Giving Credit Invisibles a Chance: How AI Can Help
by Kristina Drye, on May 11, 2022 11:45:00 AM
Credit and its numerical representation, the credit score, are the most important means of access to financial stability in the United States. Three main credit bureaus (Equifax, Experian, and TransUnion) determine who has access to consumer credit opportunities- and, more importantly, who does not.
However, credit scores present a problem for financial institutions. According to the Consumer Financial Protection Bureau, one in ten adults, almost 26 million Americans, do not have credit. This is known as “credit invisibility”. If you increase this number to include people with “thin files” (not enough credit information to produce a score, but not a complete absence of credit activity), that number increases to over 100 million U.S. consumers.
All Credit Invisibles Are At Risk...
Without a credit history, a person is invisible to a lender. Because credit is a record of information regarding financial trustworthiness, providing a rough risk threshold to the lender, someone with no credit has little information for a lender to use when deciding to issue a credit card, mortgage, loan, or even rentable capital. Many times, lack of credit can affect the most basic daily requirements, like getting employment or attending college. To compound the issue, credit unworthiness is concentrated in already-underserved populations, resulting in disadvantaged groups that are unable to build wealth and achieve financial success.
Because credit scores are used to automate lending decisions, anyone who is credit invisible cannot be automatically approved or denied for consumer credit, and the process to overcome this automation is lengthy. Emergency loans, or any loan, can be difficult to achieve, and there are higher security deposits for those without a credit score, if a landlord or renter is willing to sign with anyone who is credit invisible or thin-filed.
...But Not All Credit Invisibles Are A Risk
Meanwhile, for a financial institution, not serving credit invisibles and thin-filed customers poses a threat. A company’s reputational risk relies on its ability to fairly and equitably provide services to all viable customers. Many assume that those with thin or nonexistent credit files are high-risk consumers, but that’s not true- while most are in already underserved populations, underserved populations as such don’t serve as a reliable or ethical indicator of risk.
Luckily, artificial intelligence and machine learning can help solve these problems. Because consumer credit is issued as an approximation of risk, lenders make assessments about reliability based on a pattern of behavior associated with loans, credit card accounts, and other forms of consumer credit. But formally documented behavior isn’t always the best indicator of risk- alternative sources of data, primarily internet data and negative news - are more reliable risk indicators.
AI Can Help
With AI and tools like GOST (Giant Oak Search Technology), banks and organizations can improve their ability to identify the true risk and reliability of borrowers. This technology can help financial institutions use alternative customer data to automate decisions at scale, without denying those who deserve service. Meanwhile, an institution can also use these capabilities to advance the ESG goal of increasing service to underserved populations, while growing its viable customer base.