Credit scores reflect dramatic and troubling disparities by race. For instance, a 2019 study by the Urban Institute found that over 50 percent of white households have credit scores over 700, but only 20 percent of Black households do. The Urban Institute study is the most recent report showing these startling racial credit scoring gaps; there are a multitude of older studies with similar results, listed in this report. The racial disparities in credit scores are due to a multitude of factors, including the racial wealth gap, decades of redlining and housing segregation, historical and present-day employment discrimination, and racially biased criminal justice practices.
Black and Latinx consumers are also more likely to lack a credit history or have too scant a history to generate a credit score, referred to as being “credit invisible.” The Consumer Financial Protection Bureau (CFPB) found that about 15 percent of Black and Latinx consumers are credit invisible (compared to 9 percent of whites and Asians) and an additional 13 percent of Black and 12 percent of Latinx consumers have unscorable records (compared to 7 percent of whites).
Spurred by the Black Lives Matter protests of 2020 and subsequent racial reckoning, policymakers and advocates have focused on developing solutions to these racial disparities. One of the most touted solutions is using alternative data, i.e., any data that is not traditionally included in credit reports issued by the Big Three credit bureaus (Equifax, Experian, and TransUnion). This includes relatively conventional financial data such as rent, telecom and utility payments as well as bank account cashflow data, which are the topic of this brief. It can also include less conventional data such as web browsing, social media, educational background or “Big Data.”
But alternative data will not eliminate racial disparities in credit scores and is not a panacea for credit inequities. Alternative data may help some credit invisibles, including Black and Latinx consumers, but only if used carefully. Some forms are more promising than others. Most critically, any data that relies on financial information will still reflect racial disparities given the unequal economic positions of households of color and white households. And when financially-based data is fed into algorithms or artificial intelligence models, the results could replicate or amplify those disparities.