Connecting Rural Unserved Communities to Digital Finance

Last registered on August 10, 2023
View Trial History

Pre-Trial

Trial Information

General Information

Title
Connecting Rural Unserved Communities to Digital Finance
RCT ID
AEARCTR-0011825
Initial registration date
August 02, 2023

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
August 10, 2023, 1:19 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name
Francis Annan
Affiliation
University of California, Berkeley
Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status
On going
Start date
2022-06-01
End date
2024-06-30
Keywords
Finance & Microfinance, Firms & Productivity
Additional Keywords
Market Design, Retail DFS, Service Quality, VSLAs-as-Agents
JEL code(s)
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Connecting rural unserved, virgin communities with latent demand for retail digital financial services (DFS) is an important first order policy and commercial question. Perhaps, more important and distinct is our ability to improve the new markets once communities are connected. This project addresses major gaps in research about connecting rural, unserved communities to markets for retail DFS with (i) original data collection — on household, community, and business outcomes — and (ii) a multi-year community-level field experiment testing a scalable approach to retail agent expansions, VSLAs-as-agents. We follow the success of our 2019/2020 detailed pilot work to launch a full RCT in rural Sierra Leone—a market environment that is poorly studied—to answer the following two research questions:

1) RQ1. What are the broader, general equilibrium impacts of connecting unserved, virgin communities with retail DFS?
2) RQ2. Once connected, could letting consumers evaluate their retailers improve the quality of services, consumer usage, and business outcomes, and if so, how?

In a unique collaboration with Catholic Agency for Overseas Development (CAFOD), Orange Money, GT Bank, and Innovations for Poverty Action (IPA) - Sierra Leone, we implement a large-scale community-level field experiment that encourage Village Savings and Loan Associations (VSLAs) to set up and operate as retail DF agents across rural unserved communities in Sierra Leone — a market environment that is poorly studied. Agents are established to offer both mobile money (via Orange Money) and bank services (via GT Bank) in 3 separate phases that span 2022-2024. This is an environment where the penetration of DFS is extremely low and retail agents are “non-existent” in these rural communities, but with latent demand and supply for DFS.
External Link(s)

Registration Citation

Citation
Annan, Francis. 2023. "Connecting Rural Unserved Communities to Digital Finance." AEA RCT Registry. August 10. https://doi.org/10.1257/rct.11825-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
The study will include two interventions meant to connect unserved communities to retail agents. First, to examine the broader impacts of retail agent expansions, we encourage and onboard existing VSLAs as agents (VSLAs-as- agents) in randomly selected communities (Treatment I). Second, to test improvements in quality of services and consumer usage of DFS, we additionally implement a simple, anonymized feedback mechanism that provides monthly consumer feedback—based on digitized new user reviews or evaluations—to retailers and the community in a random set of connected communities (Treatment II).
Intervention Start Date
2022-06-01
Intervention End Date
2024-06-30

Primary Outcomes

Primary Outcomes (end points)
For RQ1, we will compare the various Treatments versus Control.
o Key outcomes:
(i) take-up (0/1 indicator for sustained presence of retail agent in community);
(ii) % of people that use financial/DF services (access);
(iii) % of people that engage in money management, such as savings, budgeting, and loan management (utilization and/or knowledge);
(iv) enterprise and VSLA groups sales revenue, number of customers, labor supply (operational hours), business income, assets, and business expenses (enterprise development);
(v) services/DFS quality;
(vi) women/adolescents’ empowerment and agency, including downstream outcomes;
(vii) household expenses/consumption; (viii) shocks mitigation;
(ix) poverty at the community level;
(x) local multiplier (and velocity of money)

For RQ2, we will compare Treatments 1 versus 2.
o Key outcomes:
(i) services/DFS quality (prices, transparency, reliability, customer service);
(ii) consumers usage of DFS/reported happiness/well-being/perceptions;
(iii) business income/outcomes.

These broader impacts allow us to measure general equilibrium effects: (i) direct effects (businesses /VSLAs) and (ii) indirect effects (households e.g., consumption expenditure, shocks mitigation; vs community e.g., poverty, local multiplier; vs commercial providers e.g., revenues).
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We explore two experimental variations: (i) randomized entry of VSLAs-as-retail agents (Treatment I) and (ii) randomized entry of VSLAs-as-retail agents, combined with a consumer feedback mechanism (Treatment II).

*Control communities (Status Quo): No introduction of DF-retail agents or consumer feedback mechanism. (n=50 communities/VSLAs);

*Treatment I: To examine the broader impacts of retail agent expansions, we encourage existing VLSAs in select communities to enrol and operate as retail agents for both Orange Money and GT Bank only. (n=75 communities/VSLAs);

*Treatment II: To test improvements in quality of services and consumer usage in DFS, we additionally implement a simple, anonymized feedback mechanism that provides consumer feedback to enrolled retailers and connected communities. (n=75 communities/VSLAs).
Experimental Design Details
Not available
Randomization Method
Computer software and simple lotteries, while ensuring balance on baseline characteristics (Bruhn and McKenzie 2009), using baseline information gathered on the communities and VSLAs by our partner CAFOD.
Randomization Unit
Randomization of treatments is at the community-level, stratified based on population of community (high vs low) and number of existing businesses in village (high vs low), yielding 4 unique strata, and all misfits (if any) resolved and randomly assigned.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
200 communities
Sample size: planned number of observations
*200 communities x 15 households / community members (respondent: hh head or customer) = 3,000 hh surveys; *200 communities x 1 VSLA (respondent: president or treasurer) = 200 businesses; *Customer reviews surveys: 75 review communities x 30 DF customers each x 3 rounds = 6,750 review surveys; *Administrative and/or transaction data from providers and CAFOD’s agent monitoring tools - 150 (treatment) communities
Sample size (or number of clusters) by treatment arms
Control (Status Quo): n=50 communities/VSLAs;
Treatment I: n=75 communities/VSLAs;
Treatment II: n=75 communities/VSLAs.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
NA
IRB Approval Date
2023-08-15
IRB Approval Number
N/A
Analysis Plan