Please fill out this short user survey of only 3 questions in order to help us improve the site. We appreciate your feedback!
An Agent-based Credit Delivery Model
Last registered on January 29, 2020


Trial Information
General Information
An Agent-based Credit Delivery Model
Initial registration date
January 29, 2020
Last updated
January 29, 2020 1:45 PM EST
Primary Investigator
Monash University
Other Primary Investigator(s)
PI Affiliation
Indian Statistical Institute
PI Affiliation
Boston University
PI Affiliation
Hong Kong University of Science and Technology
Additional Trial Information
Start date
End date
Secondary IDs
We designed and evaluated a randomized credit intervention where the task of borrower selection was delegated to local intermediaries. In one intervention, named Trader-Agent Intermediated Lending, a randomly selected local private trader was appointed as the agent. In the other intervention, named Gram Panchayat-Agent Intermediated Lending (GRAIL), the local village council selected the agent. Agents were incentivized through repayment-dependent commissions to select creditworthy clients and to enforce loan repayment. These two schemes were compared with a group-based lending (GBL) intervention where 5-member groups self-formed and received joint liability loans. We collected household survey data at 4-month intervals to evaluate the impact of these schemes on farm cultivation decisions, output and profits.
External Link(s)
Registration Citation
Maitra, Pushkar et al. 2020. "An Agent-based Credit Delivery Model." AEA RCT Registry. January 29. https://doi.org/10.1257/rct.5366-1.0.
Experimental Details
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
The outcomes of interest were: (i) cultivation decisions and input choices, output, revenue, costs and profits, especially for potatoes (ii) loan take-up, continuation and repayment rates.
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
We worked in 72 villages in two districts of West Bengal, India. Twenty-four villages were assigned to the TRAIL, GRAIL and GBL interventions each. In each TRAIL or GRAIL village, a single individual (either a trader or a political appointee) was appointed agent and asked to recommend 30 village residents who owned less than 1.5 acres of cultivable land. Ten of these 30 were randomly selected to receive the individual liability loans. The agent received as commission 75% of the interest paid by borrowers they had recommended. In the villages assigned to the GBL intervention, village residents owning less than 1.5 acres of land were invited to self-form 5-member groups. Of the groups that survived the 6-month initiation period, two were randomly selected to receive joint liability loans. All loans in all 3 interventions were offered at 6 percent interest for a 4-month duration, repayable in a single lumpsum. We conducted 4-monthly household surveys with 50 households in each village to assess the impacts on the following outcomes: (i) cultivation decisions and input choices, output, revenue, costs and profits, especially for potatoes (ii) loan take-up, continuation and repayment rates.
Experimental Design Details
Randomization Method
Villages were randomly selected by drawing randomly from a set of village councils in 2 districts of West Bengal. Villages were randomized into each of the treatment arms - randomization conducted in our office by a computer.

Individuals /groups to receive loan (treatment households) were selected through a public lottery in the presence of a village official.
Randomization Unit
1. Village
2. Household
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
72 villages
Sample size: planned number of observations
50 households in each village. Total sample size of 3600 households.
Sample size (or number of clusters) by treatment arms
24 villages in TRAIL treatment
24 villages in GRAIL treatment
24 villages in GBL treatment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
Monash University Human Research Ethics Committee
IRB Approval Date
IRB Approval Number
CF10/2368 – 2010001349
Post Trial Information
Study Withdrawal
Is the intervention completed?
Intervention Completion Date
August 01, 2013, 12:00 AM +00:00
Is data collection complete?
Data Collection Completion Date
August 01, 2013, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
72 villages
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
3600 households
Final Sample Size (or Number of Clusters) by Treatment Arms
1200 households in TRAIL Treatment arm 1200 households in GRAIL Treatment arm 1200 households in GBL Treatment arm
Data Publication
Data Publication
Is public data available?
Program Files
Program Files
Reports, Papers & Other Materials
Relevant Paper(s)