(Illustration by Adam McCauley)
Kiva.org has pioneered the use of online crowdfunding to support small no-interest loans to low-income entrepreneurs and students around the world to alleviate poverty. Its platform directly connects individual lenders to specific borrowers through either mobile-payment systems or Kiva partners on the ground. But like many such online ventures, it faces a participation problem: Many Kiva members sign up only to never lend or to make just one loan and then never return.
But the lending-teams program Kiva created to boost participation may solve this problem. In the past eight years, 38,957 teams have been created based on such member affiliations as religion, organization, and geography. If team participation increases lending, then if Kiva got more members to join teams, it could spark more lending.
With this question in mind, Yan Chen, a professor at the University of Michigan’s School of Information, and colleagues studied whether they could find a method to drive greater member participation in Kiva’s teams and whether team participation increases lending. An experimental economist working on social-identity theory, Chen understood that people join teams because they identify with the group’s members, characteristics, and aims. The challenge was how to prod Kiva’s members to join teams that would motivate them to give.
Chen and her co-researchers, computer scientists working on machine learning and recommender systems, decided to combine social-identity theory with the predictive accuracy of machine learning. They sent e-mails personalized by recommender algorithms that suggested teams to join—based on geographical proximity to the team’s lead lender and history of lending to the same borrowers—to a select group of the nearly 65,000 Kiva members who had yet to join a team. They found that geographic proximity and the status of the team in Kiva lending rankings made recipients more likely to join teams, while other factors had no positive effect.
Kiva members benefit from joining teams, and not simply from social-identity affirmation. “The cost for a prospective lender of searching for an appropriate borrower is significant,” Chen says. “The information sharing available through a team, which enables coordination, and the inter-team competition, which encourages goal setting, are useful mechanisms.”
Teams are helpful not only to Kiva’s individual members as information sources but also to the whole organization as social networks. “Crowdfunding success is most affected by the desirability of a crowdfunding appeal and the ability of the crowdfunder to circulate that appeal to a large number of people,” says Thomas Allison, assistant professor at Washington State University. “The ability to circulate the appeal to that many people is generally beyond the prosocial entrepreneur’s control. Given this, teams might be a big factor in the success of prosocial crowdfunding campaigns such as Kiva’s.”
The researchers also found that team membership did boost lending. Kiva members who opened the personalized e-mails and joined a team increased their loans more than 15 times the median $25 lifetime contribution of Kiva lenders, to $392. This result, however, evaporated after one week—perhaps because lenders waited for money to be repaid before relending. Chen is analyzing a two-year follow-up to see if team members’ active lending resumed after their initial loans were repaid.
Relatively few prior crowdfunding and microfinance studies have used experimental methods as Chen and her colleagues did, Allison points out: “This field experiment is valuable in developing better evidence of causality, versus mere correlation, in what influences funders to provide money to crowdfunding campaigns.” Online and offline social influence may play an important role in the success of crowdfunding in general, he suggests.
Wei Ai, Roy Chen, Yan Chen, Qiaozhu Mei, and Webb Phillips, “Recommending teams promotes prosocial lending in online microfinace,” Proceedings of the National Academy of Sciences, 113(52), 2016.
Read more stories by Marilyn Harris.