For decades, there has been a reasonably durable consensus across administrations led both by Republicans and Democrats: discrimination in lending is wrong, and federal law should prevent not only obvious bigotry but also practices that, in effect, lock out communities and people of color, women, older Americans, and others from the credit they need to thrive. That consensus has never been perfect, and enforcement has sometimes waxed and waned, but the foundation has been clear. The Equal Credit Opportunity Act (ECOA) and Regulation B, interpreted in light of their legislative history and case law, prohibit both intentional discrimination and policies that have discriminatory effects, bar practices that discourage applicants, and explicitly allow thoughtfully designed Special Purpose Credit Programs (SPCPs) to close enduring gaps.
In our comment letter on the proposed changes to ECOA, we are clear. “Discrimination, both intentional and not, remains pervasive and persistent, despite significant advances over the past 50 years. Eliminating how we measure, identify, protect against, and act upon discrimination in lending does not eliminate discrimination; it only makes it invisible and unchecked.”
Most of the worst discrimination in lending over the last half‑century has not taken the form of someone saying, “We will not lend to Black borrowers,” or “We won’t serve women business owners.” It has taken the form of policies, models, and marketing strategies that are written in neutral language but produce sharply unequal outcomes.
A branch network that clusters only in White suburbs. A credit scoring model that relies on data reflecting generations of residential segregation and wealth gaps. A digital marketing strategy that targets ads in ways that show prime credit offers almost exclusively to White households. A pricing grid that quietly yields higher interest rates for Black and Latino borrowers even after controlling for credit risk. None of those examples require a single explicit statement of bias. But their impact is unmistakable if we measure and acknowledge it. And once we measure and know it, it cannot be unknown. The onus is on the institution to correct the bias, and on regulators to enforce the law.
That, at its core, is what disparate impact analysis does. It seeks to find if a facially neutral policy has unjustified, adverse effects on a protected class. Under ECOA’s long‑standing interpretation, regulators, advocates, and lenders have been able to examine denial rates, pricing, and access across race, gender, age, and other protected characteristics and ask two simple questions: Does this policy fall more harshly on a protected group, and if so, is there a legitimate business necessity that cannot reasonably be achieved in a less discriminatory way?
But the Trump administration has proposed a rule that would effectively gut ECOA by eliminating the very methods it uses to identify and correct for discrimination. It would insist that only intentional discrimination counts and that discriminatory outcomes, no matter how consistent or severe, do not warrant oversight or enforcement. In doing so, it would ignore the plain meaning of the House and Senate reports that accompanied the 1976 ECOA amendments, which explicitly envisioned courts and agencies examining the effects of creditors’ practices, and it would depart from decades of consistent regulatory practice and court decisions that have treated disparate impact as cognizable under ECOA.
That is not an adjustment. It is a repudiation of the way fair lending has actually been enforced for 50 years.
There’s an old adage that a regulator can be a lap dog, a guard dog, or an attack dog. Like many of my colleagues, I find the guard dog model works best. In the previous administration, that was most definitely not the model. But recent regulatory actions by the Trump administration on fair housing, civil rights, and consumer protection would result in no dog at all. The foxes will have free range of the chicken coop, and the chickens will have to fend for themselves. We know how that story ends, and it is not a happy one for the chickens. The proposed changes to ECOA allow the discriminatory foxes to have the run of the market, harming not just consumers but lenders as well.
In the vacuum created by the Consumer Financial Protection Bureau’s retreat from the accepted interpretation of discrimination under ECOA, as well as the abandonment of its oversight and enforcement duties to both consumers and the market, bad actors will become more aggressive, introducing more risk and undermining the stability of the overall market and economy.
The same is true for the legal issue of “discouragement,” which like disparate impact, is stripped of its meaning by the proposed rule. We all know that a “statement” is not just what you print in a brochure. It is which neighborhoods you open branches in, who sees your ads, and how your creative choices make someone feel welcome—or not.
The proposal would narrow discouragement to “oral or written statements,” and even suggests that “selective encouragement” of one audience is not discouragement of others who never receive the message. In a world where digital platforms can target with razor precision, we should be clear‑eyed about how exclusion can happen without a single ugly word ever being spoken.
One of the most important tools lenders have used to better serve underserved markets are Special Purpose Credit Programs, or SPCPs. These programs are one of the clearest examples of how fair lending and business growth reinforce each other. ECOA expressly authorizes SPCPs because Congress understood that meeting special social needs—like closing persistent gaps in access to credit—sometimes requires targeted, well‑documented programs.
For years, lenders have used SPCPs to expand market share responsibly, manage risk, and reach qualified consumers they were missing. The proposed rule would sharply limit that flexibility for lenders and raise the evidentiary bar so high that many would simply walk away. That would be a loss for lenders, for consumers, and for communities that benefit when more neighbors can buy a home, build equity, and participate fully in the local economy.
The Supreme Court decision in Loper Bright Enterprises v. Raimondo ended the requirement for courts to defer to federal agencies’ interpretations of ambiguous statutes and instead defer to judicial interpretation. Judicial interpretation on these ECOA provisions has been clear and consistent in federal courts across the country. But the end of mandatory deference does not relieve agencies of responsibility; it heightens it. When statutory text, legislative history, and decades of judicial and regulatory practice point in the same direction, an agency should be especially cautious about announcing a new reading that contracts civil rights protections. Courts will no longer rubber‑stamp those interpretations. They will look more closely at whether the agency is straying from what Congress intended.
In the case of ECOA, that record is clear: Congress knew about disparate impact theories, explicitly endorsed looking at the effects of credit practices, and repeatedly amended ECOA without disturbing that understanding. It authorized SPCPs in statute, across all protected classes, and gave lenders a safe harbor to use them. It understood discouragement as broader than just explicit statements. To read ECOA now as if none of that mattered is not textual fidelity; it is racial revisionism.
The Trump administration asserts that discrimination is a thing of the past, and fair housing and civil rights enforcement is an onerous, even dangerous relic. That would be great if it were true, but the data say otherwise. Refusing to consider the data does not alter that reality. While minority homeownership rates have increased since these and other anti-discrimination policies were put into place more than 50 years ago, the homeownership gap between Black and White households is worse than when segregation was legal. In 1960, the Black homeownership rate was 38 percent compared to 65 percent for white households (a 27 point gap), and today (3Q2025) the Black homeownership rate is at 45.7 percent compared to 74 percent for White households (a 28.3 percent gap). It is clear that anti-discrimination protections, as well as targeted programs and products, are still necessary to identify and protect against intentional, unintentional, and covert discrimination, and to achieve equality in lending.
In 2022, the Justice Department secured a groundbreaking settlement with Meta over the delivery of housing ads on Facebook, which required the company to stop using certain tools and build a new system to reduce disparities in who actually sees an ad. That case did not accuse a product manager of writing a slur into an algorithm; it asked whether the output delivered by a complex, adaptive system reproduced protected‑class disparities. That is exactly the kind of question an effects‑based lens is built to answer—and exactly the kind of issue we’ll see more often as AI becomes standard in underwriting, marketing, and servicing.
Another factor is even more insidious: the multigenerational impact of racial discrimination. I know this from my own experience. I grew up in a home my parents owned, who instilled in me a belief that I was expected to own my home as well. When I was ready, I received valuable guidance from my parents that made the entire process easier. I also received help from what I call the “Daddy Downpayment Assistance Program.” My parents helped me pay for my downpayment and closing costs—help the lender was happy to accept without considering the fact that I had saved less than other applicants.
That was a great opportunity for me, and I am deeply grateful for it. But it is not one shared by everyone. My father was the first person in his family to own their own home—ever. His father was a junk dealer, and my dad worked in his junkyard. My dad’s life changed, and with it the economic trajectory of our family, when, after serving in World War II, he attended college and bought his first home with government benefits of the GI Bill of Rights.
But unlike him, another corporal in the U.S. Marine Corps, a Black corporal he never met because they were not even allowed to serve together, was unable to use these benefits because of discrimination by the government, the banking industry, and society at large. His son did not get the Daddy Downpayment Loan, made possible from the wealth earned with a government‑paid‑for college education and home equity built by a government‑backed and subsidized loan. Nor did he get advice and encouragement from experienced parents. He is the second‑generation victim of explicit and intentional discrimination. ECOA does not offer them something people like me didn’t get. Quite the opposite. They provide the same support I already got and they never did. SPCPs and fair housing protections make fairness and equal opportunity possible.
To its credit, this administration has been explicit about wanting to expand housing supply, grow homeownership responsibly, and reduce barriers that keep families from building wealth. Those are goals NHC shares. If the Bureau’s objective is clarity, predictability, and fairness, there is a path to get there: keep the parts of Reg B that work, like disparate impact, discouragement provisions, and SPCPs.
We know what happens when lenders, advocates, researchers, and public officials sit down together and try something new. Over five years ago, NHC launched the Black Homeownership Collaborative’s 3by30 initiative with partners across the housing ecosystem. We did not agree on every policy. We did agree to put goals on paper, measure what works, and adjust when it doesn’t. That approach—evidence‑driven, collaborative, and focused on outcomes—is exactly what we need now.
We can get there. But we will not do it by weakening the tools that have helped us make real, if incomplete, progress. We will do it by refining those tools, aligning the rules, and doubling down on the simple idea that equal access to credit is good for families, good for lenders, and good for the country.