Optimizing network referrals to identify and recruit creditworthy entrepreneurs

Recent studies have shown that while microfinance initiatives (MFIs) encourages business creation, their overall impact on business profits is modest at best. Microfinance institutions interested in giving out Small and Medium Enterprise (SME) loans often have sub-optimal recruitment strategies, leading to low volume of applications and a lower quality of applications. Thus, SME loans generate both high risk and high lending costs. Many MFIs use word-of-mouth referrals to evaluate applications. As more central individuals are both better connected and have more information about others’ creditworthiness and entrepreneurial capacities, targeting central individuals in a referral process can potentially generate gains both in terms of volume of applicants and identifying those entrepreneurs with higher potential growth.

The researchers will examine if referrals work in the context of SMEs, and why they work. Their intervention is based on social network theory about how information is aggregated at a community level. Surveying several hundred villages, they will target businesses and ‘central’ individuals to analyse the impact of focusing on such individuals when seeking referrals to MFIs; through analysing the change in volume and quality of applications due to referrals from these individuals. The researchers will be operating in partnership with a local MFI operating in the Indian State of Uttar Pradesh, which will be providing administrative data to verify repayment schedules and profiles of good clients. Information gathered will further be used to provide the MFI with a tool to predict loan default, predicated on the MFI not having to collect more data than it already does.

A broader discussion in development economics has considered community driven development: the idea that the community may contain more information than the policymaker, and this can be important in screening and targeting. The latter immediately falls into the domain of network economics: which members of a community have the best information and how can it be tapped into for policy? This project tackle these issues head-on.


Arun Chandrasekhar

Stanford University

Francisco Munoz

Stanford University