Unpacking the Gender Gap in Small Business Microlending: Demand and Supply Experiments in Ghana
How to cite
Jung, Youngjoo; Ayesu, Enock; Kamau, Simon. 2026. "Unpacking the Gender Gap in Small Business Microlending: Demand and Supply Experiments in Ghana." AEA RCT Registry. AEARCTR-0017894. https://doi.org/10.1257/rct.17894-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
The gender gap in microfinance exists in both the extensive (rejection) and intensive margins (smaller loan), and smaller loan sizes for female entrepreneurs prevent them from scaling up their businesses. Female entrepreneurs often request and receive smaller loans, partly because they tend to run smaller businesses. However, if a gender gap in loan size persists even among firms with similar marginal returns to capital, it may suggest capital misallocation.
Our research aims to unpack the gender gap in microfinance loan sizes for small businesses in Ghana by examining both the demand and supply channels. Understanding the underlying mechanisms will inform future experiments designed to reduce this gap. To mitigate selection bias in analyzing gender differences, we focus on male and female entrepreneurs who co-own the same firm but are not family members. This within-firm comparison helps hold business characteristics constant. In addition, for the loan officer evaluation, we randomize the applicant’s gender to identify causal differences in approval decisions and loan terms.
We conduct a three-stage experiment for entrepreneurs and loan officers: (1) Need – Do female entrepreneurs require less money than males? (2) Ask – Do they need the same amount but request smaller loans? (3) Receive – Do they need and ask for the same amount but receive less? On the demand side, we elicit each entrepreneur’s requested loan amount and stated financing needs by exploiting within-firm variation by comparing the responses of male and female co-owners of the same enterprise.
On the supply side, we will randomize the applicant’s gender. For each firm profile, loan officers will evaluate four applications: a male- and female-named version of both the male and female co-owner. This design allows us to isolate the causal effect of applicant gender within the same firm. We will examine how loan officers’ evaluations vary by applicant gender along three margins: (i) the approval decision; (ii) the maximum loan amount granted if approved; and (iii) if rejected, the maximum loan amount they would be willing to approve instead.
Intervention
On the demand side, similar to a twin-study design, we compare male and female co-owners of the same business who are not family members. By holding firm characteristics constant, this within-firm comparison helps mitigate selection concerns. We elicit gender differences in financial information, perceived capital needs, and loan demand by asking each co-owner to respond separately to identical business scenarios. These scenarios vary by financing type (grant versus loan), amount, allowing us to isolate gender-based differences in stated needs and financial decision-making within the same firm context.
On the supply side, we randomize the applicant’s gender across loan officers. The same application profile is presented as male to some loan officers and as female to others. This design allows us to estimate how loan officers evaluate identical applications differently based on the applicant’s gender.
Experimental design
This study uses a within-firm (twin-style) design to examine gender differences in small business financing decisions on both the demand and supply sides. We recruit firms that are jointly owned by two non-family co-owners of different genders, allowing us to compare male and female entrepreneurs within the same enterprise while holding constant firm-level characteristics.
On the demand side, each co-owner independently responds to identical, standardized business scenarios. These scenarios vary by financing type (grant vs. loan), amount, and loan terms. We elicit detailed information on (i) financial conditions of the business, (ii) perceived capital needs, (iii) requested loan size, and (iv) intended use of funds. This design enables us to measure within-firm gender differences in financial perceptions and loan demand.
On the supply side, we randomize the applicant’s gender across loan officers. For each firm, loan officers evaluate four applications, allowing us to compare decisions for otherwise identical profiles that differ only in the applicant’s gender. We measure (i) whether the loan is approved or rejected; (ii) if approved, the maximum loan amount granted; and (iii) if rejected, the maximum loan amount they would be willing to approve instead.
Design details
Cited by
No citation data available for this registration yet.