creating a fairer system
Proper assessment of fair lending and credit models demands a better yardstick
Current techniques are misclassifying
the race of more than 22 million Americans
Zest Race Predictor
An open-source algorithm that improves your ability to predict race in fair lending analysis
see our resourcesZRP: Validated on a national sample in 2021
ZRP is an ML model that more accurately predicts someone’s race and ethnicity than the current standard (BISG). All you need is a full name and location.
Case Study
26%
More African-Americans correctly identified
“This is a free tool any lender can use when they undergo a fair lending or CFPB review, or release a new model. People will feel better and more responsible knowing they got a better race and ethnicity assessment.“
Jay Budzik
Chief Technology Officer
Chief Technology Officer
35%
Fewer African-Americans identified as non-AA
60%
Fewer Whites identified as non-White
Case Study
The results with Zest AI are impressive, increasing approvals 26% for low-income designated loans, meaning we could deliver more to GreenState members who deserve better.
Amy Henderson
Chief Consumer Services Officer
Chief Consumer Services Officer
32%
Increase in approvals for protected classes
$132M
Annual increase in originations
$11M
additional profit per year
Mislabeling millions of Americans
One simple fix could have an immediate impact on the Black homeownership rate
The era of the fairer credit score
Public sentiment has shifted and raised expectations of financial institutions' role in equality.
New ways to reduce bias
Zest's fairness innovation helps eliminate discrimination in lending.
