The Secret Bias Hidden in Mortgage-Approval Algorithms

By MNR News posted 09-16-2021 07:21


illustration of a happy white male and a confused black maleNon-profit newsroom The Markup recently investigated racial disparities in federal mortgage data from 2019. 17 variables were used, including:

  • Race
  • Sex
  • Whether the application had a co-applicant
  • Age
  • Income
  • Loan amount
  • Property value
  • Debt-to-income ratio

After analyzing the data, The Markup found that comparing loan applications to White applicants, lenders were:

  • 40% more likely to turn down Latino applicants for loans
  • 50% more likely to deny Asian/Pacific Islander applicants
  • 70% more likely to deny Native American applicants
  • 80% more likely to deny Black applicants

The Markup’s report also looks at Fannie Mae and Freddie Mac’s credit scoring algorithm. The algorithm was developed from data from the 1990s and is closely guarded. No one outside Fannie Mae or Freddie Mac knows exactly how the factors in their underwriting software are used and weighted. Not even the Federal Housing Finance Agency, which is the agency that regulates Fannie Mae and Freddie Mac, knows exactly how the software scores applicants.

chart of debt ratio being the most common reason for loan denial among racial groups

Read the full report here, and be sure to review The Markup’s process for how they investigated racial disparities in federal mortgage data here.