Impacts of AI-based Microfinance Lending on Microentrepreneurs in Colombia

Last registered on September 25, 2026

View Trial History

Pre-Trial

Trial Information

General Information

Title

Impacts of AI-based Microfinance Lending on Microentrepreneurs in Colombia

RCT ID

AEARCTR-0019726

Initial registration date

September 17, 2026

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

September 25, 2026, 9:44 AM 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

Suzanne Duryea

Affiliation

Inter-American Development Bank

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Yuri Suarez

PI Affiliation

Inter-American Development Bank

Contact Investigator

PI Name

Gabriela Aparicio

PI Affiliation

Inter-American Development Bank

Contact Investigator

PI Name

Jaime Millán-Quijano

PI Affiliation

University of Barcelona

Contact Investigator

Additional Trial Information

Status

In development

Start date

2026-09-21

End date

2028-03-15

Keywords

Finance & Microfinance

Additional Keywords

Financial inclusion, Artificial intelligence, Credit scoring, Microenterprise, Gender bias, Randomized controlled trial, Microfinance

JEL code(s)

C93, D14, G21, J16, O16

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Microbusinesses face severe constraints in access to formal credit, even when they exhibit adequate repayment capacity. These constraints arise from both supply-side factors—such as difficulties in assessing credit risk—and demand-side factors, including self-exclusion driven by fear of rejection or misaligned financial products. Given that a large share of microbusinesses are led by women, gender-based barriers further exacerbate credit access gaps.
This study evaluates the impact of credit allocation using an artificial intelligence (AI)–based credit scoring model on microbusiness outcomes in Colombia. The intervention leverages alternative data and debiasing techniques that explicitly exclude gender and race from credit decisions. Using a randomized controlled trial with delayed access to credit, the study examines the effects of credit access on formal financial inclusion, repayment behavior, and proxies of economic activity. The analysis further explores heterogeneous impacts by credit score and by gender, assessing whether AI-based models can both expand access to credit and mitigate gender disparities without increasing credit risk.

External Link(s)

Registration Citation

Citation

Aparicio, Gabriela et al. 2026. "Impacts of AI-based Microfinance Lending on Microentrepreneurs in Colombia." AEA RCT Registry. September 25. https://doi.org/10.1257/rct.19726-1.0

Sponsors & Partners

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

Request Information

Experimental Details

Interventions

Intervention(s)

The intervention evaluates access to credit granted through an AI-based credit scoring model that relies on alternative data sources and incorporates debiasing techniques to eliminate gender and race as decision variables.
Among eligible applicants, individuals are randomly assigned to one of two groups:
Treatment group: receives a credit offer following Quipu’s standard approval timeline.
Control group: receives a notification indicating that credit cannot be approved at that time but is invited to reapply after a specified period.

Applicants who score above the minimum AI-based scoring model are not eligible for credit and thus comprise a second control group for the study.

After follow-up data collection is completed (between 6 and 18 months after application), individuals in the control groups are proactively contacted and offered access to credit. The intervention therefore delays—but does not deny—access to the financial product.

Intervention Start Date

2026-09-21

Intervention End Date

2027-10-31

Primary Outcomes

Primary Outcomes (end points)

* Access to formal credit: whether the applicant receives a loan and subsequently accesses additional formal credit products, as measured through Quipu administrative records and credit bureau data.
* Credit performance and repayment behavior: indicators of delinquency, arrears, and default, capturing whether AI-based credit allocation expands access without increasing repayment risk.
* Economic activity of the microbusiness: proxied by estimated revenues and transaction volumes derived from digital financial data (e.g., electronic wallets and open banking sources), reflecting changes in business performance attributable to credit access.

Primary Outcomes (explanation)

The primary outcomes are designed to capture the causal impact of access to credit allocated through an AI-based credit scoring model on microbusiness performance and financial inclusion. These outcomes directly reflect the core research questions of the experiment and are measured using high-frequency administrative data complemented by credit bureau information.
These outcomes are measured at baseline and during follow-up periods (approximately 6 and 12 months after the initial credit application) and constitute the primary endpoints for evaluating the effectiveness and sustainability of AI-driven credit provision for financially excluded microbusinesses.

Secondary Outcomes

Secondary Outcomes (end points)

Impacts by Gender

Secondary Outcomes (explanation)

The impacts of credit access differ by gender, and AI-based credit models that exclude gender-related information reduce gender gaps in access to credit and related outcomes.

Experimental Design

The study population consists of individuals who apply for a loan through Quipu, a fintech platform that provides credit to microbusinesses with limited or no formal credit history. Quipu has issued credit to approximately 7,000 microbusinesses since 2022 and operates nationwide in Colombia. Credit applications are submitted via WhatsApp, and approved applicants typically receive funds within a few days.
Eligible participants are applicants who obtain an approved credit score according to Quipu’s AI-based credit model.
The target sample size is approximately 1,500 applicants, with 750 assigned to the treatment group and 750 to the control group.

Experimental Design Details

Not available

Randomization Method

Randomization by a computer

Randomization Unit

Among eligible applicants, individuals are randomly assigned to one of two groups:
Treatment group: receives a credit offer following Quipu’s standard approval timeline.
Control group: receives a notification indicating that credit cannot be approved at that time but is invited to reapply after a specified period.

Was the treatment clustered?