Greetings!
Welcome to our fall newsletter. This edition is focused on new technologies and how they are being applied around the world in various attempts to address different development challenges. We take a look at how new ideas are beginning to transform the worlds of microfinance and entrepreneurship in developing countries.
As ever, we want to hear from you. Please do get in touch if you have any questions, comments or new ideas.
Randomly yours,
Delia Welsh
Managing Director
|
|
|
|
Text Message Reminders
Behavioral economics suggests that that it may be possible to "nudge" people to make better choices through very small interventions - for example, simply changing the ordering of choices can help consumers make healthier decisions.
Maggie McConnell, Dean Karlan, Sendhil Mullainathan, and Jonathan Zinman, set out to test an idea they had that people in poor countries might be "nudged" into saving  more. They decided to try the simplest nudge they could think of - a simple monthly reminder to save for people who had signed up to a new goal savings account. They set up an RCT in three countries: Bolivia, Peru, and the Philippines.
In all three countries, clients who had signed up to the new savings account were randomly assigned to either treatment or control. The treatment group received a reminder to save once per month. In the Philippines and Bolivia the reminders were sent by text message, in Peru where cell phones were less widespread it had to be the physical mail. And these reminders worked. Total savings increased by 6 percent, and people were 6 percent more likely to reach their savings goals. The effect was even bigger for savers randomly assigned to receive mention of their specific goal (versus a vague reference to savings) in their reminder.
These messages cost the banks pennies apiece. It was practically a free lunch.
To read more about this project you can download the full paper here, or a 4 page executive summary here.
|
|
Fingerprinting Microfinance Borrowers
Estimates from the Financial Access Initiative suggest that half of the world remains unbanked. The risk of default remains a major challenge inlending to poor individuals. Can fingerprinting help? Lenders are limited in their ability to punish bad repayment behaviour if they lack an effective way to link an individual to his or her loans. In addition the timing of repayment may affect some borrowers ability to make good on loans. While strategies such as frequent repayment schedules and group liability are popular with MFIs that lend in urban and peri-urban areas, they are often not feasible in an agricultural context due in part to the seasonal nature of income.
In Malawi, where 85% of the population lives in rural areas and the economy is largely based on agriculture, addressing these challenges is of great interest. Lenders who are not confident in their ability to identify good borrowers are likely to lend less than would be optimal. In rural Malawi, access to finance is limited, and the lack of available credit prevents farmers from financing potentially profitable investments such as fertilizer and improved seeds.
The Malawi Rural Finance Corporation (MRFC) offers loans to farmers that allow them to purchase paprika grower "starter kits" that are available from Cheetah Paprika Limited. Unfortunately, farmers were sometimes able to avoid paying back their loan
in one season and then re-borrow in another season using a different name.
Researchers Xavier Giné, Dean Yang, and Jessica Goldberg worked with MRFC to create a more reliable system for identifying borrowers and tracking their loan repayment; a borrower fingerprinting program. Loans were made to individuals within agricultural clubs, which were randomly assigned to have their members fingerprinted as part of their loan application process. All agricultural clubs received an identical training session from bank staff emphasizing the importance of credit history to future borrowing.
The results of the fingerprinting program suggest that the additional layer of security acted as a credible threat to poor repayment. Fingerprinted borrowers were up to 39% more likely to repay their loan than non fingerprinted borrowers. Borrowers who were fingerprinted also took out smaller loans, perhaps to ensure their ability to repay. Fingerprinting worked especially well on those who were predicted to have worse repayment rates based on pre-intervention characteristics.
The researchers also looked into the cost effectiveness of the program in order to determine whether implementing it makes financial sense for lenders. Even with conservative estimates of the benefit of increased repayment and the cost of outsourced fingerprint matching, the results suggest an attractive cost-benefit ratio.
To read more about this project you can see the project summary on the IPA website here, or download the full paper here: Identification Strategy: A Field Experiment on Dynamic Incentives in Rural Credit Markets |
|
Computer-assisted learning
Technology is often viewed as a way to connect students in developing countries with learning resources that are not otherwise available to them locally. Much excitement has been generated by initiatives such as One Laptop Per Child and Sugata Mitra's"hole in the wall" projects. Is technology the solution to educational challenges? IPA affiliates Abhijit Banerjee, Esther Duflo and Leigh Linden worked with Pratham, an education-oriented Indian NGO, to evaluate a computer-assisted learning intervention in primary schools in Gujarat. They found that the program did improve students' skills - but they also concluded that it was not the most cost-effective way to give students a leg up. 
The intervention took advantage of a government program that placed four computers in each of 80% of primary schools in the city of Vadodara in Gujarat. Pratham designed a program that supplemented regular classroom instructionwith computer-assisted learning. Students in treatment schools received basic instruction on how to use the computers, and then spent two hours per week of shared time (two students per computer) working independently with educational software. The study sought to evaluate the effects of computer-assisted learning on student skill level, and to determine if the effects would persist over time and be cost-effective compared to alternatives.
The program worked. In the first year, math scores for treatment students increased by 0.36 standard deviations, a substantial achievement when compared to other education interventions. This improvement also persisted to some extent after a year, but further research is needed to fully access long-run impacts.
This is not the end of the story, however. The computer-assisted learning program was tested at the same time as a remedial tutor-based program, Balsakhi. Although computer-assisted learning was shown to be highly effective in raising students' skill levels in math, it was less cost-effective than the tutor-based Balsakhi program. This is why IPA is considering remedial tutors in our list of best-buy "tested and proven" ideas (the Proven Impact Initiative), whereas computer-assisted learning is not. Cost-effectiveness is important as well as impact!
To read more about this project you can see the project summary on the IPA website here, download the full paper here, or read news coverage here.
Leigh Linden has also worked on a similar project in Colombia with Felipe Barrera-Osorio of the World Bank. The evaluation found neglible impacts of placing computers in schools, the main reason seeming to be a failure to integrate the computers into the curriculum. All of which goes to show the importance of paying attention to the details of an intervention design, and of replicating our studies in different contexts.
You can read a project summary here or download the full paper here.
|
|
| Contributions to this newsletter were made by Lee Crawfurd and Meredith Startz. |
|
|
|
| IPA in the News |
|
A number of IPA research affiliates discuss the recent crisis in microcredit in India in an op-ed in The Indian Express.
Melinda Gates recently announced funding of $500 million for financial services for the poor. A $7 million grant to Yale and IPA was announced at the recent Global Savings Forum, for new research into micro-savings and payments technologies.
Dean Karlan was quoted in the Financial Times on the impact of mobile phones on development: "to help make markets more efficient people need information ... we've seen clear evidence of market prices being disseminated through cell phones."
IPA is one of the first grant recipients from USAID's new Development Innovation Ventures Fund. The money will go to a project in Sierra Leone offering finance and storage for produce to farmers.
IPA Liberia Country Director Tricia Gonwa and Research Affiliate Chris Blattman were interviewed by the BBC about research with street youth in Liberia.
Business Week profiled IPA Research Affiliates Esther Duflo and Abhijit Banerjee in an article The Pragmatic Rebels.
Research by IPA Research Network Member Mushfaq Mubarak on Bangladeshi garment factories was featured in the New York Times.
The November/December issue of the Boston Review features a debate on "Can Technology End Poverty" with contributions from our board member Kentaro Toyama, and Dean Karlan.
|
| |
| International ICT for Development Conference |
|
IPA President and founder Dean Karlan will be speaking at the International Conference on Information and Communication Technologies and Development, held in London on 13th-16th December 2010.
|
| |
| We're hiring! |
|
IPA is currently recruiting for a number of positions based in the US, Africa, Latin America and Asia.
Check out the IPA website for details.
|
| |
|
|
Print this page
