The Economic Returns to Social Interaction: Experimental Evidence from Microfinance

Principal Investigators: Benjamin Feigenberg, Erica Field, Rohini Pande
Research Team: Emmerich Davies, Sitaram Mukherjee and Anup Roy
LEAD Centre: Centre for Microfinance
Focus Area: Credit
Project Geography: West Bengal
Partner: Village Financial Services
Status: Completed


The penetration of formal insurance among low-income populations in developing countries remains low. Under such circumstances, some researchers theorize, social capital can lead to better informal risk-sharing among households. By increasing the social ties between individuals, development organizations may be able to expand or deepen their informal risk-sharing networks and help the poor hedge against individual or household level-risks. Some experts believe that greater social capital can lead to other advantageous economic outcomes for the poor such as increased cooperation. For these reasons, many development organizations attempt to build social capital through programs that focus on community interaction and group activity.

In this study, CMF researchers tested whether one development intervention that involves constant group interaction, standard microfinance lending, builds social capital. By closely tracking borrowing outcomes for groups that meet more frequently, researchers attempted to quantify the returns to enhanced social capital.


Researchers used several randomised experiments to test whether or not increased microfinance meeting frequency leads to increases in social capital. The study was conducted in the impoverished urban and peri-urban parts of West Bengal, India, with the help of Village Welfare Society (VWS), an MFI operating in the state. Researchers randomized traditional microfinance borrowing groups into either a monthly repayment cohort or a weekly repayment cohort.

There were broadly two kinds of experiments:
1.     Testing for effects of more frequent interaction – Researchers chose 100 first-time loan client groups, each consisting of 10 members. The members of a group generally lived in close to each other or were neighbours. Surveyors randomly selected 30 of these groups to weekly meetings and the rest (70 groups) to attend monthly meetings. The MFI provided approximately Rs. 4000 in loans to each member. The first group had to make weekly repayments of Rs. 100 starting two weeks after loan disbursal and the second group was required to make monthly repayments of Rs. 400 after the first month.
2.     Testing for trust among group members – Researchers ran another experiment more than a year later in which select loan clients from the same groups were given lottery tickets to thank them for participating in the first loan experiment. The client could choose any number of her group members to receive lottery tickets. If the client trusts that the other members would share their winnings with her either directly or through reciprocal behaviour at a later point in time, this increases her expected gain and hence should motivate her to give out more tickets. The tickets are either of Rs. 200 or divisible into Rs. 50, these were randomly assigned to filter the effect of trusting others in the group from altruistic motives.


The study finds that weekly group members were 26% more likely to meet fellow members outside the group meetings compared to women who met monthly. The effects were stronger for women who lived closer to each other. Also, researchers found that while only about 10% of monthly group members had met everyone in their group other socially in the last 30 days, 100% of members in the weekly group had met socially.

The increase in social interaction among members who met more frequently was accompanied by increases in risk-sharing. The researchers found that weekly group members were 29% more likely to make transfers to their friends and distant family after the loan cycle. Though researchers could not directly trace borrower to borrower transfers, they observed increased risk sharing through the trust game. Weekly clients were 48% more likely to distribute Rs. 50 vouchers, an outcome which suggests that weekly clients, more so than monthly client, saw their fellow group members as individuals who would make reciprocal exchanges.

Most significantly, default risk decreased for the groups that met more frequently – weekly group members were 3.5 times less likely to default on their loans compared to the monthly group members. Researchers conclude that in this case, default decreased, not because of the pressure to repay, but rather because individuals with enhanced social capital (weekly members) were more likely to share risk and make monetary transfers.

Policy Implications
–     Higher meeting frequencies result in reduced default risks, which should theoretically reduce costs for the lending organization. However, the cost of conducting more frequent meetings is also high for the MFI and for the individual. Because both monthly meetings and weekly meetings have significant costs and benefits associated with them, it is hard to conclude that one approach is clearly better than the other. Given the short-term and long-term increases in social capital associated with weekly meetings, organizations holding monthly meetings should consider moving to a weekly schedule. To further inform the policy debate, academics should conduct on the costs borrowers incur in attending repayment meetings.
–     Many have discussed the merits of individual liability versus joint liability methodologies in microfinance. This study builds on literature that suggests that the joint liability feature of JLG is not the mechanism which guarantees repayment: rather this study finds that more frequent group meetings leads to increases in social capital which improve risk sharing, decreasing default .
–     Researchers should investigate whether other development programs are also successful at boosting social capital and whether increases in social capital led to economically significant outcomes.