Beyond the FICO Score: How Alternative Data Is Reshaping Credit Underwriting

For decades, access to credit in the United States has hinged largely on a three-digit number built from data held by the major credit bureaus: payment history on credit cards, loans, and mortgages. But an estimated large share of U.S. adults are considered "credit invisible" or have files too thin to generate a reliable score, according to research the Consumer Financial Protection Bureau has cited in its own policy work. That gap has pushed banks and fintech lenders toward alternative data, a broad term for financial information that sits outside the traditional credit report.

What Counts as Alternative Data

Alternative data generally includes any information not part of a standard credit bureau file that a lender might use to estimate a borrower's ability and willingness to repay. Common categories include bank account cash-flow data such as deposit regularity and account balances, rent payment history, utility and telecom bill payments, and in some cases subscription or payroll data accessed through connected accounts. Cash-flow underwriting, in particular, has gained traction because checking and savings account activity can reveal income stability and spending discipline even when a consumer has never had a credit card or loan.

Why Lenders and Regulators Are Interested

In December 2019, the CFPB joined the Federal Reserve, FDIC, OCC, and National Credit Union Administration in issuing an interagency statement encouraging responsible use of alternative data, noting it could expand access to credit and potentially lower costs for consumers whose files are too sparse for conventional scoring models. The agencies were careful to note that use of such data must still comply with existing consumer protection laws, not that it is automatically safe or neutral.

One concrete example of alternative data entering the mainstream mortgage market is on-time rent payment history. Fannie Mae has outlined a program allowing qualifying borrowers' rent payment histories to be factored into certain underwriting assessments, an acknowledgment that a strong rent-payment record can be a meaningful substitute for, or supplement to, traditional tradelines.

The Legal Guardrails That Still Apply

• The Equal Credit Opportunity Act (ECOA) and its implementing Regulation B prohibit credit discrimination based on protected characteristics such as race, color, religion, national origin, sex, marital status, or age, regardless of what data or model produces the decision.
• Lenders must still provide adverse action notices when denying credit or taking other unfavorable actions, explaining the specific, principal reasons for the decision.
• In its Circular 2022-03, the CFPB stated that creditors using complex algorithms or AI-based models, including those built on alternative data, cannot use the complexity of the model as an excuse to provide vague or boilerplate adverse action reasons; specific and accurate reasons are still required.
• The Fair Credit Reporting Act governs certain alternative data providers if the data is used to make eligibility decisions and the provider meets the definition of a consumer reporting agency.

In practice, this means a fintech lender cannot simply plug raw bank transaction data or rent history into a black-box machine learning model and treat the output as exempt from fair lending law. Regulators have signaled they expect lenders to be able to explain which data points drove a denial, test models for disparate impact on protected groups, and retain documentation sufficient to demonstrate compliance.

Benefits and Risks for Consumers

For a consumer with limited credit history, such as a recent immigrant, a young adult, or someone who primarily uses debit cards and cash, alternative data can open doors that a traditional score would keep closed by demonstrating consistent income and bill-paying behavior. On the other hand, the same data can cut the other way. Irregular income, overdrafts, or inconsistent rent payments could count against an applicant even if that same person would have had no credit score at all under a conventional model, meaning there was previously nothing negative on file.

There are also data-access and privacy considerations. Many alternative data underwriting products rely on consumers connecting bank accounts through data aggregation services, which involves sharing transaction-level account information with third parties. The CFPB's ongoing work on personal financial data rights, grounded in Section 1033 of the Dodd-Frank Act, is directly relevant to how consumers can control, access, and revoke that kind of data sharing.

The Bottom Line

Alternative data underwriting is neither inherently predatory nor automatically a credit-access panacea. It is a tool that, used carefully and in compliance with fair lending law, can extend credit to people the traditional system overlooks. Used carelessly, it can simply encode old biases into new data streams. As more banks, buy-now-pay-later providers, and fintech lenders adopt cash-flow and rent-based underwriting, the practical question for consumers is less about whether alternative data is being used and more about whether they understand what data was used, how it affected the decision, and what rights they have to contest it.