Domain: Financial Services/Lending
Description: Bartlett et al. analysis of HMDA dataset revealed systematic discrimination in mortgage lending, with Black and Latino borrowers paying higher interest rates despite similar risk profiles.

Ethical Challenges:

• Racial/Ethnic Bias: Black and Latino borrowers pay 7.9 and 3.6 basis points more respectively
• Economic Impact: Results in $765M+ annually in excessive interest payments
• Digital Redlining: Algorithmic discrimination in historically disadvantaged neighborhoods
• Intersectional Bias: Particularly affects African American women

Public Datasets:

• Primary: Home Mortgage Disclosure Act (HMDA) Dataset
• URL: https://www.consumerfinance.gov/data-research/hmda/
• Data Portal: https://ffiec.cfpb.gov/data-publication/
• Historical Data: https://catalog.data.gov/dataset/home-mortgage-disclosure-act-hmda-public-data-from-2007-2017
• Source: Federal Financial Institutions Examination Council
• Content: 88-90% of all U.S. mortgage loans since 1975, with enhanced data since 2017
• Includes: Loan applications, approvals, denials, applicant demographics, loan terms, property location, lender information

Additional Dataset:

• Supplementary: Fannie Mae and Freddie Mac Loan-Level Dataset
• URL: https://www.kaggle.com/datasets/thedevastator/2016-fannie-mae-and-freddie-mac-loan-level-datas