The miracle of microfinance: Evidence from a randomized evaluation

Asia

India

Labour Market

Microfinance

Banerjee, Abhijit, Esther Duflo, Rachel Glennerster, and Cynthia Kinnan

2015

Banerjee, Abhijit, Esther Duflo, Rachel Glennerster, and Cynthia Kinnan. 2015. “The miracle of microfinance: Evidence from a randomized evaluation”. American Economic Journal: Applied Economics 7, no. 1 (January 2015): 22–53. © American Economic Association
View publication

-

2015

Spandana

The microfinance intervention examined in the study was rolled out in selected treatment areas roughly between April 2006 and April 2007. It was rolled out in control areas from May 2008.

For-profit provision of microcredit

Individuals are eligible if they are female, aged 18 to 59, lived in the same area for a minimum of one year, have valid identification, and form part of a self-selected group of 6 to 10 eligible women.

The provision of loans to groups of eligible women. Loans have to be repaid in 50 weeks at an interest rate of 12%. The first loan has a value of Rs 10,000 (about US$200). When the first loan is repaid in full, applications can be made for loans of increasingly higher value. Participating groups are jointly responsible for the repayment of loans made to their group.

• Education: enrolled in school
• Work: hours worked in the past 7 days
• Other: the study looks at a host of other outcome variables including receipt of credit from various sources (first stage), various indicators for the success of the household business, various indicators for self-employment, income, consumption, hours worked by adults and teenagers (16 to 20), and female empowerment.

Cluster randomized controlled trial

Study design: 104 slums (or permanent settlements) were randomly allocated to a treatment group in which the intervention was rolled out between April 2006 and April 2007 and a control group in which roll-out did not start until May 2008
Data: Two waves of evaluation data were collected: the first from August 2007 to April 2008 and the second from November 2009 to June 2010. Baseline data were collected prior to the rollout of the intervention, but these were not used in the identification of treatment effects.
Estimation strategy: The intent to treat effect of the program is estimated by regressing the outcome variables on the indicator for treatment and vector of slum-level baseline covariates. Treatment and control variables are shown to be similar in terms of variables that are time-invariant. Estimates are shown to be robust to corrections for attrition from follow-up wave 1 to follow-up wave 2.

The first stage effects of the program are limited. By survey wave 1, the probability that households in treatment areas received micro-credit were about 8 percentage points higher in treatment areas than in control areas. By survey wave 2, the probability of receiving micro-credit are not significantly different between the treatment and the control group, but the loans taken out in the treatment areas were larger than those in control areas. Effects on the probability of running a household business are limited. However, there is evidence that the intervention positively affects investment in businesses and business profitability, especially in the upper tail of the profitability distribution. Adults (household heads and their spouses) increase hours worked in self-employment activities. Effects school enrollment are not significant for children (5-15) teenagers (16-20), either at the first or second follow-up. There is some evidence that the program reduces hours worked by boys aged 5-15 in the second follow-up survey wave and girls aged 16-20 in the first follow-up survey wave. The intervention affects "inter-temporal consumption choices", as households from treatment areas increase purchases of durables (for their business) and lower the consumption of "temptation goods"

Child Labour

Households

Local

There is some evidence that the program reduces hours worked by boys aged 5-15 and girls aged 16-20.

Evidence that the intervention positively affects investment in businesses and business profitability. Adults (household heads and their spouses) increase hours worked in self-employment activities. The effects school enrollment are not significant for children (5-15) teenagers (16-20). The intervention affects “inter-temporal consumption choices”, as households from treatment areas increase purchases of durables (for their business) and lower the consumption of “temptation goods”.

-