NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Loan Officer Incentives in Microfinance Institutions
Initial registration date
November 06, 2019
January 30, 2020 4:29 PM EST
This section is unavailable to the public. Use the button below
to request access to this information.
Harvard Business School
Other Primary Investigator(s)
Harvard Business School
Additional Trial Information
In this study we seek to understand the extent to which incentive structures within a microfinance organization may hinder opportunities for clients to graduate to better loan products. In the first stage, we began by looking at the incentives structures in place that we suspected could affect whether loan officers recommended their borrowers for the individual loan product. In this context, recommendations enable group loan borrowers to be eligible for the individual loan product, which allows higher loan amounts and encourages high return investments, and without a recommendation it is highly unlikely that a client can receive an individual loan. Therefore, in order to understand the effects of incentives, we collected recommendations and studied any effects on the number and quality of recommendations after changing loan office incentive structures. We first implemented a new structure meant to mitigate identified disincentives. The second modification involved rewards for recommendations that led to loans with good repayment behavior.
In the second stage we will pool subjects who were recommended with those that were not and randomize who receives an individual loan. We will then study the impact of this loan on business indicators and compare the results between those in the recommended group and those not in order to assess the quality of recommendations made and, moreover, the value of the recommendation itself. Registration Citation
Stage 1 of this project aims to evaluate two incentive schemes:
1. The effect of removing the disincentives created by FE’s current incentive scheme on the quality of recommendations (the Mitigation Scheme)
2. The effect of a new points scheme that we designed along with FE on the quality of recommendations (the Points Scheme)
The research team first collected LO recommendations at baseline. Months later Fondo Esperanza implemented the Mitigation scheme and the team collected recommendations once again. One month after that, FE implemented the Points Scheme and the research team collected recommendations again. Once all rounds of collecting recommendations were completed, the research team pooled total recommendations to create a dataset of all partners against an indicator of whether they were recommended or not. This leads into stage 2. In stage 2, the research team will randomly select ~40k partners from the list of all partners. Each one of these partners will be in either the "recommended" or the "not recommended." FE will pass this list through a set of internal screening processes and filter partners who are eligible to receive the individual loan. The next step in the screening process involves calling the partners who have made it through this first screening. This call, known as the pre-evaluation call, includes a set of questions whose answers further determine eligibility as well as a confirmation of whether the partner wants an individual loan. Of all the partners who pass the pre-evaluation call, we will then randomly select half the partners from this list (treatment group) to move onto the next stage of Fondo Esperanza’s usual evaluation process called the TER, an in-person evaluation performed by a Fondo Esperanza individual credit advisor. Those who pass the TER will receive an individual loan. Meanwhile,the other half (control group) will continue onto the TER stage a year from when the treatment individual loans are disbursed.
The research team will then compare business and borrowing outcomes for recommended and non-recommended partners in treatment and control to assess the impact of individual loans.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
repayment behavior, changes in income and business values, investment decisions, retention
Primary Outcomes (explanation)
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
This study has two stages: (1) changing loan officer incentive schemes in order to observe changes in the quantity and quality of recommendations and (2) randomizing who out the pool of interested and intiially eligible borrowers continues on to the in-field credit evaluation that determines whether they can get an individual loan or not.
Stage 1: randomized which loan officers we asked for recommendations in each stage (1) baseline (2) mitigation and (3) recognition/points scheme Stage 2: borrowers who pass certain filters that can be observed from the admin data were contacted in order to gauge interest in the individual credit product and assess if they pass the next set of filters. Those who pass are then randomized, at the individual level, into treatment or control. The treatment group is passed on to the individual credit loan officer (LO) who corresponds to the borrower's geography and the LO continues with the evaluation. Those who pass are given individual credits. Those in the control group will instead continue in their communal banks and be offered the chance to be evaluated in a future.
Experimental Design Details
randomization done in office by a computer
At the recommendation stage we randomized on the loan officer level
At the individual credit stage subjects will be randomize on the individual borrower level. The team plans to stratify on branch, number of group members a loan officer had, loan officer, and age of the group .
Was the treatment clustered?
Sample size: planned number of clusters
At the recommendation stages: 240 loan officers by 20 groups per officer = 4,800 communal banks
Sample size: planned number of observations
At the recommendation stage: 80,000 borrowers
At the individual credit stage: ~2000 borrowers
Sample size (or number of clusters) by treatment arms
At the recommendation stage: 123 loan officers in treatment and 118 loan officers in control
At the individual credit stage: 1000 borrowers in treatment and 1000 borrowers in control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
INSTITUTIONAL REVIEW BOARDS (IRBs)
Harvard University-Area IRB
IRB Approval Date
IRB Approval Number