The Institution Behind the Denials

Navy Federal Credit Union is not a distant Wall Street bank. It was built specifically for the people who serve in the United States military and their families. It now counts 13 million members and offers mortgage products including conventional purchase loans, rate-and-term refinances, cash-out refinances, VA-backed mortgages, and home equity lines of credit. It is, by any measure, the largest credit union in the country.

Credit unions are supposed to be different from commercial banks. They are member-owned. Their profits stay inside the organization rather than flowing to shareholders. The entire premise is that a credit union serves its community rather than extracting from it. Navy Federal’s community is people who took an oath to serve the country. Many of those people are Black and Latino veterans.

The case now working its way through federal court asks a straightforward question: Did Navy Federal build a system that systematically directed those same members toward denial while directing White members toward approval? The answer, according to the complaint filed in the Eastern District of Virginia, is yes. The mechanism, according to the complaint, is a proprietary algorithm that the company keeps secret.

The Algorithm Nobody Outside Navy Federal Can See

Every mortgage applicant at Navy Federal starts by filling out a Uniform Residential Loan Application, the same standardized form used across most of the mortgage industry. That form collects name, Social Security number, citizenship status, property address, employment history, income, assets, liabilities, military service history, and more. This information, along with other undisclosed inputs, feeds into Navy Federal’s underwriting system.

The complaint describes the process as “semi-automated.” Human loan officers play a role, but a proprietary algorithm drives a critical portion of the decision. Navy Federal has not disclosed what that algorithm actually measures, how it weights different factors, or at what point human discretion enters the process. The complaint argues that some of the data fields collected by the application form “can be proxies for race”: geography, employment type, asset types, and relationship history with the institution can all correlate with race in ways that embed historical discrimination into a formula that appears neutral on its surface.

This is the core of the legal theory. The plaintiffs are not alleging that Navy Federal executives sat in a room and said, “Deny the Black applicants.” They are alleging something more structurally dangerous: that Navy Federal built and continues to run a system that produces racially discriminatory outcomes while hiding the mechanism inside a proprietary black box that nobody outside the company is allowed to inspect.

The NCUA, the federal regulator that oversees credit unions, analyzed mortgage lending data across the credit union industry in 2020 and 2021. Its economists found that “the estimated denial likelihood was higher for minority applicants by about 1.2 to 1.9 times the rate of White borrowers.” The NCUA issued its own caveat: because the public mortgage dataset excludes credit scores, work history, and bankruptcy history, the disparity alone cannot prove discrimination. But the disparity is real and documented at the federal level.

CNN’s 2023 investigation went further. Reporters specifically analyzed Navy Federal’s data rather than the industry overall. They found the same gap, concentrated in one institution serving 13 million military families. CNN acknowledged the same limitation: the public dataset available to outside researchers is stripped of key credit-risk variables that Navy Federal itself holds and will not share. The company possesses the data that could exonerate or condemn its algorithm. It has chosen to keep that data private.

The Non-Financial Ledger

The nine named plaintiffs in this case are not statistics. They are real people who went through the process of trying to buy a home, submitted the required paperwork, waited, and received a denial or an offer with worse terms than similarly situated White applicants received. Each of them came to Navy Federal as a member. The institution was built to serve them.

Consider what a mortgage denial costs somebody who is not wealthy. It does not just mean renting instead of owning. It means not building equity while property values increase. It means paying someone else’s mortgage through rent. It means not having a stable asset to borrow against in an emergency, to pass to children, or to sell at retirement. The racial wealth gap in the United States is not abstract. It is constructed, one denied application at a time, across decades of policies and practices that prevented Black and Latino families from accumulating the wealth that homeownership generates.

The named plaintiffs had varying financial profiles. One made several hundred thousand dollars per year and still faced adverse terms. Another put down $80,000 in income. Some had received mortgages from other lenders after Navy Federal denied them, demonstrating that the denial was specific to Navy Federal’s process, not to the applicants’ creditworthiness in the broader market. Six of the nine who were denied went elsewhere and got loans. Navy Federal’s algorithm said no. The market said yes.

One named plaintiff, Constantina Batchelor, was not denied outright. She was approved, but at a higher interest rate than similarly situated White applicants received. Over the life of a 30-year mortgage, a higher interest rate is not a minor inconvenience. A single percentage point on a $300,000 loan adds up to tens of thousands of dollars in additional interest paid over 30 years. That is money extracted from a minority borrower’s household that goes directly to the institution and ultimately functions as a penalty for race.

The proposed class extends well beyond these nine individuals. It encompasses all minority mortgage applicants to Navy Federal from 2018 to the present whose applications were denied, approved at worse terms, or processed more slowly than comparable White applications. The population is enormous. Navy Federal is the country’s largest credit union. The harm, if the algorithm operates as alleged, has been distributed across thousands or tens of thousands of households that were trying to build wealth and were instead turned away or charged extra for the privilege of borrowing.

None of this has been adjudicated. Navy Federal denies the allegations. But the court record makes one thing clear: the company has not offered any public explanation of why its approval rates diverge so dramatically by race. The algorithm remains secret. The weights remain undisclosed. The data that could resolve the question sits inside Navy Federal’s servers, unavailable to the families who may have been harmed by whatever the algorithm actually does.

Legal Receipts

The following passages are drawn directly from the court record in Oliver v. Navy Federal Credit Union, No. 24-1656 (4th Cir. Feb. 9, 2026). These are the court’s characterizations of the complaint’s allegations and the district court’s reasoning.

The majority of the Fourth Circuit disagreed with the district court’s dismissal of the injunctive class. The panel ruled that plaintiffs had stated enough, on the face of the complaint, to make a prima facie showing that a common question exists: does Navy Federal’s single algorithm, applied to all applicants, produce the racial disparity the data shows? That question cannot be answered without discovery. The district court killed the case before the plaintiffs could gather a single document from Navy Federal to find out.

Societal Impact Mapping

Environmental Degradation

The environmental dimension of algorithmic mortgage discrimination is indirect but real. When Black and Latino families are systematically denied homeownership in particular areas, the resulting concentration of renters rather than owners in those neighborhoods correlates with lower political power, less civic investment, and a reduced ability to fight industrial siting decisions. Research on environmental justice consistently finds that communities of color are more likely to be located near toxic facilities, in flood plains, and in areas with degraded air quality. Homeownership is a mechanism for community stability and political voice. Denial of homeownership across a class of military families compounds existing environmental inequalities by keeping targeted communities in lower-leverage positions when fighting polluters, highway expansions, or industrial development.

Public Health

Housing stability is one of the strongest predictors of health outcomes. Families who are denied mortgages and remain renters face greater housing instability, more frequent moves, and higher rates of overcrowding. Each of these factors correlates with worse physical and mental health outcomes: higher rates of stress-related illness, more limited access to primary care tied to stable addresses, worse outcomes for children in school, and greater susceptibility to respiratory illness tied to inadequate housing conditions. A Navy Federal mortgage denial is not just a financial loss. It is a health event whose consequences extend across the lifetime of the applicant and, through housing instability, across the childhoods of their children.

Economic Inequality

The racial wealth gap between Black and White households in the United States is structural, durable, and directly tied to homeownership patterns. Black households have been excluded from wealth-building through homeownership by explicit redlining, by discriminatory lending, by racially restrictive covenants, and by other government and private policies across the twentieth century. Those exclusions compound across generations because wealth builds on itself. A family that could not buy a home in 1960 could not build equity. A family that cannot build equity cannot help their children with down payments. A family whose children cannot receive intergenerational housing wealth has to start from zero in a market where everyone else has a head start.

If Navy Federal’s algorithm is producing what the complaint alleges, it is functioning as a private continuation of that historical pattern: a system that takes the inputs of racial geography, employment type, and institutional relationship history and produces denial at a rate nearly double that seen for White applicants, while keeping the mechanism secret from regulators, researchers, and the applicants themselves.

The 13 million members of Navy Federal include an enormous portion of the active-duty military and veteran population, a group that is disproportionately Black and Latino compared to the general population. Systematic denial of homeownership to those members is not an abstraction. It is a transfer of wealth: from military families who served the country to an institution that charges them more or denies them entirely while approving their White counterparts at a dramatically higher rate.

The Cost of a Life Metric

What Now?

The Fourth Circuit’s February 9, 2026 ruling returns the injunctive class action claims to the district court for further proceedings. The majority held that the district court acted prematurely in killing the Rule 23(b)(2) class before any discovery had occurred. The damages class under Rule 23(b)(3) remains dismissed, at least at this stage, because the individual differences among class members were too apparent on the face of the complaint to support a damages class without further factual development.

What happens next depends on discovery. The single most important question is whether Navy Federal will be compelled to open its algorithm to scrutiny. If the court allows discovery to proceed on the injunctive class, plaintiffs’ lawyers will be able to demand: the algorithm’s source code and documentation; the variables and their weights; internal communications about algorithm design and testing; demographic outcome data broken down by product, geography, and applicant characteristics; and any internal audits or equity reviews the company may have conducted.

Corporate Leadership and Board Accountability

The court record does not name individual Navy Federal executives responsible for the algorithm’s design or the company’s decision to keep it secret. The company is led by its executive management and governed by a board of directors. Any accountability campaign should target those decision-makers directly: the CEO, the Chief Risk Officer, the Chief Technology Officer, and the board members who have authorized and overseen the underwriting system while maintaining secrecy about its function.