Millions of Mortgages Exposed

Same Loan.
Different Rate.

Same income. Same loan amount. Same neighborhood. The only difference?

+0.000pp
Hispanic vs White
+0.000pp
Black vs White

Higher rate spread above benchmark. On a $300K loan, that's thousands more over the life of the mortgage.

Worst B/W states: Wisconsin (+0.40pp), Louisiana (+0.40pp), South Carolina (+0.37pp), Michigan (+0.36pp), Illinois (+0.35pp)

15.3Mloans analyzed
51states + DC
2018–2023HMDA · CFPB
Since you opened this page, racial rate gaps have cost minority borrowers approximately $0 in excess interest.

The Rate Gap

Rate spread measures how much above the benchmark rate (APOR) a borrower pays. Higher spread = worse deal. Even after accounting for market conditions and loan terms, the gaps persist.

+0.225ppBlack vs White spread gap in Alabama
+0.122ppHispanic vs White spread gap in Alabama

The Cost of a Rate Gap

A fraction of a percent doesn't sound like much. But over 30 years, it adds up to tens of thousands of dollars. Adjust the numbers to see the real cost.

$300K
6.500%
+0.30pp

White Borrower

6.500%

$1896/mo

Total: $682,633

Black Borrower (same everything else)

6.800%

$1956/mo

Total: $704,079

Extra cost over 30 years:

$21,446

That's $60 more per month, every month, for 30 years.

State-by-State Lending Gaps

Mortgage rate disparities exist in every state. This map shows the average rate spread gap — how much more borrowers of color pay compared to white borrowers, even for the same loan.

AK-2ME+21WA+4MT+3ND+3MN+12WI+40MI+36NY+23VT+8NH+17OR+4ID+8WY+1SD+18IA+20IL+35IN+18OH+26PA+34MA+27RI+34NV+7UT+13CO+5NE-4KS+19MO+19KY+9WV-34VA+17NJ+22CT+25DC+33CA+10AZ+11NM+3OK+6AR+28TN+15NC+2MD+26DE+29HI-13TX+17LA+40MS+27AL+23GA+21SC+37FL+19
Rate gap (bps): ≤0 1–10 11–15 16–20 21–30 >30

Same Income, Different Rate

“Maybe they have lower incomes.” No. Even within the same income bracket, the rate spread gap persists. This chart shows the average spread above benchmark by race, at each income level.

Key Finding

The income explanation doesn't hold up. At every income level, minority borrowers tend to pay higher rate spreads. The gap often widens at higher incomes, suggesting the disparity isn't about ability to pay.

State Rankings

Where are the biggest lending gaps? States ranked by the Black-White rate spread difference. Positive = Black borrowers pay more above benchmark.

Wisconsin

#1
+0.404ppB/W
+0.297ppH/W
450,942 loans

Louisiana

#2
+0.403ppB/W
+0.147ppH/W
421,603 loans

South Carolina

#3
+0.368ppB/W
+0.180ppH/W
426,762 loans

Michigan

#4
+0.358ppB/W
+0.208ppH/W
425,052 loans

Illinois

#5
+0.347ppB/W
+0.316ppH/W
441,249 loans

Pennsylvania

#6
+0.342ppB/W
+0.327ppH/W
430,806 loans

Rhode Island

#7
+0.338ppB/W
+0.423ppH/W
147,714 loans

Washington DC

#8
+0.331ppB/W
+0.147ppH/W
73,352 loans

Delaware

#9
+0.286ppB/W
+0.404ppH/W
139,353 loans

Arkansas

#10
+0.283ppB/W
+0.131ppH/W
306,870 loans

Mississippi

#11
+0.272ppB/W
+0.070ppH/W
67,601 loans

Massachusetts

#12
+0.269ppB/W
+0.312ppH/W
418,284 loans

Maryland

#13
+0.257ppB/W
+0.372ppH/W
406,639 loans

Ohio

#14
+0.255ppB/W
+0.216ppH/W
440,675 loans

Connecticut

#15
+0.249ppB/W
+0.268ppH/W
394,424 loans

Regression Analysis

After controlling for income, loan amount, loan-to-value ratio, debt-to-income ratio, loan type, occupancy type, state, and year — significant racial disparities persist.

Race / EthnicityRate Premium95% CISignificance
Black+7.1 bps(6.2, 8.0)p < 0.001 ***
Hispanic+9.7 bps(8.7, 10.6)p < 0.001 ***
Asian-13.1 bps(-14.1, -12.1)p < 0.001 ***
American Indian / Alaska Native+0.4 bps(-3.3, 4.1)n.s.
Native Hawaiian / Pacific Islander+2.0 bps(-3.3, 7.3)n.s.

Baseline: White non-Hispanic borrowers. "bps" = basis points (1 bps = 0.01 percentage points). OLS regression, N = 1,892,859, R² = 0.0431.

+7.1 bps
Black borrowers pay 7.1 basis points more than White borrowers with the same income, loan size, LTV, DTI, loan type, state & year
+9.7 bps
Hispanic borrowers pay 9.7 basis points more with identical observable characteristics
−13.1 bps
Asian borrowers pay 13.1 basis points less — the "model minority" pattern persists in lending

What We Can — and Can't — Control For

✓ Controlled in this regression

  • • Income
  • • Loan amount
  • • Loan-to-value ratio (LTV)
  • • Debt-to-income ratio (DTI)
  • • Loan type (Conventional, FHA, VA, USDA)
  • • Occupancy type
  • • State fixed effects (all 50 + DC)
  • • Year fixed effects (2018–2023)

✗ Not available in HMDA data

  • Credit scores — the strongest predictor of interest rates. HMDA does not collect credit scores to protect borrower privacy.
  • • Down payment source & reserves
  • • Employment history & job stability
  • • Points/fees paid to buy down the rate

If credit scores fully explained the gap, Black and Hispanic borrowers would need systematically lower scores even at the same income, DTI, and LTV. Research by the Federal Reserve and CFPB finds credit scores explain part — but not all — of the disparity.

Take Action

If you believe you've been charged a higher rate because of your race, you have rights. Share this data, file complaints, and support fair lending.

Share This

Follow the Investigation

New data, updates, and analysis from Justice Index.

Know Your Rights

Shop Multiple Lenders

Getting 3+ quotes can save you thousands. Rates vary dramatically between lenders.

Request Your Loan Estimate

Lenders must provide a standardized Loan Estimate within 3 days. Compare them.

File a Complaint

The CFPB investigates lending discrimination. You can file online in minutes.

Check for Fair Lending Violations

HUD enforces the Fair Housing Act. Contact your local fair housing center.

About This Project

Data Source

This project uses data from the Home Mortgage Disclosure Act (HMDA), administered by the Consumer Financial Protection Bureau (CFPB). HMDA requires most mortgage lenders to report detailed data about each application, including the borrower's race, income, loan terms, and the interest rate charged.

Methodology

We analyze originated mortgage loans (home purchase and refinance) for primary residences across all 50 states and DC, spanning 2018–2023. Rate spreads (above the benchmark APOR rate) are compared by race, stratified by income bracket, loan type (conventional, FHA, VA, USDA), and state. An OLS regression on 1.9 million loans controls for income, loan amount, LTV, DTI, loan type, occupancy type, state, and year — see the Regression Analysis section above.

Limitations

  • HMDA does not include credit scores (the strongest predictor of rates)
  • Unobserved factors (down payment source, reserves, employment history) may differ
  • Rate differences shrink but persist after controlling for available factors
  • This is observational data — we document disparities, not prove discrimination

Justice Index · Three Investigations

Bias doesn't stop at lending. It follows people from the traffic stop to the courtroom to the bank.