Fair Lending Data

Fair lending is central to the principles under which the U.S. housing finance system operates and is a requirement of law. FHFA’s fair lending activities include monitoring the regulated entities for fair lending and fair housing risk and conducting examinations on their policies, programs, and activities. FHFA monitors loan application accept rates and other trends for fair lending risk and compliance.

The data are provided for public transparency and to promote fair lending, but do not by themselves prove or disprove unlawful discrimination. 

Statistics Reported

The accept rate represents the proportion of applicants who were approved by the AUS and whose loans are eligible for purchase based on their credit characteristics. It does not represent final credit decisions concerning applicants, which are made by lenders. The accept rate is influenced by the population of borrowers applying for mortgage credit. The number borrowers who are submitted to the AUS is constantly changing and fluctuates with changes in market interest rates, affordability, household formation, lender selection, and other market factors. Additionally, lenders may use one or both AUS systems to assess an applicant.

The accept rate gap represents the raw difference between the protected class accept rate and the comparison group (Non-Hispanic white) accept rate. The accept rate ratio represents the protected class accept rate divided by the comparison group accept rate. These statistics help FHFA to assess the impact of changes in the population applying for credit and policy changes in the AUS.

Loans that are accepted by the AUS and originated by lender may or may not result in an Enterprise loan acquisition, or funding. Enterprise loan acquisitions are influenced by a variety of factors such as the lender’s propensity to securitize or retain loans on the balance sheet, pricing and execution, as well as other economic variables. Loan acquisition data are displayed by race and ethnicity along with the proportion (or share) of loan acquisitions for that quarter.

Race and Ethnicity

The Enterprises define race and ethnicity using non-mutually exclusive definitions and applications are counted with each race or ethnicity reported by a borrower or co-borrower. For example, if there are two borrowers on a mortgage, one of whom identifies as Black and one as Asian, the loan information would be counted in both Black and Asian borrower categories. The exception is that white borrowers include only those reported as non-Hispanic white and no other race. Borrowers with missing race and ethnicity information are excluded from the population when calculating the proportion of loan acquisitions.

For race and ethnicity categories, FHFA’s Division of Public Interest Examinations and Office of Fair Lending Oversight uses the following naming conventions: “White” refers to Non-Hispanic white applicants or borrowers; “Black” refers to African American or Black borrowers; “Latino” refers to Hispanic or Latino borrowers of any race; “American Indian” refers to American Indian or Alaska Native borrowers (AIAN); “Pacific Islander” refers to Native Hawaiian or Pacific Islander borrowers; “Asian” refers to Asian borrowers.

AUS Data

The underlying data consist of application data for each Enterprise's Automated Underwriting System (AUS) that have been transmitted to FHFA. Fannie Mae uses Desktop Underwriter (DU), and Freddie Mac uses Loan Product Advisor (LPA) as their respective AUS system. The data represent the AUS recommendation for the last transaction submitted to the AUS, excluding applications that were not scorable, applications for FHA/VA loans, and applications identified as test cases.

Download: Fair Lending aggregate data and documentation [ZIP - 266 KB]

Page Last Updated: February 21, 2024​​​​​​