Fair lending examination procedures are the risk-based, multi-phase process federal regulators use to test whether a financial institution’s credit decisions, pricing, marketing, and servicing practices discriminate against protected groups. The Consumer Financial Protection Bureau, FDIC, OCC, NCUA, and Federal Reserve all conduct these reviews, and each one moves through scoping, an on-site or virtual review, and a formal report. The stakes are real: violations can trigger civil money penalties, borrower remediation, and a mandatory referral to the Department of Justice.
If you work in compliance at a bank, credit union, or nonbank lender, what follows is what to expect at each stage, what examiners will ask for, how they analyze the data, and where institutions most often get tripped up.
The Laws That Frame the Exam
Two federal statutes anchor every fair lending examination. The Equal Credit Opportunity Act, implemented through Regulation B, prohibits discrimination in any aspect of a credit transaction and covers both consumer and commercial loans.1Consumer Financial Protection Bureau. 12 CFR Part 1002 – Equal Credit Opportunity Act Under ECOA, a creditor cannot discriminate based on race, color, religion, national origin, sex, marital status, or age, and cannot penalize applicants whose income comes from public assistance or who have exercised their rights under the Consumer Credit Protection Act.2Office of the Law Revision Counsel. 15 U.S. Code 1691 – Scope of Prohibition
The Fair Housing Act covers the residential side. It makes it illegal to discriminate in mortgage lending, home improvement financing, or the appraising of residential property based on race, color, religion, national origin, sex, disability, or familial status.3Office of the Law Revision Counsel. 42 USC 3605 – Discrimination in Residential Real Estate-Related Transactions
The scope difference matters during an exam. ECOA reaches every credit product, including business loans, auto loans, and credit cards. The Fair Housing Act applies only to residential real estate transactions. An institution that makes no mortgage loans still faces a full ECOA review.
The Three Phases of the Examination
Examinations move through pre-exam scoping, the review itself, and a formal reporting stage. Duration and intensity depend on the institution’s risk profile, product complexity, and history with the regulator.
Scoping and the Information Request
Before examiners arrive, the regulatory team assesses fair lending risk and sets “focal points” for in-depth review. Focal points are shaped by past compliance record, product complexity, market demographics, and whether prior exams flagged concerns. Everything that follows narrows toward those focal points.
The institution then receives a detailed information request, typically well in advance of the on-site phase.4Federal Financial Institutions Examination Council. Interagency Fair Lending Examination Procedures Late or incomplete responses create an immediate negative impression and can expand the scope of what examiners review.
The On-Site or Virtual Review
During the review, examiners work through the procedures defined in scoping. They evaluate the compliance management system, interview key personnel involved in underwriting and pricing, and perform detailed sampling and analysis of loan files pulled from the focal points. The purpose is to verify that written policies are actually followed, that internal controls catch departures from those policies, and that no patterns of differential treatment emerge when similarly situated applicants are compared.
Exit Interview and Report of Examination
The review ends with a formal exit interview where examiners present preliminary findings to management and, typically, the board or a designated committee. The institution can provide context or explain apparent disparities before findings become final. The formal Report of Examination follows, documenting conclusions and any required corrective actions.
What Examiners Ask For
The information request is the single most useful document for preparation, because it tells you exactly what examiners will need to see. Expect to produce:
- Underwriting guidelines and pricing policies
- Descriptions of credit scoring systems, including cutoff scores and override tracking
- Compensation structures tied to loan production or pricing
- HMDA data
- Records of policy exceptions
- Consumer complaints alleging discrimination
- Marketing materials, including digital advertising
- Compliance program documentation, including fair lending training records
If pulling any of this together takes more than a day, that gap will show up in the exam.
How Examiners Test for Discrimination
Examiners combine publicly available data with internal records and statistical methods to test for discriminatory patterns. HMDA data, which captures information about residential mortgage applications including race, ethnicity, and gender, is the primary screening tool for identifying initial disparities.5Consumer Financial Protection Bureau. Home Mortgage Disclosure Act HMDA Examination Procedures That public data is supplemented by application logs, exception reports, and marketing records.
Statistical analysis typically involves regression modeling that tests for disparities in approval rates, pricing, and loan terms while controlling for legitimate credit factors like credit scores, debt-to-income ratios, and loan-to-value ratios. The question is whether a protected characteristic remains a statistically significant factor in the outcome after everything that should legitimately drive the decision has been accounted for.
Proxies for Non-Mortgage Products
Most non-mortgage products do not collect self-reported race and ethnicity. To fill that gap, regulators use the Bayesian Improved Surname Geocoding method, which combines an applicant’s surname with the residential Census tract to produce a probability estimate of race or ethnicity.6Consumer Financial Protection Bureau. Using Publicly Available Information to Proxy for Unidentified Race and Ethnicity The result is probabilistic rather than categorical. Institutions that originate auto loans, credit cards, or other non-mortgage products should understand that the absence of self-reported demographic data does not prevent examiners from testing for disparities.
Matched Pair File Review
Statistical models flag patterns; they do not tell the whole story. Examiners follow up with a comparative file review, sometimes called matched pair analysis, comparing an applicant from a protected group against a similarly qualified applicant outside that group. This is the stage where an examiner reads the actual file notes, sees whether an exception was granted to one applicant but not another, and evaluates whether subjective judgments were applied consistently.
The Discrimination Categories in Scope
Examinations target three broad categories of discriminatory conduct. Institutions sometimes assume that because they do not intend to discriminate, they are safe. Intent is only part of the picture.
Disparate Treatment
Disparate treatment means treating an applicant differently because of a protected characteristic. Examiners look for inconsistencies in how loan officers apply underwriting standards, set pricing, or handle exceptions for similarly situated borrowers. Proof can come from overt evidence, like a discriminatory comment in file notes, but more often it comes from comparative analysis showing that applicants with similar credit profiles received different outcomes without any legitimate explanation.
Disparate Impact
Disparate impact does not require intent. A facially neutral policy that disproportionately harms a protected group can violate fair lending laws even if no one meant to discriminate. Setting a minimum loan amount that excludes borrowers in lower-income neighborhoods correlated with a particular racial or ethnic group is a common example.
The analysis follows a three-step structure. First, a specific policy must be shown to have caused a disproportionate effect on a protected group. Second, the institution can defend the policy by demonstrating a legitimate business interest. Third, even if a legitimate interest exists, the policy still violates the law if a less discriminatory alternative could achieve the same objective. The Supreme Court confirmed this burden-shifting framework under the Fair Housing Act in Texas Department of Housing and Community Affairs v. Inclusive Communities Project (2015).
Redlining and Steering
Redlining is a geographic form of disparate treatment. An institution engages in redlining when it provides unequal access to credit in specific areas based on the racial or ethnic composition of those neighborhoods. Examiners evaluate branch locations, marketing patterns, loan volume by census tract, and Community Reinvestment Act assessment areas to identify whether an institution is avoiding or underserving minority communities. The CFPB’s 2024 fair lending report noted corrective actions requiring mortgage lenders to monitor for redlining risk and develop strategies to attract applications from underserved areas.7Consumer Financial Protection Bureau. Fair Lending Report of the Consumer Financial Protection Bureau, December 2025
Steering happens when a loan officer directs an applicant toward a different product or lending channel based on a protected characteristic rather than the applicant’s qualifications. An applicant who qualifies for a conventional mortgage but gets pushed toward an FHA loan, or a borrower steered into a higher-cost product despite qualifying for better terms, are classic examples.
Marketing and Digital Advertising
Targeted digital advertising creates fair lending risk that did not exist a generation ago. Algorithms that filter the reach of marketing based on user demographics, browsing behavior, or geographic targeting can effectively exclude protected groups from ever seeing a credit offer. Examiners evaluate whether marketing strategies, lead generation, and social media campaigns produce patterns of exclusion that mirror redlining or steering. Institutions relying on third-party lead generators or programmatic ad platforms should be able to demonstrate that targeting criteria do not exclude protected groups, even indirectly through proxies like zip code or browsing profile.
The Compliance Management System Review
Examiners do not just look at individual loan files. They evaluate the compliance infrastructure that is supposed to prevent discrimination before it happens.8Consumer Financial Protection Bureau. CFPB ECOA Examination Procedures Baseline Review Five areas draw the most attention.
Board and management oversight. Examiners look for evidence that the board receives regular updates on fair lending risks, that meeting minutes reflect discussion of compliance matters, and that the institution has dedicated staff and budget for fair lending work. An institution where fair lending is tacked onto someone’s other responsibilities raises red flags.
Policies and procedures. Written policies must cover the full life cycle of every consumer product, including any products introduced since the last exam. Examiners check whether policies address features that carry heightened discrimination risk, such as discretionary pricing, exception authority, and override procedures.
Training. Fair lending training should be tailored to specific job functions, delivered frequently enough to remain current, and extended to third-party service providers. Generic annual training that covers the same ground every year is a common weakness.
Monitoring and audit. The institution should demonstrate ongoing monitoring of lending outcomes for disparities, including regular analysis of exception and override activity. Internal audit should independently test fair lending controls, not simply confirm that policies exist on paper.
Consumer complaints. Examiners review how the institution handles complaints alleging discrimination, including whether complaints are tracked, investigated, and reported to the board.
The CFPB’s 2024 fair lending report confirms that compliance management system deficiencies remain a top finding. In that year, the Bureau directed institutions to develop standardized variable testing during credit scoring model development, adopt a standardized approach for evaluating less discriminatory alternatives when using alternative data, and assess the fair lending risk of any data that might serve as a proxy for protected characteristics.7Consumer Financial Protection Bureau. Fair Lending Report of the Consumer Financial Protection Bureau, December 2025
Adverse Action Notices and Algorithmic Models
When a creditor denies an application or takes other adverse action, Regulation B requires a written notice within 30 days that includes the specific reasons for the decision and a statement of the applicant’s rights under ECOA.9Consumer Financial Protection Bureau. 12 CFR 1002.9 – Notifications The reasons must be genuinely specific. Telling an applicant that the decision was based on “internal standards,” or that the applicant failed to reach a scoring threshold without more detail, violates the rule.
The requirement gets harder when an institution uses AI or machine learning models. The CFPB has stated that algorithmic complexity is not an excuse for vague notices. A creditor that cannot explain why its model denied an applicant cannot legally use that model to make credit decisions.10Consumer Financial Protection Bureau. Consumer Financial Protection Circular 2022-03 – Adverse Action Notification Requirements in Connection with Credit Decisions Based on Complex Algorithms Some creditors use post-hoc explanation methods to approximate why a model reached a particular result, but the Bureau has cautioned that creditors must be able to validate the accuracy of those approximations. A model too opaque for validation creates compliance risk on two fronts: inadequate adverse action notices under ECOA, and fair lending exposure if the model produces discriminatory outcomes that no one inside the institution can identify or explain.
Examiners reviewing adverse action practices will sample denial notices for completeness, compare the stated reasons against the actual file data, and evaluate whether the process for generating reasons accurately reflects the factors that drove each decision. This is now a top examination priority as automated underwriting spreads.
Self-Tests and Self-Evaluations Are Not the Same Thing
Regulation B provides an incentive for institutions to test their own lending practices. A voluntary self-test that is designed to evaluate fair lending compliance, and that creates data not already available in loan files, qualifies for a privilege that prevents regulators and private plaintiffs from obtaining or using the results.11eCFR. 12 CFR 1002.15 – Incentives for Self-Testing and Self-Correction
The privilege carries conditions institutions frequently mishandle. It applies only if the creditor has taken or is taking appropriate corrective action when the self-test reveals a likely violation. Corrective action must identify the policies causing the problem and assess the full scope of the violation. The privilege disappears if the institution voluntarily discloses results to a government agency, uses the results as a defense in litigation, or fails to produce the written records documenting the self-test.
Examiners will ask whether the institution has conducted self-tests or self-evaluations, and they distinguish between the two. Results of self-evaluations that use existing loan file data are not privileged and must be shared. Institutions should know exactly which category their internal analyses fall into before an examiner asks.
Possible Outcomes: MRAs, Enforcement, and DOJ Referrals
Outcomes range from clean findings to enforcement actions.
The most common regulatory response to compliance weaknesses is a Matter Requiring Attention. An MRA is a formal directive identifying a specific deficiency and requiring the institution to develop and implement a corrective action plan within a set timeline.12Board of Governors of the Federal Reserve System. Supervisory Considerations for the Communication of Supervisory Findings MRAs are not optional suggestions. Failing to resolve them can escalate the finding in the next examination cycle.
More serious violations, particularly those showing a pattern of discriminatory conduct, can result in formal enforcement actions such as consent orders and civil money penalties. These may require the institution to pay restitution to affected borrowers, overhaul its compliance program, and submit to long-term monitoring. The CFPB’s 2024 fair lending work included corrective actions requiring institutions to fix redlining problems, ensure accurate appraisal-related disclosures, and improve HMDA data integrity.7Consumer Financial Protection Bureau. Fair Lending Report of the Consumer Financial Protection Bureau, December 2025
When an agency has reason to believe a creditor has engaged in a pattern or practice of discouraging or denying credit applications in violation of ECOA, it is required by statute to refer the matter to the Department of Justice.13Office of the Law Revision Counsel. 15 USC 1691e – Civil Liability A DOJ referral is not discretionary for pattern-or-practice findings; the statute uses mandatory language. The OCC has confirmed this obligation, noting that under both ECOA and Executive Order 12892, it must notify HUD and the DOJ when it has reason to believe a lender has engaged in a pattern or practice of discrimination.14Office of the Comptroller of the Currency. Appeal of Fair Lending Referral, First Quarter 2024 DOJ investigations and settlements can involve penalties and remediation payments well into the millions, along with consent decrees that impose years of federal oversight.
How to Prepare Before the Notification Arrives
Institutions that treat exam preparation as a recurring compliance function fare significantly better than those scrambling in response to a notification letter. The interagency examination procedures describe in detail what examiners will request, so there is no reason to encounter any of those requests for the first time during the exam itself.4Federal Financial Institutions Examination Council. Interagency Fair Lending Examination Procedures
Start with the information request list and work backward. Underwriting guidelines, pricing policies, exception tracking reports, override documentation, compensation structures, HMDA data, complaint logs, training records, and marketing materials should all be current, organized, and retrievable on short notice.
Run regular internal analysis of lending outcomes broken down by race, ethnicity, and gender. Use regression analysis or comparable methods to test for disparities in approval rates and pricing before an examiner does. Where the analysis finds no issues, it demonstrates a functioning monitoring program. Where it reveals potential problems, the institution can take corrective action proactively. Remember that internal analyses using existing loan file data are not privileged and must be shared with examiners if asked. If the institution wants the analysis to be privileged, structure it to meet the specific requirements of Regulation B’s self-testing provision before the work begins.
Prepare key personnel for interviews. Examiners will speak with loan officers, underwriters, and compliance staff to understand how policies are actually applied day to day. Inconsistent answers about exception authority, pricing discretion, or override procedures create exactly the kind of concern that expands an examination’s scope. Staff should understand the written policies, be able to explain how they apply those policies in practice, and know how to escalate fair lending concerns internally.