Would you rather grant a loan to a borrower who missed a few payments five years ago but has been on the ball ever since or a borrower who has never missed a payment until the past few months, and missed a bunch in a row?
In traditional underwriting models, you wouldn’t get this choice, much less a definitive answer to this question. In this clip from the recent Data Disrupt conference, ZestFinance CEO Douglas Merrill explains why most underwriting models today misperceive true creditor risk by taking a point-in-time approach to creditworthiness. Status quo approaches ask, “How many bankruptcies has this applicant had over the past X years? How many missed payments?”
While these are helpful indicators, this approach misses a key factor: change over time. Zest customers get to deploy advanced machine learning credit models that fully take into account change over time and thousands of other variables to produce a more nuanced and accurate view of your credit risk.
If you’re interested in learning more about why it’s “quite natural” for machine learning models to view change over time and better assess credit risk, email us at firstname.lastname@example.org and check out more videos like this on our YouTube page.