By Yan Chen
Micro and small enterprises together represent the largest employers in many developing countries. However, they often find their growth stifled by a lack of access to credit or other financial services. To address this issue, microfinance programs provide small loans and other financial services to encourage entrepreneurial development. Currently, about 10 million households worldwide are served by these microfinance programs, which help very poor households meet basic needs, improve household economic welfare and promote entrepreneurship (cgap.org). Nonetheless, this number is only a small fraction of those in need.
The Internet provides opportunities for ordinary people around the world to support entrepreneurial activities through the process of online microlending. One such opportunity is provided by Kiva. Founded in October 2005, Kiva is the world’s first peer-to-peer microfinance website. Specifically, it partners with microfinance institutions and matches individual lenders with low-income entrepreneurs in developing countries as well as selected cities in the United States.
Through Kiva’s platform, anyone can make a zero-interest loan of $25 or more to support an entrepreneur. Since its inception, Kiva has increased its membership significantly. As of March 18, 2013, more than 1.3 million lenders across 208 countries have contributed $400 million in loans, reaching over one million borrowers in more than 65 countries. Kiva has transformed the entrepreneur lending model, creating a new form of capital market enabled by information technology.
In our NSF-funded research project, we explore the structure and effectiveness of the Kiva model. In particular, we investigate the effect of team membership on pro-social lending. We do so through both an empirical analysis of naturally-occurring field data and a randomized field experiment which explores the effectiveness of goal setting and coordination within a lending team.
Using data collected through Kiva’s official API, we find that while 64% of lenders have made at least one loan, 36% have never made a single loan. This highlights the main issue that Kiva faces. The membership of Kiva, along with the number of loans made through Kiva, has increased greatly since Kiva’s inception. However, while a few lenders make many loans, many lenders make only a few or no loans. Indeed, Premal Shah, the president of Kiva, indicates that many Kiva lenders lend once, but then never come back to the site, despite the fact that their loans have been repaid and thus they could make another loan at no additional cost.
To combat this problem and increase lender engagement, in August 2008, Kiva instituted a lending teams program. Any lender is allowed to create or join any number of teams. When that lender next makes a loan, a prompt asks whether to assign that loan to one of the lender’s teams.
Once a team is created, it appears on Kiva’s team leaderboard (www.kiva.org/teams). This leaderboard sorts teams by the total amounts of loans designated by their team members. Since 2008, more than 26,000 Kiva teams have been created, many of which are organized based on lender group affiliations such as school, organization, religious affiliation, geographic location, or sports. Many of the highly ranked teams are identity-based, such as the “Atheists” and the “Kiva Christians.”
At times, a sense of competition seems to emerge across teams. Indeed, the captain of the Atheist team has stated that “[t]he whole idea of teams in the Kiva context implies there should be competition.” This competitive aspect of the Kiva teams is an intriguing area of study. A large body of experimental research in economics and social psychology, conducted almost exclusively in the laboratory, has demonstrated that the existence of teams can increase public goods provision and improve coordination. However, it is an open question whether team competition increases lending activity in a natural field setting, and if so, by how much and why.
Using field data, we find that, controlling for selection bias, lenders who join teams contribute an average of 1.2 more loans per month than those who do not join teams. Overall, our empirical analysis indicates that team competition can be an effective mechanism for increasing pro-social lending and public goods provision in an online community. Kiva instituted its team component as a way of motivating its lenders to come back and lend more. Our evidence suggests that the site has been successful in some ways.
Our analysis also shows substantial heterogeneity in team participation levels. Of the 26,000 teams, nearly half have not made a loan for the past year. To investigate what factors make some teams more effective in encouraging member lending behavior, we run a field experiment. In our experiment, we post messages to team forums (n=22,233 participants) and find that, compared to the control, lenders who are members of inactive teams make significantly more loans in a month when shown a message that sets a lending goal for their team.We also find that goal setting, combined with team coordination, is effective in increasing pro-social lending in inactive teams.
When creating online teams, Kiva developers assumed that team competition, in the form of a team leaderboard, might increase lending. Our results quantify the magnitude of the team competition effect on lending activities, suggesting some reasons for its effectiveness. These reasons may help guide Kiva’s further development of its system as well as provide helpful guidance to other online communities just beginning to create their systems. A better understanding of how team membership can enhance participation will help these communities to better fulfill their missions.
Authors of the study:
- Roy Chen: Assistant Professor of Economics, National University of Singapore, former STIET fellow
- Yan Chen: Professor, UMSI
- Yang Liu: doctoral student, UMSI
- Qiaozhu Mei: Assistant Professor, UMSI