto, a fast growing suburb of La Paz located in the highlands above the capital, is known for its politically active inhabitants whose protests over the use of natural resources have led to clashes with authorities. Cochabamba, on the other hand, is well known for its agriculture production of grains, coffee, cacao, and coca leaf. Pro Mujer, a prominent microfinance lender in Bolivia, provides loans to female clients in these areas. A majority of these clients engage in commercial activity, very often in local markets or selling goods produced in home-based businesses. Pro Mujer calls their borrowing groups “Communal Associations”, and a typical group has between 18 and 25 members.

Description of Intervention:

Researchers randomly selected 300 of the sample 400 Communal Associations to test the impacts of various forms of group liability. In these 300 groups, the Communal Associations were divided into sub-groups called “solidarity groups”.  Each solidarity group included 5-7 individual borrowers.  Rather than being responsible for the repayment of all loans in the Association, borrowers in the treatment Associations were only liable for the loans of the members within their solidarity group, effectively reducing the number of loans for which each woman was liable. 

Additionally, treatment groups were randomly allocated to be formed in different ways. In the first sub-treatment group, researchers shifted liability from the Association level to the solidarity group level without changing the self-selected composition of the solidarity groups. In the second sub-treatment, the solidarity groups were rearranged to be composed of borrowers with similarly-sized loans. In the final sub-treatment group, solidarity groups were rearranged and members were assigned to solidarity groups at random.

Researchers collected data on institutional outcomes important to Pro Mujer, such as repayment rate, loan size, dropout rate, and new member acquisition rate. In addition to analyzing the impact of group liability on borrowers’ behavior, researchers also collected data on how credit officers used their time in order to determine if the new policy had an effect on the efficiency of bank operations.  Loan officers are required to spend much of their time traveling to Communal Association repayment meetings.  If changes to the liability structure improve overall repayment, loan officers may be able to lead meetings and oversee repayment in a more efficient manner, allowing them to take on more borrowing groups.  On the other hand, if the new liability structure worsens repayment, loan officers may end up spending more of their time enforcing repayment, and have less time for other required activities. 

Results and Policy Lessons:

Due to low compliance, data analysis has been delayed.

Gharad Bryan, Xavier Giné, Dean Karlan

Clean Water in Northern Ghana

This study assesses the willingness of households in Northern Ghana to purchase a ceramic water filter. The Kosim filter is sold by Pure Home Water (PHW), a Ghana-based NGO, and has been demonstrated to be highly effective at improving water quality without needing electricity. We will also measure the health effects of household-level water treatment in areas with high waterborne disease loads.  

Policy Issue: 

Diarrheal diseases, which result poor water quality, are a leading cause of death in the developing world, killing approximately 1.8 million people per year. Achieving the Millennium Development Goals of reducing the proportion of people without sustainable access to safe drinking water is especially difficult for the rural poor. Delivering treated water through pipes has resulted in sustained health gains in developed countries and urban areas in developing countries, but is not considered feasible in rural areas with dispersed populations and weak institutions for maintenance. Community interventions, such as spring improvement or communal wells, have not produced strong results, and policy makers are increasingly interested in household and point of use treatments. However, the effectiveness of such treatments in rural environments, the role of education and marketing to encourage use, and how to expand access with limited resources remain largely unknown.

Context of the Evaluation: 

Diarrheal diseases account for 12% of childhood deaths in Ghana, and are the third largest cause of death for children under the age of 5. These diseases are caused by the ingestion of water contaminated by fecal matter, and 20% of Ghana’s population does not use an improved water source. The sparsely populated northern region of Ghana is one of the least developed parts of the country, and has even less access to clean water than the national statistics would suggest. The majority of its residents make their living in agriculture, living far away from one another. This low population density makes any state- or community-wide water treatment intervention costly and impractical.

Details of the Intervention: 

This study will evaluate the demand, use, and impacts of one household level water treatment technology. The Kosim filter is a ceramic filter marketed and sold by Pure Home Water, a Ghana-based NGO. This simple product has been proven to be highly effective at improving water quality and is appropriate for the region, since it removes particles and pathogens from water without the use of chemicals or electricity which require some form of delivery.

Researchers are measuring the willingness to pay of households for the Kosim filter by offering a random selection of households the opportunity to purchase the filter through door-to-door sales. Households will also be offered a randomized price for the filter, to determine price effects and willingness to pay for preventive health technologies.

Researchers will collect data from 1,500 households on water quality, education, income, consumption, health, diarrhea disease knowledge and water treatment and storage practices to see how these variables affect the willingness to pay for a filter. The randomized offer price provides a means to estimate the filter’s health impact and health spillovers among neighbors, without letting a set price screen out households who have a lower value for clean water. Thus, researchers can evaluate different techniques for creating behavioral change, such as the adoption of new water treatment technologies and storage techniques, and the propensity of individuals to drink treated water and provide treated water to their children.

Results and Policy Lessons: 

Forthcoming

[1] As of 2004. WHO, “The Top Ten Leading Causes of Death” (accessed Nov. 6, 2009)

[2] WHO, “Mortality Country Fact Sheet 2006, Ghana.” (accessed Nov. 6 2009)

[3] United Nations, Human Development Report 2009 "Ghana" (accessed Nov. 6, 2009)

Jim Berry, Greg Fischer, Raymond Guiteras

Trust and Microfinance in Poor Communities in Peru

We evaluate a novel microfinance model in which new customers need to gain sponsorship by an existing customer. We investigate how relationships between individuals and social networks impact repayment behaviour. This individual lending program screens clients and enforces good practices much in the same way as more traditional group lending does, but allows microcredit to be extended to those who might not qualify or be interested in a group liability loan.

See here for a similar study in the Philippines.

Policy Issue: 

Microfinance has generated worldwide enthusiasm as a potential catalyst for economic development and poverty reduction. The success of microcredit in providing access to capital without increasing default rates, despite a lack of physical collateral, was originally attributed to the group liability model, in which groups of people are jointly responsible for one another’s loans. However, as the microfinance industry grows and becomes more competitive, institutions must strive to develop new financing methodologies that keep institutional costs low while also extending access to credit. A major problem in microfinance is reaching borrowers who don’t qualify for or are not interested in communal, group liability, banks. Thus, different microfinance structures are needed that reach the poor with individual loans, while still harnessing some of the screening and enforcement benefits of group lending.

Context of the Evaluation: 

Since returning to democratic leadership in 1980, Peru has struggled to regain economic stability and growth. Currently, 44% of its 29 million people live in poverty.1 This plight has driven many rural residents to the outskirts of Lima in search of work, where they make their homes in self-built shantytowns that surround the city’s center. These shantytowns now contain a large proportion of Lima's inhabitants, and their residents have limited access to formal financial services such as savings accounts or loans. This study is located in fairly diverse shanty communities in Ancash, just north of Lima. The economy of these communities is primarily based on mining of gold, copper and zinc and fishing.

Details of the Intervention: 

In collaboration with PRISMA, a Peruvian NGO offering credit through village banks, researchers designed and implemented a new loan product and administered surveys to 9,000 shantytown households. This program sought to use social connections to screen for responsible clients, outside of the traditional group lending model, by requiring new clients to match up with sponsors who were already bank clients in order to obtain a loan.

Existing communal bank members acted as a pool of potential sponsors who can cosign small, individual loans for residents of the community who are not already bank members. The sponsor was responsible for repaying a loan if the client defaulted, and thus they were incentivized to cosign with more responsible individuals whom they could easily monitor. Each adult household member in the village received a card, which outlined the rules of the program and included a list of all sponsors in the community as well as a map of the community showing sponsor location. Both spouses of a sponsoring household and a borrowing household had to act as co-signers.

The two pilot shantytown communities consisted of 282 households with 26 sponsors, and 371 households with 25 sponsors, respectively. Social network surveys conducted in the communities before the implementation of the loan program allowed researchers to map the relationships between clients and sponsors. Researchers measured the strength of connections between individuals by time spent together per week, and whether individuals were considered trustworthy. Interest rates were randomly assigned between 3% and 5% a month across all client-sponsor pairs, in order determine whether the interest rate or the social distance from one’s sponsor had a greater impact on the likelihood of default.

Results and Policy Lessons: 

Reporting early results from the two pilot communities, researchers found that close social relationships and geographic closeness between sponsor and client effectively improves trust between agents, reducing the likelihood of default and the risk of cosigning such a loan. Estimates of the relative effectiveness of interest rates and social connections at reducing defaults suggest that lending with a close neighbor reduces the likelihood of default by the same amount as a 3-4 point decrease in the interest rate.

These findings underscore the prediction on which the program was founded, namely that borrowers with close social relationships to their sponsors allow the bank to be more certain that this new client will repay their loan. Increased information from close social relationships ensures the sponsor, and thus the lender, knows what risk each client represents, and can act to minimize that risk. This individual program therefore effectively screens clients and enforces good practices much in the same way group lending does, and allows microcredit to be extended to those who might not qualify or be interested in a group liability loan.

1 CIA World Factbook, “Peru,” https://www.cia.gov/library/publications/the-world-factbook/geos/pe.html.

Dean Karlan, Markus M. Möbius

Small Consumer Loans for the Working Poor in South Africa

Expanding access to business credit is a goal shared by microfinance practitioners, policymakers, and donors alike. However, there is less of a consensus when it comes to expanding access to consumer credit. Even as microfinance institutions increasingly move beyond entrepreneurial credit and offer consumer loans, practitioners and policymakers continue to voice their concern for “unproductive” lending. This study seeks to address these concerns by examining the impact of a consumer credit supply expansion.

We found significant and positive effects on job retention, income, the quality and quantity of food consumption, control over household decision-making, and mental outlook for these borrowers. We only find negative effects on other aspects of mental health, principally stress.

 
Policy Issue: 

An important means of exiting poverty is access to productive resources, yet many poor people lack the capital necessary to invest in higher education, smooth consumption, or start a business. Expanding access to credit is a common means to enable participation in the economy, and there is a common assumption that expanding access to productive credit makes entrepreneurs and small business owners better off. There is less consensus however, on whether expanding access to credit to support consumption helps borrowers, particularly when loans are being extended at high interest rates to higher-risk customers. 

 

Context of the Evaluation: 

Poverty in South Africa is widespread; approximately 57 percent of individuals were living below the poverty line in 2001.  Numerous impoverished areas could potentially benefit from increased access to credit to help smooth household consumption. However, the poor typically lack the credit rating and/or collateral needed to borrow from traditional institutions such as commercial banks.  Moneylenders dominate the informal lending market and typically charge 30-100 percent interest per month.

The cooperating Lender has operated for over 20 years as one of the largest, most profitable microlenders in South Africa. It offers small loans at high interest over short periods of time, frequently to the working poor who have no collateral and must make payments on a fixed schedule. 

 

Details of the Intervention: 

This evaluation examined the direct impact of small consumer loans on profitability, credit access, investment, and measures of well-being such as household consumption and physical and mental health. The sample consisted of 787 rejected loan applicants deemed potentially creditworthy by the Lender. Applicants were eligible if they had been rejected under the Lender’s normal underwriting criteria but not found to be egregiously uncreditworthy by a loan officer. The motivation for increasing credit supply for a pool of marginal applicants is twofold. First, it focuses on those who stand to benefit most from expanding access to credit, namely the unbanked poor. Second, it provides the Lender with information about the expected profitability of changing its selection process to examine marginally creditworthy individuals more closely.

A random portion of the eligible applicants were then assigned a “second look” by lender staff who were encouraged but not required to approve a randomly selected portion of these applicants for loans. Ultimately branches made loans available to 53 percent of the previously rejected applicants who had been randomly assigned to be re-examined. Accepted applicants were offered an interest rate, loan size, and maturity per the Lender’s standard underwriting criteria. Nearly all received the standard contract for first-time borrowers: a 4-month maturity at 200 percent APR.

Applicants in the treatment and comparison groups were surveyed six to twelve months after applying for a loan to examine behavior and outcomes that might be affected by access to credit, including mental health outcomes. 

 

Results and Policy Lessons: 

Impact for Borrowers: Expanding access to credit is found to significantly increase certain elements of well-being of borrowers. Economic self-sufficiency (employment and income) was higher for treated applicants than for those in the comparison group 6 to 12 months after treatment. Twenty-six percent of treated households report that the quality of food consumed by the household improved over the last 12 months. A subjective measure of “control and outlook”— comprised of factors such as intra-household bargaining power, community status, and overall optimism— is also higher for treated applicants. 

Long-Term Impact on Creditworthiness: Over a 13 to 27 month horizon, study results indicate a positive impact from having a credit score: having an ordinal score due to their credit history increased the probability of future loan approval in the sample by 19 percent, though there was no impact on the score itself. 

Impact on Mental Health: Receiving greater access to credit had mixed effects on the mental health of study participants, and results indicated that the mental health impacts of taking up a small individual loan may differ by gender. While the program had no significant effects on the mental health of women, men experienced increased symptoms of perceived stress and decreased symptoms of depression. The fact that even “good” major life events, such as starting a new job, can be stressful at times may explain why men in the treatment group experienced increases in perceived stress as they took up the loans and engaged in new economic activities. The positive impacts that increased credit access had on other areas of life (described above) may explain why symptoms of depression were reduced among men despite the increase in perceived stress.  

Profitability: Offering loans to marginal applicants, formerly rejected by the Lender’s usual screening process, is also found to be profitable for the Lender, although it is still substantially less profitable than offering loans to more creditworthy applicants. 

1 Southern African Regional Poverty Network (SARPN), “Fact Sheet: Poverty in South Africa,” http://www.sarpn.org.za/documents/d0000990/

 

Selected Media Coverage:
Microlending: It's No Cure-All - Bloomberg Businessweek
Rita Hamad, Dean Karlan, Jonathan Zinman

Savings Account Labeling and Financial Literacy Training for Susu Customers in Ghana

IPA is working with Mumuadu Rural Bank (MRB) to study the response to and impact of a new account labeling savings product. Working with Susu customers and Susu agents, the study compares the success of this new product with the current Susu savings product. The new savings product has only a psychological difference: it allows the labeling of funds within an account so that deposits can be directed to a specific goal, such as health, education or business savings.

IPA is working with Mumuadu Rural Bank (MRB) to study the response to and impact of a new account labeling savings product. Working with Susu customers and Susu agents, the study compares the success of this new product with the current Susu savings product. The new savings product has only a psychological difference: it allows the labeling of funds within an account so that deposits can be directed to a specific goal, such as health, education or business savings.

Watch Project Associate Hana Freymiller talk about this project on Youtube. 

Policy Issue:

Saving is hard for most people, rich or poor, educated or not. Setting aside even small sums of money on a regular basis requires a conscious trade-off between buying something now in favor of achieving long-term goals, and even the most prosperous struggle to translate this intention into sustained savings. Saving may be especially difficult for poor individuals, as daily needs and family obligations may distract attention from meeting savings goals.

Poor individuals not only have less income, but often face additional barriers to savings. They tend to be the least educated about their financial options, have the least access to secure financial institutions and are the least able to afford financial mistakes. Due to a variety of challenges, savings rates are quite low across the developing world and individuals often go into debt to maintain family well-being.

Context of the Evaluation:

Ghana's Eastern Region has a vibrant microfinance sector populated by a wide range of formal and informal institutions, and uniquely characterized by a prevalence of "Susu" collectors: traditional savings collectors who walk a daily path through town to collect Susu, "small small moneys", from their customers. Typically, Susu collectors return the funds to their customers at the end of the month in exchange for one day’s worth of collections.

As banks moved into rural areas, they have formalized Susu collection, paying their agents on commission and not charging their customers a direct fee for the service. Competition between banks is highly visible in the urban marketplaces where Susu agents, clothed in the bright batiks of their respective institutions, fight for the patronage of the same group customers.

Description of Intervention:

Researchers collaborated with Mumuadu Rural Bank (MRB) in the Eastern Region of Ghana to test the impact of a new type of savings account aiming to help clients save by focusing attention on savings goals. The evaluation seeks to understand if a purely psychological savings product, which encourages customers to earmark account funds for a specific financial goal, increases savings rates.

Study participants were active savings customers of Susu agents at Mumuadu Rural Bank in five urban and rural communities across Eastern Region in Ghana. Among them, half were randomly selected to receive an offer of the labeled savings account, while the remaining customers continued to access existing savings services from the bank. The new labeled account shared all the characteristics of the regular Susu account with the addition that customers could “label” funds for particular expenditures, such as buying a house or paying children’s school fees. After labeling the account, customers stated how much they planned to save over the next six-month period. The bank provided each customer with a free passbook that had the personal savings goal written on the front as a reminder.

Mumuadu Rural Bank staff were responsible for maintaining the accounts once they had been opened and Susu agents continued their normal rounds, collecting funds for the labeled account alongside the regular Susu savings accounts. Researchers tracked the take-up of the new product and savings activity over six-months among all participating customers.

Preliminary Results:

Preliminary results found that customers with a labeled Susu savings account show a 31.2 percent increase in total deposits after nine months of account operations as compared to Susu customers without the labeled account. This increase is statistically significant across the five study branches, though the effect size varied in each community.

Over the study period, withdrawals by customers with the labeled account were not significantly higher than customers without the labeled account, indicating that these funds provided a stable source of additional capital for Mumuadu Rural Bank. While customers with labeled accounts showed greater savings rates, there was no difference in their expenditure patterns from regular Susu customers.

Additional data is currently being collected and analyzed to determine if these impacts are sustained and if there are identifiable trends in the timing of deposits and withdrawals.

Dean Karlan, Edward Kutsoati, Margaret McConnell, Margaret McMillan, Christopher Udry

Group versus Individual Liability for Microfinance Borrowers in the Philippines

Microcredit has become a popular anti-poverty policy in the last decades. Now with more than 150 million borrowers, microcredit has undoubtedly increased access to formal financial services for the poor.

Policy Issue:

Microcredit has become a popular anti-poverty policy in the last decades. Now with more than 150 million borrowers, microcredit has undoubtedly increased access to formal financial services for the poor. An extensive debate exists about the advantages and disadvantages of group liability, where a group of individuals are all responsible for each others’ loans if one member defaults, versus individual liability, where only the borrower is at risk if they default. Group liability may improve repayment rates but it also raises the possibility that bad clients will take advantage of good clients in their liability group.

While individual liability lending may address some of these issues, it also has potential drawbacks in the form of less intensive screening of members, and higher default rates due to the lack of member responsibility to cover group members’ loans. Additionally, credit officers may spend more time in securing payment or using a more intensive and time-consuming credit investigation and background checks.  

Context:

In the Philippines 25% of the population live below the national poverty line[1], and many depend on small and individual enterprise for their livelihood. The islands of Leyte, Cebu, and Bohol, where this study takes place, host a wide range of economic activities, including farming, fishing, manufacturing, and commerce. As is true in much of the Philippines, most of this area has been heavily penetrated by microfinance institutions. Rural banks, cooperatives, and NGOs offer both individual- and group-liability microcredit loans and competition is strong. Most of the lending centers involved in this study are located in small towns or rural villages, though some are located in mid-sized cities. The majority of the members of the Green Bank of Caraga, the sample of this study, are microentrepreneurs engaged in small-scale sales or activities such as tailoring, food processing, and small-scale farming. The average loan size for this sample is US$116), not an insignificant amount when compared against the Philippines GDP per capita of $3,300.[2]

Description of Intervention:

Researchers examined two trials conducted by the Green Bank of Caraga to evaluate the effects of group liability relative to individual liability on monitoring and enforcing loans.

In the first trial, a randomly selected half of the bank’s 169 existing group-lending centers on the island of Leyte were converted to the individual liability model, phased in over time.Researchers could then isolate the impact of group liability on behavior through peer pressure by comparing the repayment behavior of existing clients in group-liability centers and converted individual liability centers. Centers were then assigned to comparison, individual liability or staggered individual liability (the first loan for each member is covered by group liability, but subsequent loans have individual liability). Critical to the design is the fact that individual-liability centers were converted from existing centers, and not newly created. By comparing the repayment behavior of existing clients in group-liability centers and converted centers, researchers were able to isolate the impact of group liability on employing peer pressure to mitigate moral hazard. 

In the second trial, the sample consisted of 124 randomized communities in areas where the bank was not yet operating. Once feasible villages were identified, an independent survey team conducted a business census, a household roster, and a social network survey. Each of these villages was randomly assigned into one of three treatment groups before the bank established lending centers: liability program, individual-liability program, and group-liability program converted to individual-liability after the first cycle.

Results:

After three years, researchers found that individual liability compared to group liability leads to no change in repayment rates (clients in individual liability centers were no more likely to default than their peers in group liability centers) in the short as well as the long term. The removal of joint liability resulted in larger lending groups, hence further outreach and use of credit but the average loan size was smaller, leading to no change in overall group profitability. Loan sizes in converted groups were lower because members were more likely to withdraw savings, lowering their capacity to borrow. Under individual liability, members were also less likely to be forced out of their center, because they could only be removed by credit officers—not peers. Thus, individual liability made existing centers 13.7 percentage points less likely to be dissolved.

Bank officers in new areas were lesswilling to open groups despite the fact that there had been no increase in defaults. This constrained the growth of the lending program.



[1] United Nations Human Development Report, “Human Development Indices,” http://hdr.undp.org/en/media/HDI_2008_EN_Tables.pdf (accessed August 25, 2009).

[2] As of 2008. CIA World Fact Book, “The Philippines,” https://www.cia.gov/library/publications/the-world-factbook/geos/rp.html , (accessed Nov, 20, 2009).

 

Xavier Giné, Dean Karlan

Using a Referrals Program to Test for Asymmetric Information and Borrowers’ Ability to Influence Repayment Behavior of Fellow Borrowers in South Africa

In recent years, the shift in microcredit from group-liability to individual-liability has accelerated. The most commonly cited driving force behind this trend is borrowers’ aversion to both group tension and free-riding by fellow group members. Usually, the shift to individual-liability microcredit results in the loss of group-liability's two biggest advantages: peer screening and peer pressure. In this study, however, we test a “client referrals” promotional campaign that seeks to recapture these efficiencies.

Our partner lender is interested in increasing the number of first-time applicants via a client referrals promotional campaign. In addition, this lender would like to attract good applicants and encourage good repayment behavior. By randomizing the rules of the promotion, we will be able to detect whether clients have asymmetric information on the borrowing “type” of fellow borrowers, and if clients are able to influence the repayment behavior of fellow clients. Based on this information, we will be able to design an optimal referrals program going forward.

Our partner lender operates in South Africa’s “cash loan” industry, offering high-interest, short-term credit with fixed repayment schedules to a “working poor” population. As part of the promotional campaign, first-time borrowers receive two Refer-A-Friend Vouchers with their loan. These vouchers follow one of four sets of rules that dictate how the client becomes eligible for a Referral Bonus. The randomized variation in the rules will allow us to answer our research questions.

Dean Karlan, Jonathan Zinman

Cosignatory Requirement as a Barrier for Women Accessing Credit in Peru

Microfinance institutions have long targeted women as recipients due to the belief that women more reliably pay back their loans and the increased access to funds serves to improve women's decision-making power in the household. However, some institutions implement a co-signature requirement in order for women to take out a microfinance loan. This may be acting as a barrier for women to access credit.

IPA has partnered with Microfinanzas PRISMA, an MFI in Peru, to identify the implications of the cosignatory requirement. The study is set up in four agencies in the Peruvian Highlands: Huancayo, Huaraz, Juliaca and Tarma. Through a geographic randomization of districts, the study divides all new communal banks in these areas into two groups: 1. The control group, in which the cosignatory requirement remains the same, and 2. The treatment group, in which the cosignatory requirement is removed. The study will use a baseline and follow-up survey to analyze whether the removal of the signature is allowing more people to join the banks and if the default rate is increasing. Other questions from the surveys will be within the realm of household dynamics and bargaining.

Dean Karlan, Melissa Tartari

Before and After Comparison of Innovative Mobile Applications for Micro Retailers in Colombia

Policy Issue:

Without a system for managing finances, small businesses may miss opportunities to increase profits and trim expenses. In particular, entrepreneurs in developing countries who rely on informal businesses to meet basic consumption needs may benefit from formal record keeping systems. Many of these entrepreneurs, often with little formal education and low levels of financial literacy, do not maintain records of business expenses or sales.  Providing small business owners with tools to manage their finances may be a way to improve both business outcomes and household consumption levels.

For additional information on current SME Initiative projects, click here.

Context of the Evaluation:

Colombia has an estimated 400,000 micro and small stores or "tiendas”[1]. In Colombia, these retailers comprise 52% of food and retail  sales.[2] While tienda entrepreneurs sell hundreds of different products, and have managed to keep contact with wholesalers, most continue to use unsophisticated business administration tools (writing down sales and purchases on notebooks) or none at all. Colombia also has one of the most penetrated information communication and technology (ICTs) markets: there are 92.3 cell phone subscriptions per 100 people, and 45.5% of the population uses the internet[3].

Frogtek, a firm specialized in creating business tools for entrepreneurs in emerging markets, identified that a major challenge for micro-retailers was managing their perishable inventories and figuring out how often they would need to restock them.  While several non-governmental programs have tried to address these issues through financial education and literacy training programs, shopkeepers continue to use unsophisticated methods, or none, and cannot easily determine the size of their inventory or actual profits.

Description of Intervention:

To address these challenges, Frogtek created “Tiendatek”, a smart phone application to help shopkeepers systematize their business by managing their accounting, inventories, sales, payments to suppliers, expenses and earnings. The tool is distributed through personal sales associates who visit the shopkeepers, demonstrate the product, and provide training on its use during the length of the pilot. All data generated by the shopkeeper is uploaded and stored on a Frogtek web server. Additionally the Tiendatek application creates simple and advance reports on sales, purchases, credit, inventory, and break-even points based on the data uploaded by the user.

This study pilot targeted clients of a microfinance institution, Bancamía, who are shopkeepers with sales of about 4.5 million COP (about 2500 USD) a month. Frogtek staff interviewed the candidates, assessed their interest, delivered the phone and provided training in one or two visits. Fifty participants were selected to receive a baseline survey approximately ten to twenty days after receiving the new phone and an endline survey, eight to ten months after the initial phone delivery.  Shopkeepers participating in the pilot received a phone, charger, SIM card with data access, and a TiendaTek manual free of charge. Those shopkeepers who become frequent users of the application by the end of the study, will be able to keep the phone and receive a three month data plan.  By collecting data on the use of the application, store production, sales, satisfaction, and perceived improvements, this comparative study will assess the use of Tiendatek, possibly for further analysis in a randomized evaluation.

Results and Policy Lessons:

Results forthcoming.
 

 

[1]Diaz, Alejandro, Jorge Lacayo, and Luis Salcedo. 2007. ”Selling to ‘mom-and-pop’ stores in emerging markets” The McKinsey Quarterly,

[2]De Jacobs, Alicia. “Colombia Retail Food Sector” USDA Foreign Agricultural Services Global Agricultural Information Network Report, October 2010.

[3]United Nations Conference on Trade and Development 2010 Information Economy Report

Antoinette Schoar

Increasing Savings and Reducing Reliance on Credit Card Debt for Low-Income Individuals in Washington DC

This project will evaluate the impact of commitment contracts and reminder messaging on savings behaviors among low- and medium-income credit union members in Washington DC.  Traditional financial products which dominate the consumer finance market tend to operate under the assumption that consumers act in a rational manner and fail to take into account cognitive biases which can impede the realization of financial goals. Here we test a savings product that includes two features designed to overcome these biases. A built-in commitment contract attempts to encourage consumers to forego present expenditures in lieu of future payoffs.  Regular messaging attempts to overcome limited attention, which may result in an inability to stick to a budget or savings plan.

Dean Karlan, Eldar Shafir, Jonathan Zinman

Insurance, Credit, and Technology Adoption in Malawi

Policy Issue:

The classic economic view of poor farmers is that their lack of a savings cushion causes them to prefer agricultural approaches with more reliable, but lower average returns. Farmers may see adoption of new technologies as risky, especially early in the adoption process when proper use and average yields are not well-understood. Weather and environmental factors can also pose significant risks. Risk and uncertainty can lead to low levels of technology adoption, particularly where resources to help farmers deal with risk, such as insurance, are not available. However, few studies have evaluated whether providing insurance can increase adoption of profitable agricultural technologies.

Context of the Evaluation:

Nearly all Malawian households (97 percent in 2004-2005) are engaged in maize production, but only 58 percent use hybrid maize varieties. Hybrid maize adoption in Malawi has lagged behind adoption in Kenya, Zambia, and Zimbabwe. Aversion to risk, credit constraints and limited access to information are among the most cited reasons why hybrid seeds and other technologies have failed to take hold in Malawi. Within Malawi, varying rainfall risk is by far the dominant source of production risk in Malawi, followed by pests.

Description of Intervention:

Researchers conducted a randomized field experiment to determine whether bundling rainfall insurance with a credit program (intended to finance adoption of a new crop technology) increased demand for credit. Researchers randomly selected 16 localities in central Malawi where farmers were offered credit to purchase high-yielding hybrid maize and groundnut seeds for planting in the November 2006 crop season. In another 16 localities, farmers were offered a similar credit package, but if taking the loan were also required to purchase (at actuarially fair rates) a weather insurance policy that partially or fully forgave the loan in the event of poor rainfall.

The microfinance institutions Opportunity International Bank of Malawi (OIBM) and Malawi Rural Finance Corporation (MRFC) offered loans for the hybrid seeds based on group liability contracts for clubs of 10-20 farmers. Take-up of the loan was an individual decision, but the subset of farmers who took up the loan were told that they were jointly liable for each others’ loans.

The weather insurance policy was customized to each of the four project regions (Lilongwe North, Kasungu, Nkhotakota, and Mchinji). Payouts were based on the rainfall readings at the closest weather station to the individual in question, and premiums were lower in places where the likelihood of a bad rainfall shock was lower. Compared with the annual interest for the uninsured loan (27.5 percent), a farmer taking out an insured groundnut loan faced an effective interest rate ranging from 37.8 percent to 44.4 percent, depending on the area.

All farmers in the study were administered a household survey that covered income, education, assets, income-generating activities (including detailed information on crop production and crop choice), measures of risk aversion, and knowledge about financial products such as credit and insurance.

Results and Policy Lessons:

Loan Take-Up:Take-up was 33 percent among farmers who were offered the basic loan without insurance. Take up was lower, at only 17.6 percent, among farmers whose loans were insured against poor rainfall. A potential explanation is that farmers already were implicitly insured by the limited liability inherent in the loan contract, so that bundling a loan with formal insurance (for which an insurance premium is charged) is effectively an increase in the interest rate on the loan.

It is also possible that farmers may have been uncertain about the risks associated with the hybrid seeds. For those in the treatment group, the fact that they were offered insurance may have served as a signal that the seeds were a risky investment. Lower take-up of the credit plus insurance product would then be a rational response.

Analysis indicates that farmers who are wealthier and more educated were more likely to take up the insured loan. By contrast, there is no indication that farmer education, income, or wealth is related with loan take-up in the uninsured loan group.

Xavier Giné, Dean Yang

Impact of Accountability- GlobalGiving Mentoring

Can communicating beneficiary feedback to donors augment philanthropic gifts or improve the operations of charitable organizations? It is unclear whether an organization with a higher level of accountability is more appealing to donors than one with a lower level of accountability.  From an institutional perspective, it is also unclear if greater transparency within an organization helps to increase its effectiveness in achieving its mission.  By providing direct feedback from program recipients to institutions and their donors, the effects of transparency on donor participation and organization functioning will be assessed.

Context of the Evaluation:

GlobalGiving is a US-based non-profit and online marketplace that allows donors to give directly to projects in developed and developing countries through targeted online donations. GlobalGiving serves as an intermediary between donors and local partner organizations that implement and monitor projects on the ground. The idea behind this type of targeted online giving is to reduce the number of intermediaries between the donor and the recipient so as to provide a stronger and more efficient vehicle for development finance. The model also allows donors to have greater control over the types of projects their donations will support and provides information about how specific dollar amounts are used to make a difference for each program.

Description of Intervention:

IPA’s intervention targets randomly selected donors from GlobalGiving’s online platform and organizations in three countries in which GlobalGiving currently operates: Ecuador, Peru and Guatemala.

The first phase of the study measures the impact of direct SMS/text messages from beneficiaries of local organizations on the level and frequency of donations from GlobalGiving’s online supporters. Existing donors were randomly assigned to one of two treatment groups. Those in the treatment groups were presented with information regarding the new beneficiary feedback measure at the same time that they were solicited for another donation with or without specific beneficiary messages.  Those in the comparison group received an email asking for another donation to a previously supported project.

New visitors to GlobalGiving’s website were also randomly assigned a treatment group, viewing project descriptions advertising feedback from clients, or a comparison group, viewing standard project descriptions. Researchers will analyze the giving behavior of individuals in treatment and comparison groups: the likelihood of an additional gift, average gift size, and whether or not donors give to additional projects.

The second phase of the study measures the impact of direct client feedback regarding the services provided by organizations on institutional behavior. Client satisfaction surveys and/or an external audit were used to measure changes in the effectiveness of the organization or any shifts in organizational behavior that result from the introduction of client feedback. In this phase, the treatment organizations received information about the client feedback program and were asked to encourage their clients to send SMS messages about services they receive. Clients were informed that that they would be reimbursed for the cost of sending the SMS message and would receive a small payment as an additional incentive. Beneficiaries of comparison group organizations were not eligible to give feedback to GlobalGiving. The goal of this phase of the project is to a) collect evidence on how organizations change their practices when feedback from donors is provided directly to donors and b) learn how providing donors with feedback affects organizations’ participation  in the GlobalGiving community. 

Results and Policy Lessons:

Results forthcoming.

Dean Karlan, Margaret McConnell, Jonathan Zinman

Using Behavioral Economics to Help Households Reduce Debt

This project will evaluate the effectiveness of innovative decision-making devices based on lessons from behavioral economics and psychology.  Financial outcomes in retirement are the product of a lifetime of decisions.  For many households, the most important financial decisions involve borrowing, not saving.  Rather than the rational creatures that standard economic models of behavior assume, humans are subject to numerous biases that may impede our abilities to make good choices.  This project will use successes from United States and developing countries to implement behaviorally-motivated interventions for helping U.S. consumers reduce their reliance on credit card debt and increase savings, with an eye towards the importance of early retirement planning. 

Dean Karlan, Jonathan Zinman

Starting a Lifetime of Saving: Teaching the Practice of Saving to Ugandan Youth

Policy Issue:

Programs promoting financial literacy and savings among the poor have become popular in recent years, in hopes that they may enable individuals and families to meet their financial demands. Developing a financially educated population may help promote large-scale change in a country’s economic situation by increasing the savings rate and thereby smoothing individual consumption and increasing investment in productive resources.  At this point, however, very little information exists as to whether or not financial education significantly improves savings behavior.  Particularly, the impact which financial literacy programs can have relative to other means of promoting savings, such as designing group savings accounts that leverage peer pressure, is not known. Quantifying this impact can help researchers understand what drives financial decisions, and enable policymakers to fund programs that will have the largest impact on savings behavior.

Context of the Evaluation:

The population of Uganda is disproportionately young: 52 percent of Uganda’s population is under 15 years of age and 29 percent of the country’s adult population is between 15 and 34 years of age[1].   In addition to being very young, the population of Uganda has extremely low savings rates, even relative to its sub-Saharan neighbors, which on average have the lowest savings rates in the developing world. Between 2001 and 2003, for example, Uganda’s average savings rate was 5.2 percent.   Its neighbor Kenya, by comparison, had an average rate of 12.7 percent for the same period.[2]  Uganda’s current savings rate remains alarmingly low, at 10 percent[3].

Founded in 1984, the Foundation for International Community Assistance (FINCA)International’s mission is to provide financial services to the world’s lowest-income entrepreneurs .  FINCA Uganda was founded in 1992, and hasexperience and expertise in providing financial assets, as well as, youth-oriented savings products to the world’s lowest-income entrepreneurs.

Description of the Intervention:

The Church of Uganda maintains a large network of youth fellowship groups, based out of village churches around the country. These groups were targeted for this intervention because they offer a high level of trust among members, as well as, a high degree of consistency across the different groups, relative to other youth group structures in Uganda.  Each group has an average of 25 members with a well balanced mix of genders and occupations.

This evaluation examines two interventions: a financial education curriculum (a knowledge-based intervention) and a specially-designed youth group savings account (an access-based intervention). The curriculum was developed in partnership with Straight Talk Foundation – a highly successful Ugandan organization specializing in communication to youth – based on the Your Future, Your Moneycurriculum from the Global Financial Education Program and materials from Binti Pamoja, an organization that promotes the rights of teenage girls.

The ten-session, fifteen-hour curriculum focused primarily on teaching concepts and skills for improving savings behavior, ranging from role-playing the differences between saving and borrowing to achieve a goal, to how to keep a budget, to strategies for successfully discussing sensitive topics around money.

The group savings account was designed without fees and with simple account-opening procedures to minimize the barriers that were found in focus group discussions to most discourage young Ugandans from opening accounts.

Two hundred forty church groups, representative of all of Uganda’s regions, were selected based on their level of activity and access to district capitals. The sample population was randomly assigned into four groups, including one group which received neither intervention and served as the comparison group:

 

Offered/encouraged to open youth group savings account

No account offered

Financial literacy training

Treatment Group A

Treatment Group C

No financial literacy training

Treatment Group B

Comparison Group (Group D)

In total 120 groups were offered the financial education program by FINCA-hired and IPA-managed financial educators and 120 groups were offered the group savings account.

Results and Policy Lessons:

Results forthcoming.


[1] Uganda Bureau of Statistics (UBOS) and Macro International Inc. 2007. Uganda Demographic and Health Survey 2006. Calverton, Maryland, USA: UBOS and Macro International Inc. Page 11.

[2] Bank of Uganda research department, Sept. 14, 2005.  Found in “Savings Habits, Needs and Priorities in Rural Uganda.” Prepared by Richard Pelrine, Olive Kabatalya. Rural SPEED and Chemonics International.  Produced by USAID, September, 2005.

[3] Dovi, Efam, “Africa: Boosting Domestic Savings on the Continent.”

Julian Jamison, Dean Karlan, Jonathan Zinman

Saving Incentives for Low and Middle Income Families: Evidence From a Field Experiment with H&R Block

Policy Issue: 

A significant share of low- and middle- income American families are thought to have inadequate savings for retirement. Families with income below $40,000 typically have low rates of coverage under employer-provided pensions, and tax deductions and other incentives often do not provide much incentive to contribute to Individual Retirement Arrangements (IRAs) because these households pay relatively little in taxes. Only two in five elderly Americans have a retirement savings account of any kind. Insufficient savings can reduce elderly households’ ability to cope with financial shocks.  

 
Context of the Evaluation: 

An alternative to tax benefits, government incentives such as savings matching may offer a more promising way to bolster savings among low- and middle-income households. Matching, where a percentage of savings is matched by the government in the form of additional deposits, is independent of an individual’s tax rate, and thus may provide an incentive for saving even for people in lower tax brackets. The Saver’s Credit, a federal program offering a tax reduction of up to 50 percent of funds contributed to a 401(k) or an IRA, offers one example of a matching program. Other matching contributions are present in many employer-sponsored 401(k) plans. Matching programs theoretically offer significant encouragement and benefits to families who desperately need to save for retirement, but the actual effects of matching on take-up and contributions are not well understood, and the willingness of families to contribute to retirement has remained low even in the presence of these programs. 

 
Details of the Intervention: 

The field experiment was conducted in conjunction with H&R Block, one of the world’s largest tax services providers. H&R Block offers its clients the option to split their tax refund between their IRA accounts and money for immediate use. Clients who engaged with this program in 60 H&R Block locations in the St. Louis metropolitan area from March 5th to April 5th, 2005, constituted the sample for this intervention study. 

Each client electronically prepared their tax return and was randomly assigned one of three match rates for their IRA contributions.  When a client was deciding how much to contribute to their IRA, they were told a (randomly chosen) rate at which their contributions would be matched, potentially influencing their contribution amount. Roughly one third received a 20 percent match rate, another third received a 50 percent match rate and the rest received a match rate of 0 percent, functioning as a comparison group. Individuals could receive up to $1,000 in matched funds. 

 
Results and Policy Lessons: 

Higher matching rates significantly increased IRA participation and contributions. Take-up rates were 3 percent, 8 percent, and 14 percent, respectively, for the comparison group, the 20 percent match group and the 50 percent match group. Conditional on take up, average contribution levels excluding the match funds were $765, $1102, and $1108, respectively. The 50 percent match rate raised take-up and aggregate contributions the most and did not reduce average contributions among participants, even as their number rose. The effects of matching on take-up rates and amounts contributed were particularly large for married tax filers.

Controlling for other factors, clients with a large tax refund, positive investment income, or higher overall income were more likely to respond to the matching offer.  Taxpayers were much more responsive to matching in this experiment than in the aforementioned Saver’s Credit program, perhaps due to differences in information, program structure, or in the framing of the programs. The combination of financial incentives, tax preparer assistance and the opportunity to use part of an income tax refund to save could generate increases in both the efficacy of federal tax incentives and the willingness of households to contribute to retirement savings accounts.

Esther Duflo

Group vs. Individual Micro-Lending in Peru

One of the most famous innovations of microfinance was the idea of “social collateral” – a way to guarantee the loans of people who have limited physical assets. However, it’s not clear that requiring group liability is actually a good thing. For instance, it can drastically raise the cost of a loan for a good client if she is forced to cover for other loans. Furthermore, it can force someone to guarantee people who take out much larger loans, which may prove to be impossible. It’s possible that, at least for some clients, individual liability loans may be better if the other mechanisms of microfinance (such as social embarrassment of being a debtor) ensure high repayment rates.

IPA ran a study in the Philippines testing this question, and found that repayment rate under individual liability did not go down, while growth increased. We are replicating the study in Peru.

The first phase of the study, in 2010, is to convert pre-existing communal banks (depending on if the group is in a rural area, the associations range from around 8-20 clients who all guarantee each other) to individual liability products, maintaining the rest of the group structure. Depending on the results, it’s possible that as Pro Mujer expands to new regions we’ll test impact with new associations.

We hope to implement financial diaries in the field in order to see, among other things, if clients under different liability structures have different approaches towards repaying their debts.

Dean Karlan

Using Encouragement to Overcome Psychological Barriers to Saving in Peru

This research examines whether bank marketing and communication tools can help individuals save more and, in particular, switch from informal savings vehicles to formal sector methods (e.g., a bank account). In conjunction with Caja Municipal de Ica (CMI), IPA examines various methods of product design, beyond the financial incentive, of encouraging clients to complete their savings commitment.

Policy Issue:

Microfinance has generated worldwide enthusiasm as a possible strategy to help people living in poverty get the resources they need to start a business, receive additional education, or make investments. While much of the focus of microfinance has been on microcredit, formal savings services can also have a dramatic impact on the lives of the poor. Savings are important both as insurance in the case of illness or other economic shocks, and as a way to purchase productive assets. Savings can also substitute the need for loans among clients who have enough funds to finance their expenditures themselves. But savings strategies are less tested than credit services, and microfinance institutions struggle to effectively expand their savings services.

Context of the Evaluation:

The semi-urban poor living in Ica and Ayacucho, cities in southern Peru, often earn income through small enterprises and self-employment. In Ica, agriculture represents the most important industry, while Ayacucho is well known for its artisans and handicrafts. Many of the poor in this part of Peru save through informal means. They often keep savings in their own homes, a practice referred to as a colchón banco (mattress-bank), or join Merry-go-Round savings groups called ROSCAs, where members pool their money into a pot, and each week or month a different member takes home the pot.  Due to their informal nature, both of these savings practices can be risky and unreliable.

Details of the Intervention:

Researchers will examine whether an initiative to promote savings can help individuals save more and switch from informal savings to formal sector methods. The study is implemented by the Caja Ica, a bank designed to serve the needs of poor clients with microsavings and microcredit programs, with program support from Catholic Relief Services (CRS) and technical assistance from COPEME. The Caja Ica is offering a new commitment savings product called “Ahorro Programmado”. Clients who choose to participate in this service commit to saving an amount of their choosing, amounting to at least 20 soles (US$6.50) per month for 6, 12, 18, or 24 months. As an incentive for meeting their savings commitment, clients receive a preferential interest rate of more than twice what the normal interest rate is for savings accounts.

This research will examine various product designs, beyond the already increased financial incentive, to see which are more effective at encouraging clients to complete their savings commitment. Each of the estimated 5,000 clients expected to enroll in the program will be randomly assigned to receive one or more of the following: (1) reminder letters before the due date of their payment, (2) token gifts upon payment to bring forward the "benefit" of saving, (3) positive or negative incentive messages on each deposit slip, or (4) no services, serving as a comparison. This study will determine the commitment device that most effectively encourages clients to meet their savings goals.

Results and Policy Lessons:

Clients began opening bank accounts in February 2006. The last group to be tracked completed their savings commitments in Oct 2007.  Preliminary results indicate that sign-up gifts, letters, and deposit slips all increased the probability that clients would reach their savings goal by several percentage points.  However, negatively framed messages appear to be more effective than the corresponding positive messages in getting people to save.

 

Dean Karlan, Sendhil Mullainathan, Jonathan Zinman

Financial Education and Commitment Savings Contracts to Reduce Credit Card Reliance and Mobilize Savings among Low-Income U.S. Households

This project will evaluate the effectiveness of financial education and commitment contracts in promoting higher levels of saving, reduced reliance on credit card debt and healthier financial portfolios among low-income individuals in the United States.  U.S. households in the bottom quartile of wealth spend, on average, more than they earn, and many low-income consumers lack formal savings accounts. Consumers tend to have time-inconsistent preferences for savings and consumption; they tend to be more impatient in the near-term than in the long-term and thus have a propensity to make purchases that are later regretted.  This project will evaluate the impact of commitment devices as a mechanism for mitigating time-inconsistent tendencies in spending, borrowing and saving. 

Dean Karlan, Jonathan Zinman

Chlorine Dispensers for Safe Water in Kenya

Two million children die of diarrheal disease each year and contaminated water is often to blame. Treating water with chlorine could substantially reduce this toll. The most common approach to chlorination in areas without piped water infrastructure is to offer small bottles of chlorine for sale to consumers.

However, chlorine use has been slow to catch on in this system. In this Kenyan study area, for example, less than 10% of households regularly use chlorine at a monthly cost of approximately US$0.30, despite several years of vigorous social marketing that has raised awareness about the product.

Based on this finding, the research team has developed a way to drastically cut the cost of chlorinating water by reducing packaging and distribution costs -- which account for the majority of the price of chlorine sold in individually-packaged bottles -- by installing chlorine dispensers at communal water sources. Users turn a knob on the dispenser to release a pre-measured dose of chlorine appropriate to treat the volume of water typically collected. The presence of a dispenser provides a reminder to treat water and harnesses peer effects to help increase take-up.

A randomized evaluation, in which provision of chlorine dispensers is phased in over time, is demonstrating the impact of the intervention on child health outcomes and will shed light on how the technology can be sustainably managed in a variety of settings. So far dispensers have been provided to 5,000 people at 20 rural water points.

Results:

During an unannounced visit three to six months after the installation of the dispensers, 61% of households in communities with a dispenser had detectable chlorine in their drinking water, compared to 8% of households in a comparison group. The percentage of households who use the dispensers was rising over time.

A second round of pilots is underway, with dispensers at a variety of settings, including schools, unprotected springs, and several urban sites. Work is underway to refine the dispenser hardware to further lower costs and develop strategies for marketing, cost recovery, and sustainable scale-up. The second round of the study will be completed in 2010, and further work to understand how to finance and maintain dispensers will be ongoing until the end of 2011.

Efforts are also underway to expand the program in Kenya and throughout the world. Chlorine dispensers could be appropriate for up to 2 billion people globally. Scaling up this approach globally could drastically alter the rural water landscape and save the lives of 100,000 – 250,000 children each year.

Do you want to help support IPA's work in providing safe water to people in Kenya?  Donate here and select the "Chlorine Dispensers for Safe Water" fund.

Michael Kremer, Jessica Leino, Edward Miguel

Text Message Reminders for Malaria Treatment

Increasing mobile phone ownership in developing countries presents exciting new opportunities for delivering public health and other social programs. Mobile phone subscriptions in developing countries increased from 7.9 per 100 inhabitants in 2001 to 78.8 per 100 inhabitants in 2011[1].

Increasing mobile phone ownership in developing countries presents exciting new opportunities for delivering public health and other social programs. Mobile phone subscriptions in developing countries increased from 7.9 per 100 inhabitants in 2001 to 78.8 per 100 inhabitants in 2011[1]. Rigorous studies of health promotion efforts delivered by mobile phones and other technologies can inform effective and efficient health programs The impact of text message reminders on patient adherence to short-term treatments, such as that for malaria, remains largely unexplored. This study is a first attempt to evaluate the impact of text message reminders on patient adherence to malaria treatment through a randomized controlled trial in Tamale, Ghana, from May to October 2011.

Results from this study, which is still underway, along with those from studies in Uganda (here and here), Kenya, and Zambia conducted in partnership with the Clinton Health Access Initiative, will be used to inform global policy on malaria diagnosis and treatment.

Policy Issue:

Despite the massive international efforts made over the past decades, malaria continues to be one of the primary causes of under-5 mortality worldwide. An estimated 1.24 million malaria deaths occurred in 2010, more than half of which were among children[2]. Of malaria deaths, 92% occurred in sub-Saharan Africa, where Plasmodium falciparum, the most virulent form of the malaria parasite, is most common[3]. P. falciparum has developed widespread resistance to several classes of antimalarial drugs, leaving artemisinins as the only known class of antimalarials that are effective at the population level.

P. falciparum becomes resistant to antimalarial treatments when parasites develop rare, random genetic mutations that prevent drugs from being effective. Artemisinin-based combination therapies (ACTs) are the first line treatment recommended by the World Health Organization because combination therapies lessen the likelihood that P. falciparum will develop resistance; parasites that mutate to become resistant to one drug should be killed by the other drug and there is a substantially lower likelihood that random mutations will confer resistance to both treatments.  Patients must complete the full dose of ACTs to most effectively prevent P. falciparum from developing resistance to artemisinins, but many patients do not finish their drugs. This study investigates the impact of text message reminders on adherence to ACT regimens. 

Evaluation Context:

Malaria is one of the predominant causes of illness in Ghana, especially among young children. Ghana is a pilot country for the Global Fund’s Affordable Medicines Facility – malaria (AMFm), which aims to expand access to ACTs by highly subsidizing their cost. Ghana has also been rolling out a National Health Insurance Scheme since 2004, which allows registered members to receive ACTs free of charge. This study took place in and around Tamale, the capital of Ghana’s Northern Region.

Description of the Intervention:

Data enumerators recruited respondents at public and private hospitals, clinics, pharmacies, licensed chemical sellers, and other vendors and followed up all patients that could be reached by motorcycle within 30 minutes of leaving the town center. Vendors identified individuals purchasing malaria medicine and gave them a flyer to enroll in a mobile malaria information system and directed them to data enumerators. Willing and eligible participants received baseline participation questionnaire soliciting the participant’s mobile number, directions to his/her home and possible times to locate the patient (or his/her primary caretaker) at home. Participants who enrolled in the text messaging system were randomized to a treatment group or a comparison group. Participants randomized to the treatment group received one reminder for each of the six doses of ACT over the course of 60 hours. Those assigned to the treatment group were randomized to receive a short message, “'Please take your MALARIA drugs!” or a long message, “'Please take your MALARIA drugs! Even if you feel better, you must take all the tablets to kill all the malaria.”

Data enumerators made home visits between 72 and 96 hours after the in-vendor recruitment, when the course of the ACT treatment was supposed to be completed. The main outcome of the study was adherence, which was assessed by detailed, per-dose self-report and by data enumerator observation of pill packets. Enumerators also asked respondents about malarial symptoms, care-seeking patterns, awareness of malaria and malaria medications. In cases when a home-visit could not be made in the requisite time frame, participants were followed-up over the phone.

Results forthcoming


[1] International Telecommunications Union. (2012). Mobile cellular subscriptions per 100 inhabitants. Retrieved from: http://www.itu.int/ITU-D/ict/statistics/.

[2] Murray, C., Rosenfeld, L., Lim, S., et al. (2012). Global malaria mortality between 1980 and 2010: A systematic analysis. Lancet, 379(9814): 413-431.

[3] Murray, C., Rosenfeld, L., Lim, S., et al. (2012). Global malaria mortality between 1980 and 2010: A systematic analysis. Lancet, 379(9814): 413-431.

 

 

Gunther Fink, Julia Goldberg

Responses to Degree of Control over Remittances in El Salvador

How do migrants decide how much money to send home in remittances? Would they like to have some control over how much of the money is spent and how much is saved? This study offered a variety of special bank accounts to migrants from El Salvador living in Washington DC, offering the sender varying degrees of control over an account held in the receivers name. Migrants offered greater control sent significantly more. Those offered some control over bank accounts roughly doubled their total savings in the combined trans-national household (migrant plus remittance recipient).

How do migrants decide how much money to send home in remittances? Would they like to have some control over how much of the money is spent and how much is saved? This study offered a variety of special bank accounts to migrants from El Salvador living in Washington DC, offering the sender varying degrees of control over an account held in the receivers name. Migrants offered greater control sent significantly more. Those offered some control over bank accounts roughly doubled their total savings in the combined trans-national household (migrant plus remittance recipient).

Policy Issue:

By the year 2000, individuals living outside their country of birth had grown to nearly 3% of the world’s population, reaching a total 175 million people. The money these migrants send home, called remittances, is an important but relatively poorly understood type of international financial flow. Recent research into the economics of migration has documented several beneficial impacts of remittance flows on household well-being and investments. However, research has only just begun to look at how migrants make their remittance-sending decisions, particularly if they desire greater control over how family members back home use the remittances they receive, and whether that impacts the amounts remitted. 

Context of the Evaluation: 

El Salvador is highly unusual among developing countries in its number of overseas migrants relative to the national population. After the 1980 civil war, large flows of Salvadorans emigrated, and continued to do so at a remarkably steady pace. At least one in seven Salvadorans now lives outside of the country, primarily in the United States. The total income of the approximately 1 million Salvadorans living in the U.S. was roughly equal to the El Salvador’s total GDP in 2001. Concurrent with the expansion of Salvadoran communities overseas, the dollar value of remittances sent to El Salvador has grown dramatically, from $790 million in 1991 to $3.8 billion in 2008.

Remittances appear to have significant benefits for recipients – households in El Salvador receiving more remittances have higher rates of child schooling, for example. But the lack of control migrants have over how remittance funds are used at home could be reducing the amount that they choose to send home. The fact that migrants report far higher preferences for saving, about 21% of income, relative to recipient households, who prefer to save less than 3%, supports this assumption. 

Details of the Intervention:

Researchers, in collaboration with Banco Agrícola, designed a field experiment that offered a way for Salvadoran migrants to directly channel some fraction of their remittances into savings accounts in El Salvador. To isolate the importance of migrant control over savings, researchers tested the demand for different products that offered migrants varying levels of control over remittance use. Baseline surveys were administered to both migrants in the U.S. and their corresponding remittance-receiving households in El Salvador.

The sample consisted of Washington DC area migrants who first entered the U.S. in the past 15 years and sent a remittance in the last 12 months. All 898 migrants received a marketing visit, where a marketer described the uses and benefits of savings, and were encouraged to save. They were also randomly chosen to be offered one of three new accounts in El Salvador that they could remit to. The account would either be (i) opened in the name of an individual in El Salvador, granting the recipient full control, (ii) a joint account where the recipient and the remitter would have access through ATM cards, allowing monitoring, but no enforcement on the part of the migrant, or (iii) an account only in the name of the migrant, providing full ability to control funds in the account. In the case of the first two accounts, project staff arranged by phone for the El Salvador remittance recipient to meet with the bank manager of the nearest Banco Agrícola branch to complete the final account-opening procedures. To help track migrants’ remittance behavior after the marketing visit, all visited treatment migrants were given a special card (called a “VIP card”) that provided a discount for sending remittances via the partner institution’s remittance locations. A final group of migrants received the marketing visit but were not offered an account, serving as the comparison group.

Results and Policy Lessons:

Results indicate that a desire for control over remittance uses – in particular the fraction that is saved in formal savings accounts – was large, and had significant influence on migrants’ financial decision making. When offered an account in the name of the recipient, allowing no formal control of the remittances, migrants were 16.4 percentage points more likely than the comparison group to open accounts. When offered joint control, migrants were 21.4 percentage points more likely to open accounts than people in the comparison group, and 34 percentage points more likely when offered exclusive control.  

While effects on savings at Banco Agrícola were substantial, there was also a substantial increase in savings outside of the partner bank, including U.S.-based banks. Researchers interpret this result as due to the financial advice offered as part of the treatments. Migrants implemented savings strategies suggested by the marketers but using savings facilities at other banks.

This resulted in large treatment effects in savings. Compared to a base of roughly $430 in reported comparison group savings, offering joint or exclusive control of bank accounts roughly doubles total savings in the combined trans-national household (migrant plus remittance recipient).

 

Nava Ashraf, Dean Yang

Supply Chain Financing for Dairy Farmers

Policy Issue:
Policy Issue:

Farming entails long cycles of production which require up-front investment in animals, equipment, seeds, fertilizers, and other inputs. However, small farmers may have problems securing access to credit if they are located in remote areas that are not served by traditional financial institutions. Many small farmers manage their businesses informally and frequently do not have records or financial information that banks require for lending. Some microfinance institutions have tried to expand their usual urban activities to rural clients, but the costs of doing business in rural areas are still high and limit their scope. However, farmers often have stable relationships with agriculture processing companies and traders who purchase their crops, and these relationships may provide an opportunity to facilitate loan distribution and repayment.

For additional information on current SME Initiative projects, click here.

Context of the Evaluation:

In Colombia, less than 8% of rural households and enterprises are thought to have access to formal loans[1]. Nevertheless, there are about 400,000 families engaged in small and medium agribusiness for the production of milk in rural Colombia.[2] Milk production requires daily contact between the producer and the buyer, with payment occurring frequently. Buyers tend to attract the best dairy producers by providing access to inputs for production, or other services, or rewarding quality with higher prices.

Description of Intervention:

Bancamía, a bank specializing in microfinance, partnered with Alquería, a Colombian dairy company, to offer an individual loan product to small dairy farmers. Four hundred thirty five small dairy farmers who sell milk to Alquería via three intermediaries were randomly assigned to a treatment group, receiving a loan product offering, or a comparison group, receiving no product offering.

Bancamía offered the farmers in the treatment group a micro-loan with  a unique repayment process. Each month when the loan installments were due, farmers did not have to travel to the bank office to make the payments. Instead, the Alquería dairy deducted the value of the monthly installment from the farmer’s milk transfer payment and paid the bank directly. This scheme reduced risk for the bank as well as transportation, planning and transaction costs for the farmers. Loans ranged from about one to five million COP (about 560-2,800 USD) and were granted over a one to three year period.

To promote the credit program, meetings were held at the offices of three of the Alquería buying intermediaries to introduce farmers to the program, the benefits of the product, the rules and obligations, and the loan application and repayment process.  At the end of the meetings, milk farmers in the treatment group were provided with bank contact information and the opportunity to file a credit application.  Both those who participated in the meetings and selected participants who could not attend the promotional meeting received a phone call reminder about the program.  Before, during, and after the program, surveys were administered to collect socioeconomic information, household wellbeing, and loan data.

Results and Policy Lessons:

Low levels of loan product take-up led to a discontinuation of the evaluation.  Two main issues affected the implementation.  Firstly, two Alquería intermediaries offered their own loan products, very similar to the product offered by Bancamía, which lowered the demand. Secondly, the beginning of the rainy season was unexpectedly devastating that year, causing major flooding, damage to farmland, and death of livestock.  While the evaluation was discontinued, Bancamía, Alquería, and the intermediaries have continued to service the 33 clients who applied and received loans.


[1]Colombia Rural Finance: Access Issues, Challenges and Opportunities. Rep. no. 27269-CO. World Bank, Nov. 2003. Web. 16 Mar. 2010

[2]Rivera, José Félix Lafaurie, “TLC…la batalla no ha terminado.” Carta Fedegan N. 117.

Antoinette Schoar

Expanding Access to Formal Savings Accounts in Malawi, Uganda, Chile, and the Philippines

Policy Issue:

Expanding financial services to reach the poorest of the poor helps to broaden their savings and investment options. Yet the vast majority of people in the developing world remain unbanked [1].  In the absence of formal savings products, many in the developing world depend on costly devices like "susus", agents who charge a fee to collect money for secure keeping, or illiquid devices such as rotating savings and credit associations (ROSCAs), groups that pool members’ regular contributions for lump-sum distribution. A majority of rural households store cash at home, “under the mattress” [2], where it is prone to loss, theft, and the demands of neighbors and kin.

Early experimental evidence from Kenya suggests that small business owners can benefit from access to a savings account in a formal bank [3].  By understanding the channels through which savings accounts might impact the lives of poor households, financial institutions—and the products they offer—can be better poised to serve this vulnerable population. 

Context of the Evaluation:

This study is ongoing in three countries: Malawi, Uganda, Chile. We are currently exploring options to add a study site in the Philippines. The aim is to understand the causes and consequences of the lack of access to banking services in a variety of contexts.

Malawi: Malawi’s population is largely rural, with agriculture supporting over 85 percent of the population [4].  A 2009 study found that 55 percent of Malawians do not have access to any type of financial institution, formal or informal; furthermore, only 19 percent of the overall population uses a formal bank [2].  In Malawi, the study is being conducted in the Southern Region, which has lower than average levels of bank usage and financial inclusion [2].

Uganda:A 2006 study found that 62 percent of Ugandans do not have access to any type of financial institution and only 16 percent of the population uses a commercial bank [2]. Total savings remains low: from 1999-2009, Uganda's gross domestic savings averaged 9.3 percent, compared to 21 percent worldwide and 12 percent for the least developed countries [3]. In Uganda, the study is being conducted across several districts in the southwest of the country.

Chile:Although Chile boasts an upper-middle-income economy, such prosperity is not universally shared.  In 2009, Region IX – the target region for the study – reported the highest incidence of poverty in the country at 27.1%, nearly double the national average [3].  And although Chile has a vibrant financial sector, an IPA-led pilot exercise in the target region revealed that 33% of the population does not use a savings account.

Philippines:The Philippines has reached a medium level of deposit account usage, with around 50 accounts per 100 people in 2009.  (Globally, there are more savings accounts than people.) [1] While this is to be celebrated, financial inclusion for rural populations has lagged, and 37% of municipalities did not have access to banking services in that year [7].  Recently, however, the mobile money sector has developed, enabling anyone with a mobile phone to conduct basic transactions.  Mobile banking services have great potential to provide access to the rural poor in this context.

To evaluate the impact of access to a formal savings product, the researchers have partnered with a commercial bank or credit union at each site.    

Description of Intervention:

In each site, IPA identified a partnering financial institution and selected rural areas in which the partnering institution is operating. A probabilistic sampling strategy was then used to enroll into the study a representative sample of unbanked households in those areas.  Upon enrollment and completion of a baseline survey, study households were randomly assigned to either a treatment or comparison group. Those assigned to the treatment group received a voucher enabling them to open an account, at no cost to themselves, with the local branch of the partner institution. They also received procedural assistance with account opening. Take-up of the account offer varied from 20% in Chile to 78% in Malawi.

Follow-up surveys at 6-, 12- and 18-month will be used to estimate the impacts on a range of household activities, including agricultural and business practices, expenditures, household income, response to shocks, and savings and credit practices.Qualitative data will be collected to understand the mechanisms through which access to a bank account affected (or not) the study participants.

Results and Policy Lessons:

Results forthcoming. 

 

1. CGAP. 2009. Financial Access 2009: Measuring Access to Financial Services around the World.  World Bank Group.

2. FinScope Malawi funded by the UK’s Department for International Development (DFID).  Demand Side Study of Financial Inclusion in Malawi.  2008. http://www.finscope.co.za/documents/2009/Brochure_Malawi08.pdf

3. Dupas, Pascaline and Jon Robinson.  2009.  Savings Constraints and Microenterprise Development: Evidence from a Field Experiment in Kenya.  NBER: Working Paper 14693.

4.  Reserve Bank of Malawi.  2008.  The Malawi Economic and It’s [sic] Banking System.   http://www.rbm.mw/documents/basu/MALAWI%20ECONOMY%20AND%20BANKING%20SECTOR.pdf

5.  FinScope Uganda. 2007. Results of a National Survey on Access to Financial Services in Uganda. Final Report. August. http://www.fsdu.or.ug/pdfs/Finscope_Report.pdf.

6.  World Bank. 2010. World Development Indicators 2010 [Online Database]. Washington, DC: The World Bank.

7. Jimenez, Eduardo C. 2010. Financial Access: An Essential Condition. Presentation at the OECD-Banque du Liban Conference on Financial Education.  http://www.oecd.org/dataoecd/57/21/46256657.pdf

Pascaline Dupas, Dean Karlan, Jonathan Robinson

Feasibility and Impact of Malaria Diagnostics in Ugandan Drug Shops

Malaria is one of the most common causes of illness in Sub-Saharan Africa. The standard first response to a suspected malaria episode is to purchase over-the-counter medication from a local pharmacy, bypassing the formal health care system altogether. Evidence is emerging that a large share of illnesses for which antimalarial medication is taken are not in fact malaria, but are rather bacterial or viral infections. A high rate of inappropriate treatment is problematic because it delays proper diagnosis and treatment for the true cause of illness, wastes precious resources (such as antimalarial subsidies) and possibly accelerates antimalarial drug resistance.

This study explores a method to increase access to affordable malaria diagnostics through retail sector drug shops. We investigate supplier incentives to sell and customer incentives to purchase rapid diagnostic tests (RDTs) for malaria in drug shops in Eastern Uganda. We sell heavily subsidized RDTs to drug shops and allow them to set the price, while simultaneously experimenting with methods to increase customer demand through behavior change communication messaging and social learning. The study also experimentally varies training modules for drug shop owners, in some cases emphasizing their role as primary health care providers in remote communities with poor access to the formal public health system. Finally, we exploit Uganda’s underlying variation in malaria endemicity to explore how financial incentives to sell RDTs are influenced by expected malaria positivity in an effort to understand the circumstances in which RDT subsidies can be most cost-effective.

Jessica Cohen, Gunther Fink
Syndicate content
Empowered by:
Advomatic
Birocreative
Copyright © 2011 Innovations for Poverty Action. All rights reserved.