The Michigan Engineer News Center

How a team-based approach boosts charitable lending

Charitable lenders who belong to a team contribute significantly more than those who operate on their own, according to new studies of the microlending site| Short Read

Charitable lenders who belong to a team provide significantly more loans than those on their own, according to new studies led by economists and computer scientists at the University of Michigan.

The researchers say the findings have implications for charitable giving as well.

The project team, which also include researchers from the National University of Singapore and Kidaptive, studied the behavior of more than 60,000 members of the online lending community Kiva.

Kiva matches citizen lenders with low-income entrepreneurs in developing countries. Loans are zero-interest. Since 2008, the site has allowed members to create and join teams, and it aims to spark competition by posting team stats on a leaderboard.

The researchers say team membership appears not to simply correlate with greater contribution, but to outright cause it. The studies used randomized field experiments and other statistical techniques to show that the findings aren’t results of the most active lenders being more likely to join teams.

Charitable organizations typically focus appeals on individuals, but the researchers say their studies highlight the underexplored potential of team-based approaches.

The researchers examined Kiva members’ natural behavior and also performed two large-scale randomized field experiments involving emails and Kiva forum messages. They initially turned to the website to study what motivates people to give, but early in their investigations they noticed that team membership seemed to have a huge effect.

One study, to be published this month in a special issue of Games and Economic Behavior, reports that lenders who join teams contribute 1.2 times more loans per month than their solo counterparts. The study also determined that loaning increases when team leaders post messages on Kiva forums that: include a link to a chosen borrower, set a goal, and urge the team to work together to meet the goal. When a forum post included all those factors, teams produced an average of 11 more loans per month.

Another study, published in Proceedings of the National Academy of Sciences in December 2016, reported that Kiva members who weren’t on teams could be most effectively nudged to join them with email messages that suggest three teams to join, and also say why Kiva recommended these teams. Location matching worked best. Kiva used algorithms developed at U-M to match members with teams. This study found a spike in donations in the first week a lender joined a team. The average loan of a new team member was $392 more than someone who did not join a team, more than 15 times greater than the $25 lifetime contribution of the median Kiva member.

“The indications are that team membership is effective in increasing member contributions among lenders and that making recommendations to lenders to join teams is an effective and inexpensive mechanism to engage community members and increase their contributions,” said Yan Chen, the Daniel Kahneman Collegiate Professor of Information in the U-M School of Information and research professor at the U-M Institute for Social Research.

The research also has other implications in the field of economics. It demonstrates that teams can discourage free-riding and prod people toward so-called “pro-social” behavior – conduct that benefits the public good. While this had previously been demonstrated in a lab, it hadn’t been shown in the real world. Social scientists and economists are interested in understanding what strategies can increase pro-social behavior in situations where there isn’t a central authority.

“We see this as going beyond Kiva,” Chen said. “We’re now looking at Wikipedia, for example. While it has more than 1,000 editors coming on board every week, most of them don’t continue to contribute and keep editing. It boils down to the economic principle known as the free-rider problem in public goods provision. You want to encourage the overcontributors and somehow penalize or call out the undercontributors. We think this team structure can help accomplish that.”

In the December study, the researchers had Kiva send various emails to more than 64,000 members. Roughly 20,000 members opened the emails, which either: told the recipient that teams exist but said nothing more; suggested three teams for the recipient to join; or suggested teams based on location, lending history, or team status and included an explanation for that particular recommendation.

In the January study, the researchers set up 50 new lender profiles, gave $25 loans to nearly 500 teams, then posted on the team’s forum a note introducing themselves in four different ways: A simple note that they had just joined the team and made their first loan; a note that also included information about their chosen borrower and a link so others could give too; a note that urged others to give so the team could increase its ranking; or a note that included both a link to the chosen borrower and a call to action to improve the team’s ranking. The last category proved to be the most effective. Here’s an example post from the paper:

“Hi, I’m new to the team. I just credited my first loan to the team. I loaned to Hranush from Armenia. She requested a loan of $3,000 to help her purchase wheat to feed the livestock. Here is the url to her request…. If each of us gives a $25 loan in the next month, we will improve our rank.”

Forums often lit up after such posts, with lender after lender thanking the new member for calling their attention to the borrower and the rankings.

“We obtained these experimental results even though our messages were relatively short and neutral compared to actual member postings,” said Qiaozhu Mei, an associate professor of information and also of electrical engineering and computer science at U-M. “This suggests that the effects of real competition or coordination messages on lender activity may be even greater than those observed in the study.”

The researchers hypothesize that teams offer more than healthy competition. They also provide helpful coordination. Members can share loan requests that seem to be good fits for the team’s identity. This likely helps lenders narrow down who to give to among the site’s thousands of active borrowers.

The December paper is titled. “Recommending teams promotes prosocial lending in online microfinance.” The January paper is titled, “Does team competition increase pro-social lending? Evidence from online microfinance.” An online version was published last year, but it appears in print this month. The research was funded by the National Science Foundation.

Portrait of Nicole Casal Moore


Nicole Casal Moore
Media Relations & Research News Director

Michigan Engineering
Communications & Marketing

(734) 647-7087

3214 SI-North

  • Yan Chen

    Yan Chen

    Daniel Kahneman Collegiate Professor of Information

  • Qiaozhu Mei

    Qiaozhu Mei

    Associate Professor of Information

The outside of the Ford Robotics building

U-Michigan, Ford open world-class robotics complex

The facility will accelerate the future of advanced and more equitable robotics and mobility | Medium Read