A Randomized Evaluation of an Agent Intermediated Loan

This study evaluates a more flexible system of lending that involves individual rather than group liability, exploring variations where local intermediaries like shopkeepers, traders and lenders function as commissioned loan agents.

Background
Microfinance in India has been successful on several fronts such as improving access to finance for the poor, empowering women or aiding in smoothing consumption, but it has failed to deliver in boosting entrepreneurship, increasing productivity and reducing poverty.
While there is significant demand for credit among the rural poor for diversification into high-risk, high-value cash crops, MFIs do not successfully meet this need. Some reasons for this have been found to be time-consuming and inflexible repayment schedules that do not account for cropping cycles and the active discouragement of risk-taking ventures by loan agents and other members in loan groups. Interest rates charged by MFIs are often even higher than informal lenders and far too high for cash crops to be adequately profitable.

This study tests out a more flexible system of lending that involves individual rather than group liability, 4-month durations that synchronize with cropping cycles, interest rates lower than the informal lending market and does not require collateral, savings, group meetings or intensive monitoring. It explores variations of the RBI-recommended Business Correspondent/Business Facilitator (BC/BF) Model where local intermediaries like shopkeepers, traders and lenders function as commissioned loan agents.

Approach

In randomly selected villages in West Bengal, India, the researchers implemented trader-agent-intermediated lending (TRAIL), in which local trader-lender agents were incentivized through repayment-based commissions to select borrowers for individual liability loans. In other randomly selected villages,  a group-based lending (GBL) program was organised in which individuals formed 5-member groups and received joint liability loans. The randomized intervention was conducted in 72 villages in 2 districts (Hugli and West Medinipur) of West Bengal in India. The main credit intervention involved providing agricultural loans, with repayment due in 4 months (120 days). 

Key Findings

TRAIL loans increased the production of the leading cash crop by 27% and farm incomes by 22%. GBL loans had insignificant effects. The researchers developed and tested a theoretical model of borrower selection and incentives. Farmers selected by the TRAIL agents were more able than those who self-selected into the GBL scheme; this pattern of selection explains at least 30–40% of the observed difference in income impacts.

Implications

The results suggest that TRAIL is effective (TRAIL agents recommend safe clients and there is no evidence of collusion); confirms predictions that: TRAIL agents select households with intermediate landholdings, while GBL selection is biased in favor of low landholdings; repayment rates are higher in TRAIL as are takeup rates, though the differences are not statistically significant in this case. Additionally the study finds find that the agent intermediated lending models (both TRAIL and GRAIL) are working well
at least in terms of the conventional MFI metrics of takeup and repayment rates. If anything they are doing better than GBL. Comparing TRAIL and GBL in terms of targeting is hard, because GBL is more pro-poor (more likely to select landless households) but TRAIL and GBL both appear to be able to target safe borrowers.

Thematic area

Financial Inclusion

Project Leads

Dilip Mookherjee, Pushkar Maitra, Sandip Mitra, Alberto Motta, Sujata Visaria

Location

West Bengal

Partners

Shree Sanchari

Status

Completed