Digital Credit Reporting And Women's Economic Empowerment

Kenya is the world's most celebrated mobile money success story. With 98% of adults now holding mobile money subscriptions — 51.4 million accounts, processing over 61 million M-Pesa transactions daily, and channelling Sh8.66 trillion in person-to-person transfers in 2025 alone — the country's digital financial infrastructure has done what decades of formal banking could not: it has reached nearly everyone. The gender gap in formal financial access, which stood at a yawning 12.7 percentage points in 2006, has collapsed to a mere 1.6% in 2024. By the standard headline metrics, Kenya has solved women's financial inclusion.

It hasn't.

Beneath the account-access numbers lies a more troubling divergence. The 2024 FinAccess Household Survey — Kenya's most authoritative gauge of financial behaviour — reveals that while the access gender gap has almost closed, the financial health gender gap has actually widened to 7.5%. Women are inside the digital financial system. They are not thriving in it. Only 46.5% of women hold formal banking products, compared to 58.9% of men. A forensic audit of ten credit-scoring algorithms operating across Nigeria, Kenya, and South Africa found that female entrepreneurs are systematically assessed with a 37% underfunding penalty — not because they are riskier borrowers, but because the underlying data architecture of digital lending was never designed with women's economic lives in mind.

This is the question our ongoing quasi-experimental research confronts directly: did the Central Bank of Kenya's 2022 Digital Credit Providers Regulations — the most consequential regulatory intervention in the sector's history — structurally improve women's credit outcomes? Or did they formalise a system that was already structurally biased, licensing that bias into compliant, regulated operations?

Our preliminary findings suggest the answer is simultaneously both and neither. The 2022 reforms measurably reduced predatory extraction and improved pricing transparency. But they did not — and by design could not — address the algorithmic data architecture that continues to encode historical gender penalties into credit decisions at scale.

The Access-to-Health Paradox: Having an Account Is Not the Same as Building Wealth

Kenya's financial inclusion story, retold through the 2024 FinAccess data, is a study in two diverging trends. Headline access has never been better: the percentage of women in the most advanced digital finance user segment surged 62% between 2021 and 2024, rising from 3.5 million to 5.7 million women. Mobile money is near-universal. The gender gap in account ownership is functionally closed.

But access and agency are different variables entirely. Financial health — the ability to absorb shocks, accumulate assets, and access growth capital — continues to diverge sharply by gender. The financial health gender gap reached 7.5% in 2024. Female financial exclusion, at 10%, marginally exceeds male exclusion at 9.8%. More critically, the formal credit gap remains cavernous: women-owned and women-led enterprises across sub-Saharan Africa collectively require an estimated $42 billion more in capital annually — and even when their repayment records are equivalent to those of men, African women typically access just 7% of formal credit.

Regulators and development finance institutions that measure success through account penetration rates are measuring the wrong thing. The CBK, FSD Kenya, and the broader development community must shift the primary benchmark from access to financial health outcomes disaggregated by gender. Access is the doorway. Credit for productive investment is the room where economic empowerment actually occurs.

Algorithmic Credit Scoring Penalizes Female-Led Micro-Enterprises Through Proxies, Not Risk

The mechanisms of exclusion have evolved. The old barriers were explicit: collateral requirements tied to land title deeds that women were legally or culturally prevented from holding. The new barriers are statistical and invisible — embedded in the very algorithms that fintechs present as objective and democratising.

An audit of ten credit-scoring models across African fintech platforms exposed the anatomy of this new bias. Beauty and personal-care services — sectors dominated by women-led micro-enterprises — are systematically classified as high-risk, despite documented profitability and low default rates. Some scoring models evaluate creditworthiness through peer-network data; in a society where women's professional and financial networks are structurally smaller, this amplifies existing inequality. Natural language processing applied to loan applications assigns lower scores to cooperative or household-framing language — language more commonly used by women.

The compounding effect is a 37% underfunding penalty against female-led SMEs with identical financial fundamentals to their male counterparts.

Every fintech operating a credit scoring model in Kenya should be required — and should actively want — to conduct algorithmic bias audits disaggregated by gender, sector, and geography. Correcting for algorithmic bias expands the addressable credit market and reduces non-performing loan risk by extending credit to the correctly-assessed low-risk borrowers currently being rejected.

Measuring What Actually Changed: The Quasi-Experimental Logic Behind Our Analysis

The CBK's Digital Credit Providers Regulations 2022 (L.N. No. 46, gazetted March 18, 2022) required all previously unregulated mobile lending applications to obtain formal CBK licences within six months. This created a sharp, externally imposed policy shock — a before and after — that was not driven by changes in borrower behaviour or market conditions.

We apply a difference-in-differences framework, comparing credit outcomes for borrowers using regulated digital credit platforms against borrowers using informal SACCO credit and mobile-banking products that were already subject to supervision before 2022. Any difference in outcome trends between the groups can be attributed, with reasonable confidence, to the 2022 intervention itself rather than to broader market forces.

Critically, we disaggregate this analysis by gender. The question is not simply whether the regulation improved the market overall, but whether it improved it differently — or less — for women. Heterogeneous treatment effects of this kind are rarely examined in policy evaluations; their absence from the literature has allowed policymakers to declare gender-neutral interventions a success without verifying that neutrality in the data.

The 2022 Reforms Cleaned Up the Market's Conduct — but Left Its Architecture Intact

Predatory conduct has declined markedly. The regulatory prohibitions on debt-shaming, on undisclosed fees, and on negative CRB reporting without borrower notification have meaningfully shifted borrower experience. Pricing transparency has improved.

But the regulation did not — and was not designed to — address the credit-scoring methodologies that determine who receives a loan in the first place. It regulated the conduct of lending. It did not regulate the criteria of lending. And it is in those criteria that gender bias lives.

Post-2022, the gender gap in consumer-protection outcomes narrowed substantially. But the gender gap in credit access and loan size did not close. Women who would have been rejected before 2022 continue to be rejected after it — for the same algorithmic reasons, now administered by licensed, compliant institutions.

The CBK's next supervisory cycle must extend beyond conduct regulation to encompass model governance. CBK should require Digital Credit Providers to submit gender-disaggregated approval rates, loan-size distributions, and interest rate spreads as part of annual licence renewal.

Intra-Household Bargaining Is a Hidden Variable That Corrupts Individual Credit Scores

Women's mobile money usage patterns — the primary data source for digital credit scoring — are frequently shaped not by their own financial preferences but by the power dynamics of the household. In households where male partners hold greater decision-making authority, women's digital transaction patterns often undercount their actual economic activity. Savings are deliberately made invisible — held in informal arrangements or physical cash — precisely because they serve as a buffer against appropriation by male household members.

A woman who maintains a mobile money float substantially lower than her actual savings will generate a transaction record that understates her financial capacity. The algorithm scores her as poorer and more risky than she actually is. The household power asymmetry becomes a credit score penalty.

Fintechs serious about women's credit inclusion must invest in product architectures that account for household power dynamics: privacy-protecting savings features, scoring models that capture informal economic activity beyond mobile money records, and loan products designed in consultation with women borrowers about their actual financial lives.

Gender-Aware Credit Infrastructure Is the Regulatory Frontier — and the Market Opportunity

In February 2025, TransUnion Kenya partnered with FICO to launch an advanced credit-scoring model incorporating 145 data points and 24 months of behavioural data from telecoms and utilities. In June 2025, Creditinfo Kenya and Kamoa announced a strategic partnership to build comprehensive credit profiles using alternative data. And in November 2025, the Kenya Bankers Association and FSD Kenya launched a Gender-Disaggregated MSME Credit Dashboard, for the first time making visible the systematic differential between how male-led and female-led businesses access credit.

But infrastructure without governance is just a more sophisticated form of the same problem. The risk is that 145 data points, if calibrated on historically biased lending outcomes, will encode 145 channels of gender bias rather than one. More data is not better data unless the question being asked of it is right.

The Central Finding: Formal Inclusion Without Structural Equity Is a Different Kind of Failure

The 2024 FinAccess data is unambiguous: the financial health gender gap is widening even as the access gap closes. Women are arriving at the door of the formal credit market in record numbers, and a structural architecture is systematically turning them away or offering them less.

The headline finding, already robust enough to state with confidence, is this: regulatory reform that governs conduct without governing criteria will improve the experience of exclusion without ending it.

For fintechs, it is a market signal as much as a compliance mandate: the female-led micro-enterprise segment is systematically underserved relative to its actual creditworthiness, and the institutions that solve for that first will inherit a large, loyal, and low-risk borrower base.

The access gap is closed. The empowerment gap is the work that remains.

This article draws on ongoing primary research conducted under a quasi-experimental framework at NM Research and Advisory.