Report Number 425
It’s virtually impossible to get through an issue of a credit union trade publication without encountering coverage about the latest person- to-person (P2P) lending initiative. Whether these new programs come from within our own industry or from the innovators of Silicon Valley, P2P lending apps have captured the attention of credit union leaders.
And, while today’s newest solutions leverage cutting- edge techniques such as data analytics and machine learning to make pricing and underwriting decisions using data beyond a three- digit FICO Score, the roots of lending are anything but high tech. In fact, the practice of relying upon shared experience, word- of-mouth feedback, and trust to make lending decisions goes back to the ancient tradition of lending circles. These circles have given way to more formalized networks, such as merchant credit associations and the omnipresent FICO Score, that allow lenders to share credit and repayment experiences for more informed decision making.
As our ability to store and analyze vast repositories of data has rapidly increased, machine learning is opening up affordable new avenues for making credit decisions based on new types of data. Product usage, member behavior, and even community- level data can now be rapidly digested to offer appropriately priced products to consumers who might not otherwise be deemed credit worthy. For credit unions, which, as an industry, have member service as a primary philosophical objective, the value in these data can extend far beyond simply making credit decisions. Using machine- learning techniques, credit unions can build algorithms that support and reward positive financial behaviors and enhance member loyalty, while they further differentiate themselves from profit- driven competitors.
Filene thanks its generous sponsors for helping to make this research possible.