We can be both fast and fair
This article originally appeared in Credit Union National Association
AI-based underwriting creates better, more inclusive lending decisions.
This century has now seen two of its biggest-ever financial crises. The first, the Great Recession in 2008, was arguably one of our own makings.
The second, COVID-induced, happened to us. But both provide important lessons as the credit union movement prepares for the future and looks to better serve its members.
All lenders, especially credit unions, should be able to look at borrower data from 2008 and quickly see what we could have been done better. While more risk-averse credit unions fared better than other lenders, for many the consequences of the Great Recession resulted in people losing their homes and members’ lives being turned upside down.
The credit union movement promised to do better during the pandemic—and it has. It advocated for forbearance and skipped payments that kept people afloat. Credit unions made emergency loans and offered financial counseling.
Even the regulators this time said, “Go out and help your members; don't just let them sink or swim.”
We now have an opportunity to bake this experience and empathy into our ongoing practices so we can continue supporting people through financial challenges.
To do this successfully and proactively we need to get out of our comfort zone and embrace new approaches. Hundreds of credit unions, for example, are obtaining the community development financial institution (CDFI) designation, which will allow them to do more to help the underserved.
Programs like CDFI, along with new tools and policies, make it possible to find the people who need help but aren’t getting it.
But nothing will happen unless we invest in change. Six of 10 credit union professionals surveyed at the annual National Association of Credit Union Service Organizations (NACUSO) conference in November said their underwriting technology is either outdated or very outdated.
Technology is now at the point where predictive risk models can make far more accurate and faster loan decisions when the numbers are clear. That frees up lending teams to work on cases that aren’t so clear.
Say a family experiences a financial event in which one of the breadwinners is sick for an extended period.
“What do we do to keep them going so when that event is over, they're still successful? The loyalty that comes with solving dilemmas like these is huge,” says Alice Stevens, vice president of credit administration at $3.5 billion asset Truliant Federal Credit Union in Winston-Salem, N.C., which recently earned a CDFI designation to improve community outcomes. “And we can do it if we upgrade our tools and techniques.”
According to the NACUSO survey, 80% of credit union executives believe the increased use of artificial intelligence (AI) and machine learning would lead to better credit scoring. Many are taking action to modernize.
Most of those surveyed said they’re making AI-based underwriting an investment priority this year.
Truliant Federal recently updated to AI-based risk models for its lending. Through a pilot program, the credit union found the tool could help increase automated approvals by as much as 30% across diverse segments of the member communities it serves.
Credit unions and other lenders now realize they need to deliver as fast a turnaround with a decision as their members are used to getting with their ride-sharing and food-delivery apps. But credit union leaders owe it to their organizations to ensure the decisions are prudent, fair, and consistent.
I’d like to see the movement’s biggest credit unions share how they’re implementing new approaches to serve as a guide for smaller organizations that may not have access to the same resources.
We don’t want another crisis. But if it comes, we want the movement to be more efficient and nimble in how we respond.
We should devote less time to the work we can automate so we have more time to treat people facing a tough spot with the respect they deserve.
Those who have been around the movement a long time know it can be done better. It’s finally possible: We can be both fast and fair.
MIKE de VERE is CEO at Zest AI, a CUNA Strategic Services alliance provider.